Using 139 fb⁻¹ of 13 TeV proton-proton data, ATLAS simultaneously unfolded 24 correlated Z+jets kinematic observables with the machine-learning method OmniFold and released the particle-level result as an unbinned event dataset. The measured total fiducial cross section is 1,808 ± 42 fb; most illustrative differential bins have 3–5% total uncertainty, rising to about 15% in tails. The measurement is more precise than the compared predictions for many distributions, especially Sherpa. The event-level release also supports derived observables, including an infrared-and-collinear-safe τ₂₁ measurement in a specified fiducial region.
Technology: OmniFold iterative machine-learning unfolding
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Using 139 fb⁻¹ of 13 TeV proton-proton data, ATLAS simultaneously unfolded 24 correlated Z+jets kinematic observables with the machine-learning method OmniFold and released the particle-level result as an unbinned event dataset.
- Still unknown: This is a new high-dimensional precision measurement of known Z+jets production, not a new particle or beyond-Standard-Model discovery. The released measurement is unbinned in 24 measured variables; displayed plots may be binned for presentation. The authors report mostly 3–5% uncertainties in illustrative bins, up to about 15% in tails, with dominant unfolding-prior/hidden-variable uncertainties for some observables. Pseudo-data closure is generally good, with exceptions including a p-value of 0.038 for leading-jet transverse momentum; users are given phase-space and covariance-use guidance.
- AI role: OmniFold used neural-network classifiers to iteratively reweight simulated events so the detector-level simulation matched collision data, then propagated the learned correction to particle-level events. That multivariate unfolding handled 24 correlated observables simultaneously and produced event weights with uncertainty variations, enabling a high-dimensional unbinned measurement and event-level construction of additional derived observables. The paper validates closure on pseudo-data and reports its uncertainty components.
Evidence sources: https://journals.aps.org/prl/abstract/10.1103/PhysRevLett.133.261803
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 10% of works of the same type, year and subfield (OpenAlex work metrics, 44 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A virtual-screening workflow combined XGBoost and graph-neural-network models to screen 11,576 DrugBank compounds. Researchers selected 15 candidates for evaluation; four met the study’s activity cutoff against Mycobacterium tuberculosis H37Rv. Aldoxorubicin and quarfloxin were further tested against multidrug-resistant isolates, intracellular bacteria and recombinant DNA gyrase. These are laboratory repurposing results, not evidence of patient benefit or clinical safety. A same-team patent-family publication predates the journal article; this record makes no priority claim over patent disclosures. Concentration values are omitted because of inconsistent units in the rendered tables.
Technology: Ensemble virtual screening with XGBoost and graph neural networks
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: The article reports follow-up testing of 15 selected compounds, four meeting the authors’ activity cutoff against H37Rv, and additional assays of aldoxorubicin and quarfloxin against multidrug-resistant isolates and recombinant DNA gyrase.
- Still unknown: All reported results are in vitro. The study does not establish patient efficacy, clinical safety or the compounds’ cellular mechanism. Precise concentration values are omitted because table units are inconsistent; an earlier patent-family publication is noted without a priority claim.
- AI role: Multiple learned models, including XGBoost and graph neural networks, predicted antitubercular activity in a DrugBank screen and contributed to selection of candidates for experimental evaluation.
Evidence sources: https://academic.oup.com/bib/article/26/1/bbae696/7934648
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review date: 17 Nov 2026. Review scope: Primary Oxford Academic full indexed article, including abstract, Methods, results, article history and author affiliations; patent-family chronology checked from the patent family page. A-version patent text was not reviewed..
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A deep protein language model used prior single- and multiple-mutant measurements to guide combinations across two design/test rounds. The 13-substitution variant 13M4 reportedly retained nearly wild-type catalytic activity and increased apparent melting temperature by 10.19 °C. This assay does not establish long-term stability across production conditions, and the model recombined previously measured mutations.
Technology: Deep protein language model
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Publisher article and PubMed metadata report 13M4, near-WT catalytic activity and apparent melting-temperature increase.
- Still unknown: Prior exact-sequence reporting and independent replication not established.
- AI role: Protein language model guided combinations of previously tested mutations for experimental stability and activity assays.
Evidence sources: https://doi.org/10.1002/mlf2.12151
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 16 Nov 2026. Review scope: Primary source checked for identity, date, model role and bounded reported result..
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Researchers used a deep-learning composition model to prioritize a sigma-phase ternary compound, synthesized Mo20Re6Si4, and measured bulk superconductivity with a critical temperature of 5.4 K. Magnetic susceptibility and a specific-heat anomaly support the bulk transition. The paper reports mixed model performance on its small test set and describes the model as guidance alongside chemical intuition.
Technology: Deep-learning neural network that classifies compositions as superconducting and predicts critical temperature
Publication: Preprint · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Preprint
- Test type: direct test
- Tested: Researchers used a deep-learning composition model to prioritize a sigma-phase ternary compound, synthesized Mo20Re6Si4, and measured bulk superconductivity with a critical temperature of 5.4 K.
- Still unknown: The result is experimental and supported by bulk measurements, but the preprint reports a small, specialized evaluation set and mixed predictor performance. Chemical intuition also guided candidate selection; AI's contribution was a search hint, not independent autonomous discovery. The compound is a materials-physics finding rather than a new general superconducting mechanism.
- AI role: Provided composition-based hints that helped researchers select Mo20Re6Si4 for synthesis; laboratory resistivity, magnetic-susceptibility and specific-heat measurements established the superconducting result.
Evidence sources: https://arxiv.org/abs/2412.13012
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review scope and date are not recorded; source checked does not establish full-text review.
Editorial assessment: Impact 2/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Deep learning clustered cognitive and functional trajectories in 283 people with early dementia probably due to Alzheimer disease into fast- and slow-progressing groups. The trajectory groups were externally validated and independently replicated in a 2,779-person dementia cohort. A classifier using cross-sectional data at diagnosis predicted progression group in external validation with AUC 0.70 ± 0.01.
Technology: Deep-learning clustering of longitudinal cognitive/functional trajectories, followed by machine-learning classification from cross-sectional diagnosis data
Publication: Peer-reviewed paper · Test context: replication or operational use. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: replication or operational use
- Tested: The trajectory groups were externally validated and independently replicated in a 2,779-person dementia cohort.
- Still unknown: These are prognostic progression groups, not established biological disease subtypes or treatment-response predictors. The classifier had modest external discrimination (AUC 0.70 ± 0.01). Trial-size reduction and cost savings were estimated by emulating hypothetical enrichment and were not demonstrated in an actual clinical trial. The study does not show improved patient outcomes.
- AI role: Deep learning grouped patients using multivariate disease trajectories; the authors then trained a classifier to predict those progression subgroups from measurements available at diagnosis. The model-supported fast/slow progression groups were externally validated and independently replicated in a separate dementia cohort.
Evidence sources: https://pubmed.ncbi.nlm.nih.gov/39713242/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 10% of works of the same type, year and subfield (OpenAlex work metrics, 23 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 5/5 · Novelty 2/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
An active-learning loop combined machine-learning prediction, surrogate optimization and experimental testing to explore lead-free potassium-sodium niobate (KNN) ceramics. After three experiment and optimization iterations, the authors report identifying a composition with a measured piezoelectric coefficient d33 of approximately 353 pC/N from thousands of possible compositions. This is a laboratory material-property measurement, not validation of a finished sensor or device.
Technology: Support vector regression with a radial basis function kernel in an active-learning experimental-design loop
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: After three active-learning-guided experiment iterations, the authors report a synthesized KNN composition with measured d33 of approximately 353 pC/N.
- Still unknown: The reported coefficient is a laboratory material-property measurement, not device-level qualification or commercial performance.
- AI role: An SVR model predicted KNN piezoelectric response and guided composition selection across three experiment-optimization iterations.
Evidence sources: https://www.sciencedirect.com/science/article/pii/S0272884224048338 · https://research.mpu.edu.mo/en/publications/multi-iteration-active-learning-for-the-composition-design-of-pot/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review date: 17 Nov 2026. Review scope: Primary publisher indexed full text/abstract and primary institutional repository metadata/abstract for article date, author affiliation, model workflow and experimental result..
Editorial assessment: Impact 2/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A deep-learning radiomic pipeline generated a pretreatment CT response score (CTRS) for advanced non-small-cell lung cancer treated with PD-(L)1 checkpoint inhibitors. In an independent real-world test cohort, high-response-score first-line ICI-monotherapy patients had longer progression-free and overall survival after adjustment for baseline covariates including PD-L1 (adjusted HR 0.46 and 0.50). In a secondary analysis of 54 participants from a prospective sasanlimab trial, the score was associated with overall survival (adjusted HR 0.33) but not progression-free survival.
Technology: Three-dimensional hierarchical convolutional neural network feature extractor combined with a Cox survival model (CT response score, CTRS)
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: In an independent real-world test cohort, high-response-score first-line ICI-monotherapy patients had longer progression-free and overall survival after adjustment for baseline covariates including PD-L1 (adjusted HR 0.46 and 0.50).
- Still unknown: The real-world data were retrospective. Association of the score with survival among treated patients does not establish that the score predicts differential treatment benefit versus an untreated or alternative-treatment control. Clinical-trial analysis was secondary and small (n=54): PFS association was not significant, while OS association was significant. Authors say larger prospective studies are needed; no improved-care or deployment outcome is shown.
- AI role: The model learned image features from pretreatment CT/PET-CT scans and combined them with a Cox survival model to produce an imaging-only response score. Researchers evaluated the score in a held-out real-world cohort and in a secondary analysis of a prospective clinical-trial cohort; it stratified PFS/OS in the first and OS in the latter.
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC11658027/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 10% of works of the same type, year and subfield (OpenAlex work metrics, 19 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 2/5 · Novelty 2/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A closed-loop workflow combined high-throughput synthesis of organic semiconductors, measured photovoltaic-device performance, molecular descriptors and Bayesian optimization to search a virtual library of about one million candidate hole-transport materials. The machine-learning-guided iterations identified high-performing molecules; after device optimization, a perovskite solar cell using a selected material reached 26.23% power-conversion efficiency (25.88% certified). The work links candidate molecular structures to measured device performance in this specific cell system.
Technology: Molecular-descriptor machine-learning models with Bayesian optimization in a high-throughput synthesis and device-testing loop
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: A closed-loop workflow combined high-throughput synthesis of organic semiconductors, measured photovoltaic-device performance, molecular descriptors and Bayesian optimization to search a virtual library of about one million candidate hole-transport materials.
- Still unknown: The 26.23% champion and certified 25.88% values followed device-specific optimization of a selected molecule; this does not establish that the molecule alone accounts for the full efficiency gain or that it is globally optimal among all possible molecules. The paper reports a materials-selection workflow for a particular p-i-n perovskite device setup, with performance also affected by other cell layers and processing. The million-member search was virtual and only a small selected subset was synthesized and measured; authors do not claim to have found the single best molecule in that space.
- AI role: Models trained on measured data from 149 synthesized molecules linked calculated molecular descriptors to perovskite-device performance. Bayesian optimization selected new structures from a virtual library of roughly one million; high-throughput synthesis, characterization and device tests supplied feedback for two closed-loop iterations. Three promising molecules then received individual device optimization. This workflow identified hole-transport materials yielding the reported high-efficiency perovskite cells; the AI contribution was candidate selection and structure–performance modeling, with synthesis and measurements performed experimentally.
Evidence sources: https://www.science.org/doi/10.1126/science.ads0901
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 201 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
The PICNIC gradient-boosting model ranked proteins for membership in biomolecular condensates using sequence-pattern and AlphaFold2-derived structural features. To test its predictions, researchers expressed 24 poorly characterized human proteins in U2OS cells: 21 formed mesoscale fluorescent foci, and 18 met the paper’s high-confidence condensate threshold. The experiments mapped examples to cytoplasmic bodies, nuclear bodies, nucleoli or P-bodies; for example, TYW5 and RAD51AP1 showed distinct localization patterns. Across the tested set, the authors found condensate localization was not limited to highly disordered proteins.
Technology: PICNIC, a gradient-boosting machine-learning classifier using protein sequence co-occurrence patterns, sequence complexity/disorder features and AlphaFold2-derived structural features
Family AlphaFold: profile · models page
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: To test its predictions, researchers expressed 24 poorly characterized human proteins in U2OS cells: 21 formed mesoscale fluorescent foci, and 18 met the paper’s high-confidence condensate threshold.
- Still unknown: The experimental test used fluorescently tagged proteins expressed in U2OS cells, so observed localization may depend on expression level, tag, cell type or condition. The study did not establish each protein’s endogenous localization or biological function; the authors explicitly state that condensate localization alone makes no functional claim. Three of 24 proteins did not form detectable mesoscale foci, and three additional results were low-confidence under the paper’s size threshold. This is targeted experimental validation of model-selected candidates, not an independently replicated atlas.
- AI role: PICNIC scored human proteins for likely cellular-condensate membership and selected poorly characterized positive predictions for experimental testing. Its rankings directly guided the 24-protein U2OS localization validation; the model was built to predict condensate members (including clients), not only proteins that self-drive phase separation.
Evidence sources: https://www.nature.com/articles/s41467-024-55089-x.pdf
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 10% of works of the same type, year and subfield (OpenAlex work metrics, 52 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
The XLuminA differentiable optical-design system rediscovered established STED beam shaping and optical-vortex approaches, then identified a previously unreported experimental blueprint combining principles from both to produce sub-diffraction imaging in the study’s computational model. This is a specific theoretical/optical-design result, not a report of an independently built and experimentally validated microscope implementing the new blueprint.
Technology: XLuminA auto-differentiating optical experiment-design framework; transformer-based code-generation model used in the discovery workflow
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: The XLuminA differentiable optical-design system rediscovered established STED beam shaping and optical-vortex approaches, then identified a previously unreported experimental blueprint combining principles from both to produce sub-diffraction imaging in the study’s computational model.
- Still unknown: The novel design is established through computational optical modeling and simulation; the paper does not report physical construction or experimental validation of this specific blueprint. The three rediscovered designs are prior work and are not counted as new. Interpret as a computationally proposed optical design, not a demonstrated imaging instrument or biological discovery.
- AI role: AI-guided computational search generated and evaluated candidate optical experimental designs; numerical simulations identified a previously unreported combination of optical principles with predicted sub-diffraction capability. Researchers analyzed and interpreted the design.
Evidence sources: https://www.nature.com/articles/s41467-024-54696-y
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: at percentile 89 of works of the same type, year and subfield (OpenAlex work metrics, 15 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 2/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
The PlantRNA-FM model, pretrained on sequence and predicted RNA structure from 1,124 plant transcriptomes, identified 5′-UTR motifs associated with translation efficiency across plant transcriptomes. Reporter assays in plants tested selected predictions: mutating a high-translation-associated RNA secondary-structure motif reduced measured translation efficiency by up to 5.3-fold, while disrupting a model-prioritized low-translation-associated RNA G-quadruplex motif increased it by up to 5.8-fold. The paper reports the rG4 motifs as translation-repressive, consistent with prior evidence on individual rG4s; the transcriptome-wide model prioritization and tested instances are the scoped findings here.
Technology: PlantRNA-FM, an interpretable RNA foundation model integrating RNA sequence and predicted structure
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Reporter assays in plants tested selected predictions: mutating a high-translation-associated RNA secondary-structure motif reduced measured translation efficiency by up to 5.3-fold, while disrupting a model-prioritized low-translation-associated RNA G-quadruplex motif increased it by up to 5.8-fold.
- Still unknown: Experimental reporter validation was limited to selected motifs and had n=3, as reported in the figure caption. The model’s transcriptome-wide associations are not individually causal validations. The repressive role of rG4 motifs was previously known for individual cases; the paper’s contribution includes model-prioritized plant 5′-UTR motifs and experimental effects for selected constructs, not first discovery of rG4 biology. Reporter assays do not establish whole-plant phenotype or crop benefit.
- AI role: PlantRNA-FM was pretrained with plant RNA sequence and structure data, then fine-tuned on translation-efficiency data. An attention-contrast interpretation framework prioritized 5′-UTR sequence/structure motifs associated with translation; researchers tested selected motif predictions by mutating reporter constructs and measuring translation efficiency in plants.
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC11652376/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 10% of works of the same type, year and subfield (OpenAlex work metrics, 65 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A random-forest classifier trained on global basalt trace-element data distinguished destructive from constructive plate-boundary signatures, with the paper reporting greater than 98% accuracy on its labeled evaluation set. Applied to basalts from Tibet’s Bangong–Nujiang suture zone, it found mixed signatures before about 108–107 Ma and only destructive-boundary signatures afterward; the authors interpret the transition as likely oceanic-slab detachment, consistent with other geological evidence.
Technology: Random-forest classifier trained on global basalt trace-element concentrations
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: A random-forest classifier trained on global basalt trace-element data distinguished destructive from constructive plate-boundary signatures, with the paper reporting greater than 98% accuracy on its labeled evaluation set.
- Still unknown: The specific age and slab-detachment interpretation are geological inferences from classified basalt geochemistry, not direct observation of a detached slab. The authors describe the detachment explanation as likely and note consistency with existing evidence; model accuracy is not independent confirmation of the tectonic interpretation.
- AI role: The classifier learned trace-element signatures associated with destructive and constructive plate boundaries and applied them to Tibetan suture-zone basalts, revealing a time-ordered change in inferred tectonic setting that supported the authors’ slab-detachment interpretation.
Evidence sources: https://agupubs.onlinelibrary.wiley.com/doi/abs/10.1029/2024JH000400
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: at percentile 46 of works of the same type, year and subfield (OpenAlex work metrics, 2 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 2/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A Swin-Transformer search of VPHAS+ Hα images identified faint, resolved planetary-nebula candidates in crowded Galactic fields that prior point-source searches missed. Follow-up spectroscopy and HASH classification of 31 targets found 16 true, two likely and four possible planetary nebulae, alongside contaminants and a supernova-remnant fragment. This is one validated search-sample result, not 22 equally secure confirmations.
Technology: Swin-Transformer deep-learning classifier trained on known planetary nebulae
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: A Swin-Transformer search of VPHAS+ Hα images identified faint, resolved planetary-nebula candidates in crowded Galactic fields that prior point-source searches missed.
- Still unknown: The 31 follow-up targets were selected from more than 800 high-quality candidates, so the sample is not a completeness or population estimate. Spectroscopic classifications comprise 16 true, two likely and four possible PNe; these confidence levels are not merged into a claim that all 22 are confirmed. Two non-emitter late-type contaminants and other emission-line contaminants were found. The result counts the paper-level AI-assisted search and validation campaign once.
- AI role: Scanned survey images and ranked candidate planetary nebulae; follow-up spectroscopy and standard HASH classification established which selected targets were nebulae and which were contaminants.
Evidence sources: https://www.aanda.org/articles/aa/full_html/2024/12/aa50998-24/aa50998-24.html
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: at percentile 55 of works of the same type, year and subfield (OpenAlex work metrics, 5 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 2/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A 0.25° model that outperformed ECMWF's ENS ensemble on 97.2% of 1,320 variable-and-lead-time combinations (99.8% beyond 36 hours) and produces a 15-day ensemble in about 8 minutes on a TPU. Physical models remain essential because they provide training data and initial conditions.
Technology: GenCast (diffusion model)
Family GenCast: profile · models page
Publication: Publication status unspecified in annotation · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Not reported
- Test type: retrospective or benchmark
- Tested: Forecasts were compared with ECMWF ENS on held-out 2019 data (trained up to 2018) across 1,320 variable and lead-time combinations; GenCast was more accurate on 97.2% of targets.
- Still unknown: Forecasts are especially uncertain beyond a few days, the wind-power analysis is only a proof of principle, and physics-based models remain needed for training data and initial conditions.
- AI role: A diffusion model generates an ensemble of 50 or more 15-day weather trajectories from the current weather state.
Evidence sources: https://deepmind.google/blog/gencast-predicts-weather-and-the-risks-of-extreme-conditions-with-sota-accuracy/
The legacy label “Not reported” does not distinguish information absent from reviewed material from information not assessed. It is not a scientific finding of absence.
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: blog. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 1072 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 4/5 · Validation 2/5 · Novelty 4/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Spatial N-glycan imaging of hepatocellular carcinoma (HCC) tissue and paired serum identified thirteen altered glycan structures; glycoproteomics linked them to more than 50 proteins, and sixteen glycoproteins were assessed in a separate patient set. Machine-learning algorithms combining glycan/glycoprotein features with clinical variables differentiated early-stage HCC from cirrhosis in a case-control cohort of 100 HCC and 101 cirrhosis-only patients, with reported AUROC values across the algorithms of 0.88–0.97.
Technology: Machine-learning algorithms using serum N-glycan and glycoprotein features, age, sex and AFP
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: Machine-learning algorithms combining glycan/glycoprotein features with clinical variables differentiated early-stage HCC from cirrhosis in a case-control cohort of 100 HCC and 101 cirrhosis-only patients, with reported AUROC values across the algorithms of 0.88–0.97.
- Still unknown: The discovery samples included 53 HCC tissues and 23 matched sera; the 201-patient testing set was a case-control cohort, not prospective surveillance. The paper notes that sialic acid was removed before analysis, excluding an important class of glycan modification. It does not establish screening benefit or prospective clinical utility. Individual glycoforms were weak classifiers; performance came from combined feature algorithms.
- AI role: Researchers used spatial glycan imaging and glycoproteomics to select altered tissue and matched-serum features, then used machine-learning algorithms to combine selected glycan/glycoprotein measurements with clinical variables for HCC-versus-cirrhosis classification. The source reports model performance in an independent case-control patient set.
Evidence sources: https://www.nature.com/articles/s43856-024-00677-7.pdf
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: at percentile 86 of works of the same type, year and subfield (OpenAlex work metrics, 17 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A representation-learning pipeline searched archived Chandra event data and identified XRT 200515, a previously unnoticed extragalactic fast X-ray transient in the Large Magellanic Cloud. Its light curve has a sharp sub-10-second hard burst followed by an approximately 800-second oscillating tail with spectral softening. The paper discusses possible origins including a low-energy giant magnetar flare or a Type I burst from a faint low-mass X-ray binary; the source's physical nature is unresolved.
Technology: Unsupervised representation-learning pipeline using sparse autoencoders, dimensionality reduction and density-based clustering
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: A representation-learning pipeline searched archived Chandra event data and identified XRT 200515, a previously unnoticed extragalactic fast X-ray transient in the Large Magellanic Cloud.
- Still unknown: The transient is a real source-reported observational event, but its astrophysical progenitor remains unresolved. The study also catalogs candidate flares and dips; those catalog candidates are not counted as separately confirmed discoveries.
- AI role: Converted Chandra photon event files into time-energy representations, learned low-dimensional embeddings and clustered similar events; anomalous transient-dominated clusters surfaced XRT 200515 for subsequent human inspection and astrophysical analysis.
Evidence sources: https://arxiv.org/abs/2412.01150
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: at percentile 86 of works of the same type, year and subfield (OpenAlex work metrics, 20 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 2/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
ApexGO uses Bayesian optimization with a peptide variational autoencoder and an activity-prediction oracle to optimize derivatives of ten peptide templates. The primary article reports 100 synthesized sequences tested in vitro and activity improvements; selected candidates showed activity in mouse A. baumannii infection models. These results are preclinical and do not establish human efficacy.
Technology: ApexGO; peptide variational autoencoder; Bayesian optimization; APEX predictor
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: The primary source reports the study-specific result summarized here.
- Still unknown: Independent replication and broad deployment are not established by this review.
- AI role: A predictor scores generated peptide derivatives and Bayesian optimization uses feedback to propose variants; researchers synthesize and test them in vitro and in mouse models.
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC13201158/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 16 Nov 2026. Review scope: Primary-source indexed article or abstract and bibliographic metadata reviewed; complete source scope is given in source_checks..
Editorial assessment: Impact 4/5 · Validation 3/5 · Novelty 4/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A backpropagation neural network fitted composition/activity and deactivation data and suggested a Ru:Ti:Mn ratio of 0.26:0.26:0.48. The synthesized electrode had measured composition near 0.24:0.28:0.48 and reported 165.2 mV overpotential at 10 mA cm−2 in the acidic assay. Model optimum and measured composition are distinct; reported long-duration tests apply only under the study’s specific conditions.
Technology: Backpropagation neural network
Publication: Preprint · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Preprint
- Test type: direct test
- Tested: Research Square v1 reports measured Ru/Ti/Mn electrode and acidic OER result.
- Still unknown: Exact prior composition overlap and whether candidate was included in training data remain open.
- AI role: Neural network fit experimental composition/activity/durability data to suggest a candidate composition, which was synthesized and tested.
Evidence sources: https://doi.org/10.21203/rs.3.rs-5208867/v1
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 16 Nov 2026. Review scope: Primary source checked for identity, publication date, AI workflow and the bounded experimental result stated..
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Regularized logistic-regression and other machine-learning analyses of 178 Olink proteins identified a 29-protein preclinical Crohn disease signature. It distinguished future cases from controls with AUC 0.85 in a Swedish discovery cohort and 0.87 in an independent preclinical cohort; a separate inception cohort produced AUC 1.0. The study also reported a less predictive ulcerative-colitis signature and found that downregulated proteins related to gut-barrier integrity and macrophage function correlated with time to Crohn diagnosis.
Technology: Regularized logistic regression and additional machine-learning methods (L1-penalized regression, elastic net, support-vector machines and random forests) applied to Olink protein measurements
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: It distinguished future cases from controls with AUC 0.85 in a Swedish discovery cohort and 0.87 in an independent preclinical cohort; a separate inception cohort produced AUC 1.0.
- Still unknown: This is a nested case-control study using biobanked samples, not prospective screening or intervention evidence. The paper reports that the Crohn signature’s AUC fell to 0.58 when preclinical twins were compared with their healthy twin siblings, suggesting inherited/shared-environment factors affect discrimination. The ulcerative-colitis validation AUC was lower (0.67). Protein associations do not establish causal mechanisms; authors call for mechanistic studies. Source checked; independent replication beyond reported cohorts was not assessed.
- AI role: Machine-learning feature selection/modeling produced preclinical Crohn and ulcerative-colitis protein signatures in the discovery cohort. The fitted signature was evaluated in an external population-based preclinical cohort, an inception cohort at diagnosis, and twin comparisons. The result is predictive association from cohort data; the paper says mechanistic studies are needed.
Evidence sources: https://www.gastrojournal.org/article/S0016-5085%2824%2905741-X/pdf
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 10% of works of the same type, year and subfield (OpenAlex work metrics, 32 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A neural-network model was used to predict metal-oxide catalyst performance for acetylene semi-hydrogenation. The authors report experimental validation of predictions for CuO, ZnO and V2O5, and report 99.6% conversion with 90% ethylene selectivity for CuO under the abstract-described conditions. The entry is limited to the publisher abstract; detailed model validation, candidate selection and full assay context were not reviewed.
Technology: Neural-network prediction of metal-oxide catalyst performance
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: The publisher abstract says CuO, ZnO and V2O5 predictions were experimentally validated; it reports 99.6% conversion and 90% ethylene selectivity for CuO at 50–100°C.
- Still unknown: Detailed selection chronology, model split/performance, full test conditions and results for ZnO and V2O5 were not established from the reviewed abstract. The date is the publisher-displayed date, not a claim of first public appearance.
- AI role: A neural-network model predicted catalyst performance and informed the metal-oxide candidates reported as experimentally validated.
Evidence sources: https://www.sciencedirect.com/science/article/pii/S0926860X24004149 · https://homepage.zjut.edu.cn/ljh_en/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review date: 17 Nov 2026. Review scope: Primary publisher abstract for identity, displayed date, model and reported experiment; official Zhejiang University of Technology faculty profile/publication list for named contributor affiliation. Full paper methods and supplement not reviewed..
Editorial assessment: Impact 2/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
The November 2024 Nature Communications paper describes six GPT-4 agents guiding copper/TEMPO alcohol oxidation, from literature search through optimization and purification. An agent-assisted scale-up produced 915 mg of product at 86% isolated yield and above 98% purity. Additional reaction tasks were demonstrated. The workflow involved chemists: thin-layer chromatography was performed manually, so this was not fully autonomous synthesis.
Technology: LLM-RDF (six GPT-4 agents)
Family GPT-4: profile · models page
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Physical synthesis, optimization, scale-up and product purification.
- Still unknown: Not assessed: supplements; manual TLC and chemist participation are explicit in the paper.
- AI role: Six GPT-4 agents supported literature, design, execution, analysis, separation and interpretation.
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC11585555/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 189 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 4/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
ClickGen combines reaction-based molecular generation, reinforcement learning and inpainting to generate synthesizable candidate structures. In a PARP1 case study, researchers selected and synthesized three generated lead molecules; two showed nanomolar PARP1 inhibition in vitro, with reported IC50 values of 19.24 ± 1.63 nM and 69.73 ± 14.57 nM. The two compounds also showed antiproliferative activity in several tested cancer cell lines and lower toxicity in the reported normal-cell assays than the most toxic lead compounds.
Technology: ClickGen deep-learning reaction-based molecular generator with reinforcement learning, Monte Carlo tree search and an inpainting generator
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: In a PARP1 case study, researchers selected and synthesized three generated lead molecules; two showed nanomolar PARP1 inhibition in vitro, with reported IC50 values of 19.24 ± 1.63 nM and 69.73 ± 14.57 nM.
- Still unknown: Three selected molecules were synthesized and tested in vitro; the article reports two with nanomolar PARP1 inhibition. Cell-line antiproliferative and toxicity assays do not show tumor efficacy or safety in animals or people. The reported IC50 values are from triplicate measurements in the study. This record captures the PARP1 experimental result, not the paper’s model benchmark comparisons.
- AI role: The system generated candidate molecules by assembling reaction-compatible synthons and using reinforcement learning/inpainting to guide novelty and binding-oriented generation. Researchers computationally filtered and selected compounds, synthesized three leads and tested them; PARP1 inhibition and cell effects were experimentally measured, not generated by the model.
Evidence sources: https://www.nature.com/articles/s41467-024-54456-y.pdf
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 10% of works of the same type, year and subfield (OpenAlex work metrics, 45 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 2/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A model trained on 56 experimental observations selected electrode-fabrication and electrolysis conditions for silver-catalyzed CO₂ reduction. Researchers experimentally validated the proposed conditions against a desired hydrogen-to-carbon-monoxide ratio and CO efficiency. This is an assay-specific process-optimization result for a known silver catalyst, not a newly discovered catalyst or evidence of industrial carbon-neutral operation.
Technology: Random-forest regression; kernel ridge regression
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: The paper reports 56 experimental observations and experimental validation of selected conditions for the target product ratio and CO faradaic efficiency.
- Still unknown: Independent replication and performance outside stated assay conditions are not established.
- AI role: ML models trained on experimental outcomes estimate selectivity across fabrication and electrolysis parameters; inverse search proposes conditions for follow-up testing.
Evidence sources: https://d-nb.info/1349018996/34
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review date: 16 Nov 2026. Review scope: Primary author manuscript PDF read through abstract, results and conditions; exact online date from source ledger, issue metadata says 2024..
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A Transformer-based decoder trained on data from a 49-qubit Sycamore processor made 6% fewer errors than tensor-network methods and 30% fewer than correlated matching in the largest experiments. It is too slow for real-time error correction on fast superconducting processors.
Technology: AlphaQubit (Transformer)
Family AlphaQubit: profile · models page
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: On Sycamore data, logical error per round was 2.901e-2 (distance 3) and 2.748e-2 (distance 5), versus 3.028e-2 and 2.915e-2 for a tensor-network decoder. Simulations to distance 11 (241 qubits); correlated matching was 1.25 to 1.4 times worse.
- Still unknown: Throughput is slower than the 1 microsecond per round target. Training at distance 11 needs more data. Accuracy beyond distance 11 and logical computations remain future work. Larger-distance results are simulated.
- AI role: A recurrent transformer neural network decodes surface-code syndromes to predict logical errors; it is the method under test.
Evidence sources: https://doi.org/10.1038/s41586-024-08148-8
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: blog. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 195 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 2/5 · Novelty 4/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
SCimilarity searched a 23.4-million-cell reference atlas spanning 412 single-cell RNA-sequencing studies using query profiles from interstitial-lung-disease macrophages and fibroblasts. It identified fibrosis-associated macrophage-like states in other fibrotic diseases, COVID-19 and selected tumour datasets, and a 3D hydrogel culture as a high-ranking in-vitro match; the authors report experimentally demonstrating that this system reproduces the queried macrophage state.
Technology: SCimilarity, a deep metric-learning model for querying cell-state similarity across single-cell atlases
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: It identified fibrosis-associated macrophage-like states in other fibrotic diseases, COVID-19 and selected tumour datasets, and a 3D hydrogel culture as a high-ranking in-vitro match; the authors report experimentally demonstrating that this system reproduces the queried macrophage state.
- Still unknown: Most cross-disease and tumour-context links are computational similarity observations in previously published single-cell datasets; they do not establish causal roles in those diseases. Some rare tissue/cancer matches are from individual datasets or samples. The in-vitro 3D hydrogel assay is the experimentally tested part; broad disease associations need further biological validation.
- AI role: The model ranked cell profiles in a 23.4-million-cell atlas by similarity to ILD-derived macrophage/fibroblast query profiles, linking these states to disease/tissue contexts and nominating the 3D hydrogel culture for experimental follow-up.
Evidence sources: https://www.nature.com/articles/s41586-024-08411-y
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 208 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A seven-layer convolutional neural network classified raw seismic waveforms recorded near Italy’s 2016 Mw 6.5 Norcia earthquake as pre- or post-mainshock and separated time-to-failure intervals with over 90% test accuracy. The authors interpret the model’s performance, frequency dependence and null tests as evidence that local waveforms encode evolving attenuation and elastic properties consistent with progressive fault damage before rupture and postseismic recovery.
Technology: Seven-layer convolutional neural network trained on local seismic waveforms
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: A seven-layer convolutional neural network classified raw seismic waveforms recorded near Italy’s 2016 Mw 6.5 Norcia earthquake as pre- or post-mainshock and separated time-to-failure intervals with over 90% test accuracy.
- Still unknown: Authors studied one earthquake sequence and state that generalization to other tectonic settings and prospective real-time precursor detection remain untested. Related changes in elastic properties were already reported by field studies; the paper's new contribution is applying deep learning to infer them from this local waveform sequence.
- AI role: Classified the waveform signals by their temporal relation to the mainshock; controlled tests using incorrect mainshock times, distant stations and filtered frequencies helped expose subtle local seismic-wave property changes associated with the fault cycle.
Evidence sources: https://www.nature.com/articles/s41467-024-54153-w
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: at percentile 78 of works of the same type, year and subfield (OpenAlex work metrics, 11 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 2/5 · Validation 2/5 · Novelty 2/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A dataset from Meta FAIR, the University of Toronto, and VSParticle covers over 600 mixed-metal catalysts made from 13 elements and tested under the same conditions, alongside analysis of over 19,000 materials. Hydrogen-evolution predictions flagged platinum-containing alloys as promising, although platinum was absent from training data. Fewer than 25% of produced catalysts matched the target; this was a pilot.
Technology: Catalyst prediction models (Open Catalyst)
Publication: Preprint · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Preprint
- Test type: direct test
- Tested: Computational predictions were compared with lab results: 441 gas diffusion electrodes tested in zero-gap electrolysis and 572 characterized samples, alongside DFT-verified adsorption energies on about 20,000 materials.
- Still unknown: Platinum was ranked as a top hydrogen-evolution candidate but was not experimentally tested, and the abstract lists no explicit limitations.
- AI role: AI-accelerated relaxations (685 million) fed a data-driven Sabatier volcano that flagged platinum-containing alloys as promising.
Evidence sources: https://arxiv.org/abs/2411.11783
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: blog. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
The CHOOSER framework fine-tuned ESM-2 protein-language-model representations for alignment-free CRISPR-Cas discovery and pre-crRNA self-processing predictions. It nominated 11 Casλ homologs, nearly doubling the known Casλ set. Wet-lab work on selected homologs confirmed pre-crRNA processing for five of six tested proteins; EphcCasλ also cleaved target DNA and showed trans-cleavage activity, with a TTR PAM reported.
Technology: CHOOSER, a protein foundation-model framework fine-tuning ESM-2 representations for Cas homolog identification and Cas12 pre-crRNA-processing classification
Family ESM / ESMFold: profile · models page
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Wet-lab work on selected homologs confirmed pre-crRNA processing for five of six tested proteins; EphcCasλ also cleaved target DNA and showed trans-cleavage activity, with a TTR PAM reported.
- Still unknown: This is a specific functional genomics result, not evidence of a ready-to-use diagnostic or gene-editing tool. Only selected candidates were purified and tested; the paper reports pre-crRNA activity in five of six tested Casλ proteins and DNA-cleavage validation for a subset. EphcCasλ activity and PAM characterization are in-vitro findings. The article has a publisher-listed author correction dated 2025-01-09; this record relies on the version-of-record text and retains limited scope.
- AI role: CHOOSER mined candidate proteins adjacent to CRISPR arrays using learned sequence representations rather than alignment-only detection and prioritized Cas12 homologs by predicted pre-crRNA processing. The study reported 11 Casλ homologs and experimentally tested selected predictions; purified EphcCasλ was validated for self-processing pre-crRNA, DNA cleavage and trans-cleavage.
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC11576732/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 10% of works of the same type, year and subfield (OpenAlex work metrics, 26 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
LLM agents built from participants' full interview transcripts reproduced their General Social Survey answers at about 85% of the rate at which participants reproduced their own answers two weeks later, a normalized score. They also matched personality and economic-game results (normalized correlations of about 0.80 and 0.66) and replicated four of five behavioural experiments. Caveat: arXiv preprint, not peer reviewed, and the results are normalized rather than raw accuracy.
Technology: Generative agent architecture with LLM and full interview transcripts
Publication: Preprint · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Preprint
- Test type: retrospective or benchmark
- Tested: Agents built from interview and survey data of 1,052 Americans were scored on held-out General Social Survey items against participants' own two-week test-retest consistency, and on personality, economic-game and experiment replications.
- Still unknown: Preprint, not peer reviewed; there is no dedicated limitations section, and combining interviews and surveys gave only modest gains over either source.
- AI role: An LLM was the engine of agents that answer questions on behalf of individuals, grounded in their self-report data.
Evidence sources: https://arxiv.org/abs/2411.10109
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review scope and date are not recorded; source checked does not establish full-text review.
Editorial assessment: Impact 3/5 · Validation 2/5 · Novelty 4/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Published in Science, Evo (Arc Institute, Stanford, and UC Berkeley) designed EvoCas9-1, a CRISPR system sharing only about 73% of its sequence with CRISPR-Cas9 but with comparable activity after eleven designs, and new IS200/IS605 transposons that cut and paste DNA. The team calls this the beginning and conducted an ethics and safety study.
Technology: Evo (DNA foundation model)
Family Evo: profile · models page
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Evo was prompted to generate CRISPR systems; EvoCas9-1 was obtained after testing eleven designs, validated in the lab by an experimental biology subteam.
- Still unknown: The page gives no success rates or error analysis, and an ethics and safety investigation addressed potential risks rather than efficacy.
- AI role: Evo, a DNA-trained foundation model, designed the protein and RNA components of EvoCas9-1 and new IS200/IS605 transposons.
Evidence sources: https://arcinstitute.org/news/evo-science
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: press release. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 737 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 4/5 · Validation 3/5 · Novelty 4/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Ro5's prospective HydraScreen structure-based deep-learning screen ranked compounds for IRAK1 and was evaluated against a 47,000-compound robotic high-throughput screen. The study reports that 23.8% of IRAK1 hits were in HydraScreen's top 1% and identifies three nanomolar scaffolds, two described in the abstract as novel IRAK1 candidates.
Technology: HydraScreen structure-based deep-learning virtual screening, with SpectraView knowledge-graph target evaluation
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: The study reports that 23.8% of IRAK1 hits were in HydraScreen's top 1% and identifies three nanomolar scaffolds, two described in the abstract as novel IRAK1 candidates. The new candidate experiments were conducted within the Ro5/Strateos discovery collaboration.
- Still unknown: This is prospective early-stage hit identification, not drug development or disease efficacy. The assay validated kinase inhibition in vitro. Authors are affiliated with the companies that developed the tools. The abstract describes two scaffolds as novel candidates, while detailed similarity analysis indicates one of three nanomolar scaffolds had a similar known active; do not call all hits unprecedented.
- AI role: SpectraView selected IRAK1 as the study target; HydraScreen scored docked ligand-pose ensembles to rank compounds before prospective screening. Researchers compared rankings with automated high-throughput assay results from a 47,000-compound library and reported nanomolar IRAK1 hit scaffolds, two described as novel candidates.
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC11566907/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: at percentile 77 of works of the same type, year and subfield (OpenAlex work metrics, 9 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 2/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A random-forest model trained on first-principles formation and decomposition energies screened more than 150,000 proposed gamma-prime compositions. Domain screening and CALPHAD narrowed the set; researchers synthesized two selected Co-based gamma/gamma-prime alloy systems and characterized their microstructures. The measured compositions differ from nominal targets, and preliminary hardness values were below the cited conventional comparator. Engine performance was not demonstrated. Count the study once.
Technology: Random-forest regression; density-functional-theory energy screening
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Researchers synthesized and characterized two screened Co-based alloy systems; XRD and electron microscopy support reported microstructures and preliminary hardness was measured.
- Still unknown: Nominal and measured compositions differ; preliminary hardness was below the cited conventional comparator; no engine performance or independent replication was reported.
- AI role: A random-forest model trained on DFT energies screened candidate gamma-prime compositions; domain experts and CALPHAD then narrowed candidates for synthesis.
Evidence sources: https://www.nature.com/articles/s41524-024-01455-8
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review date: 16 Nov 2026. Review scope: Primary npj Computational Materials article page, abstract, indexed results and author metadata..
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A machine-learning-assisted reaction-optimization study found conditions for reducing N,N-dimethylacetamide to dimethylethylamine using the known platinum complex [Pt(DMS)₂Cl₂]. The selected condition used 0.01 mol% catalyst and reached complete conversion in two hours at 39 °C without solvent, with a reported turnover frequency of 5,002 h⁻¹. The chemical catalyst identity is not claimed as new.
Technology: Sunthetics machine-learning reaction-optimization platform
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: The selected condition used 0.01 mol% catalyst and reached complete conversion in two hours at 39 °C without solvent, with a reported turnover frequency of 5,002 h⁻¹.
- Still unknown: This is a single-substrate reaction-condition optimization on N,N-dimethylacetamide. [Pt(DMS)₂Cl₂] is a known complex; the reported finding is the ML-assisted optimized condition and measured performance, not discovery of a new platinum compound. The best-performing condition is reported after human-guided interpretation as well as model recommendations. The paper does not establish transferability to other amides, industrial scale-up or general performance across catalyst classes.
- AI role: The ML platform recommended reaction conditions across iterative experiments, using observed reaction outcomes to refine proposed conditions and reduce the number of trials. The team combined those recommendations with chemist interpretation and tested the predicted conditions experimentally; the source reports the optimized platinum-catalyzed reduction as a study-specific result.
Evidence sources: https://pubs.rsc.org/en/content/articlehtml/2024/cc/d4cc05273k
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: at percentile 59 of works of the same type, year and subfield (OpenAlex work metrics, 7 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 2/5 · Validation 3/5 · Novelty 2/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
The new model, trained on 15,980 stations (up from 5,680), gives seven-day reliability comparable to the previous model's five-day forecast. Flood Hub covers over 100 countries and about 700 million people (up from 460 million). About 250,000 virtual points are unverified and intended for experts only; flash, urban, and coastal floods remain out of scope. The linked March 2024 Nature paper describes the predecessor model, also covered by the catalog's March entry; the November performance and coverage update is reported in this Google Research announcement.
Technology: Hydrological forecasting model
Publication: Announcement only · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Announcement only
- Test type: retrospective or benchmark
- Tested: Gauged basins were checked with standard cross-validation, and ungauged locations were checked against Sentinel-1 radar inundation images; reliability was compared using Nash-Sutcliffe efficiency distributions across 3,079 basins.
- Still unknown: Ungauged validation covers only a fraction of locations, flash, urban and coastal floods are out of scope, and virtual-gauge forecasts are unverified.
- AI role: An LSTM network takes meteorological time series and static catchment attributes, adds input from DeepMind's medium-range forecast model, and outputs a probabilistic streamflow distribution.
Evidence sources: https://research.google/blog/a-flood-forecasting-ai-model-trained-and-evaluated-globally/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: blog. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 580 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 4/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
The GRACE workflow combined RFdiffusion structure generation, ProteinMPNN sequence design, CLEAN enzyme classification and solubility/structure filtering to select two carbonic-anhydrase-like sequences from 10,000 candidates. The authors report experimental expression and carbonic anhydrase activity: dCA12_2 at 400 WAU/mL and dCA23_1 at 260 WAU/mL in the reported assay.
Technology: GRACE (RFdiffusion, ProteinMPNN and CLEAN, with computational solubility/structural filters)
Family RFdiffusion: profile · models page
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: The authors report experimental expression and carbonic anhydrase activity: dCA12_2 at 400 WAU/mL and dCA23_1 at 260 WAU/mL in the reported assay.
- Still unknown: The reported activity values are from the study's carbonic-anhydrase assay and do not establish purified-enzyme kinetics, physiological CO2-fixation performance, or industrial utility. The designs and validation are a two-protein proof of concept; no independent replication assessed.
- AI role: The generative-design workflow produced structures/sequences and machine-learning function classification prioritized two candidates from a 10,000-sequence pool; researchers expressed both and measured carbonic-anhydrase activity.
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC11669173/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: at percentile 89 of works of the same type, year and subfield (OpenAlex work metrics, 20 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 2/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Mobile robots move samples between a synthesis platform, a UPLC-MS spectrometer, and a benchtop NMR, while a heuristic decision-maker evaluates data and chooses reactions. Demonstrated on structural diversification, supramolecular chemistry, and photochemical synthesis. The AI is heuristic and there is no speed or accuracy metric; the announcement says robots will not have a ‘Eureka’ moment.
Technology: Heuristic decision-making with mobile robots
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Mobile robots were demonstrated on three exploratory chemistry problems (structural diversification, supramolecular chemistry, photochemical synthesis); the page reports no reaction counts or accuracy metrics, only that the AI made the same or similar decisions as a human researcher.
- Still unknown: The page reports no quantitative accuracy metrics, only two robots were used, and the system lacks a 'Eureka' moment.
- AI role: A heuristic decision-maker processes analytical data and autonomously decides whether a reaction proceeds to the next step.
Evidence sources: https://news.liverpool.ac.uk/?p=123346
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: press release. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 307 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 4/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A graph neural network augmented with quantum-inspired adsorption-site similarity reduced the DFT training data needed for a given prediction accuracy from 1,600 to 800 calculations. The model identified a five-metal Fe–Co–Ni–Ir–Ru high-entropy alloy candidate for oxygen reduction; the author group reports experimental validation and performance above a Pt/C comparison. The publisher abstract gives a predicted overpotential of 0.24 V, which is not presented here as the measured value. This is a laboratory catalyst result, not a fuel-cell deployment claim.
Technology: Similarity-augmented graph neural network for adsorption-site and alloy-composition screening
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: The author group reports experimental validation identifying a five-metal high-entropy alloy oxygen-reduction catalyst that outperformed its Pt/C comparison. The publisher describes experimental study of four additional HEAs; the 0.24 V overpotential is a model prediction, not stated here as an experimental measurement.
- Still unknown: The publisher full Methods and supplementary results were not reviewed in this pass. The abstract-level material does not provide the measured ORR performance values, test uncertainty, durability, or full experimental protocol. No independent replication or fuel-cell device performance is established.
- AI role: The graph neural network incorporated quantum-inspired similarity between adsorption sites, reducing the required DFT training-set size from 1,600 to 800 while maintaining prediction accuracy. Its predictions nominated HEA compositions for experimental synthesis and ORR evaluation.
Evidence sources: https://www.sciencedirect.com/science/article/pii/S2590238524005289 · https://clean.energyscience.ca/publications/078.html
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: other. Paper: linked separately.
Review date: 17 Nov 2026. Review scope: Primary ScienceDirect indexed abstract and metadata; primary author-group publication page. Direct publisher full-text open was blocked, so no full Methods or complete article affiliation review is claimed..
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
XRDMatch, a semi-supervised model using X-ray diffraction patterns as its descriptor, screened unlabeled crystal data and identified 38 lithium superionic-conductor predictions. The paper says 32 were supported by prior literature, six were new candidates assessed with ab initio molecular dynamics, and Li₆AsSe₅I was synthesized and experimentally confirmed as a superionic conductor. The source abstract does not establish that all 38 are new or experimentally validated; this record covers the single reported synthesized material.
Technology: XRDMatch semi-supervised learning framework
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: XRDMatch, a semi-supervised model using X-ray diffraction patterns as its descriptor, screened unlabeled crystal data and identified 38 lithium superionic-conductor predictions.
- Still unknown: The publisher abstract says six newly predicted candidates were quantified with ab initio molecular dynamics, but names only Li₆AsSe₅I as synthesized and experimentally confirmed. It reports 32 additional predictions matched recent literature. The full publisher article was inaccessible during the source review (HTTP 403); the annotation is limited to the primary publisher abstract and its article metadata.
- AI role: Consistency regularization and pseudo-labeling let the model learn from limited labeled conductivity data and unlabeled ICSD X-ray diffraction patterns. It nominated Li₆AsSe₅I for synthesis, where the researchers experimentally confirmed superionic conductivity.
Evidence sources: https://pubs.rsc.org/en/content/articlelanding/2024/ee/d4ee02970d
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: at percentile 74 of works of the same type, year and subfield (OpenAlex work metrics, 13 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 2/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
HiDDEN refines sample-level case labels at single-cell resolution. In human bone-marrow single-cell RNA-seq data, its labels tracked manual neoplastic-cell annotations in mixed precursor samples and inferred malignant-like cells in three low-purity MGUS samples that the original manual analysis had labeled as containing only healthy cells. The inferred cells were enriched for previously described malignancy-associated gene signatures. In a separate mouse demyelination dataset, HiDDEN identified an early endothelial subpopulation with expression changes associated with angiogenesis, blood–brain-barrier breakdown and extracellular-matrix responses.
Technology: HiDDEN, a machine-learning framework for refining disease/condition labels at single-cell resolution
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: HiDDEN refines sample-level case labels at single-cell resolution.
- Still unknown: The MGUS finding is a transcriptomic inference in previously collected samples, computationally supported by known gene signatures and sample-purity estimates; the paper does not report independent pathology confirmation, clinical follow-up, or validation as a diagnostic test. The endothelial subpopulation was found in a mouse lesion model; its association with blood–brain-barrier dysfunction does not establish causality.
- AI role: The model inferred which individual cells in case samples were affected rather than assigning the sample-level case label to all cells. This enabled the paper’s reanalysis of human myeloma precursor single-cell data and mouse demyelination single-nucleus data, yielding the specified malignancy and endothelial-population findings.
Evidence sources: https://www.nature.com/articles/s41467-024-53666-8.pdf
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 10% of works of the same type, year and subfield (OpenAlex work metrics, 24 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 2/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
PatternBoost alternates a local search that generates candidate constructions with a transformer trained on the best ones. The authors report the best known solutions to several extremal combinatorics problems, including a counterexample to a conjecture that had been open for 30 years. Candidates are scored mechanically, so each construction is checked by computer; the abstract does not name the conjecture and the paper was not peer-reviewed at the time of posting.
Technology: PatternBoost (transformer trained on search outputs, combined with local search)
Publication: Preprint · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Preprint
- Test type: retrospective or benchmark
- Tested: The abstract reports best known solutions to several extremal combinatorics problems, including a counterexample to a 30-year-old conjecture, but describes no verification procedure.
- Still unknown: The abstract states that PatternBoost's performance varies across problems.
- AI role: A local search generated candidate constructions, and a transformer trained on the best candidates seeded the next search round.
Evidence sources: https://arxiv.org/abs/2411.00566
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Editorial assessment: Impact 3/5 · Validation 2/5 · Novelty 4/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A support-vector-regression model trained to infer gene dependencies from RNA-sequencing profiles predicted selective MCL1 dependence in alveolar soft part sarcoma (ASPS), distinct from TFE3-translocation renal cell carcinoma. The authors tested the prediction with CRISPR knockout and the MCL1 inhibitor murizatoclax in two ASPS cell lines; the reported viability effects were greater in ASPS than comparator tRCC lines.
Technology: Transcriptome-based support-vector regression model (RBF kernel; 657 highly predictable gene dependencies)
Publication: Preprint · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Preprint
- Test type: direct test
- Tested: The authors tested the prediction with CRISPR knockout and the MCL1 inhibitor murizatoclax in two ASPS cell lines; the reported viability effects were greater in ASPS than comparator tRCC lines.
- Still unknown: bioRxiv preprint, not peer reviewed as stated by PMC. Validation was in two ASPS cell lines with CRISPR and inhibitor assays; it does not establish patient benefit or clinical efficacy. The model nominates dependencies and is not a functional screen itself.
- AI role: The model predicted per-gene dependency scores from tumor/cell-line expression profiles and prioritized selective vulnerabilities. The study then tested MCL1 dependence in ASPS using three distinct sgRNAs and murizatoclax; those experiments, rather than the prediction alone, provide the reported validation.
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC11527139/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Editorial assessment: Impact 2/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A partial-convolution U-Net trained on CMIP6 climate simulations and transferred to European station-based extreme indices reconstructed the poorly observed 1911 heatwave. The model estimates that in parts of southern France, more than 40% of September days exceeded the 90th percentile of maximum temperature, extending the heatwave signal into fall. The reconstructed pattern is similar to 20th Century Reanalysis Version 3 and aligns spatially with higher French mortality in 1911; the HadEX3 gridded record does not show the same pronounced heat.
Technology: Climate Reconstruction AI (CRAI), a U-Net with partial convolutions trained on CMIP6 historical simulations and transferred to the HadEX-CAM European extreme-index dataset
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: A partial-convolution U-Net trained on CMIP6 climate simulations and transferred to European station-based extreme indices reconstructed the poorly observed 1911 heatwave.
- Still unknown: The September 1911 values are AI reconstructions rather than newly recovered thermometer readings; the paper states the source observations are sparse or missing. CRAI is trained on CMIP6 simulations and applied to European station-based extreme-index fields, so model-prior dependence remains. The article reports a similar spatial pattern in 20CRv3 and correlation with 1911 mortality, but these support the reconstructed pattern rather than directly validate each missing grid cell. The reported >40% refers to predicted percentage of days above the local 90th percentile for maximum temperature.
- AI role: The deep-learning reconstruction inferred missing monthly temperature-extreme indices from spatial context and learned climate-model patterns. This enabled a detailed estimate for the poorly observed September 1911 period; the model output exposes hot extremes in southern France that are not clearly represented in the standard HadEX3 map and supports comparison with independent 20th Century Reanalysis and historical mortality patterns.
Evidence sources: https://www.nature.com/articles/s41467-024-53464-2
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 10% of works of the same type, year and subfield (OpenAlex work metrics, 36 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 2/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
The CODA platform used the Malinois deep-learning model and sequence-search algorithms to design synthetic cis-regulatory elements (CREs). In massively parallel reporter assays, 94.1% of the synthetic sequences met the authors’ cell-type-specificity criterion, versus 73.6% of model-selected natural sequences and 40.6% of chromatin-mark-selected natural sequences; synthetic sequences also had higher measured target/off-target separation across K562, HepG2 and SK-N-SH cells. In transgenic zebrafish, two of three liver-directed and two of three neuron-directed CREs showed the intended activity. One neuron-directed sequence, synN1, retained cortical and neuronal expression in mice, including postnatal neocortical layer 6 and subplate neurons.
Technology: CODA (Computational Optimization of DNA Activity) with Malinois deep neural network and AdaLead, simulated annealing, and Fast SeqProp sequence search/design algorithms
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: In massively parallel reporter assays, 94.1% of the synthetic sequences met the authors’ cell-type-specificity criterion, versus 73.6% of model-selected natural sequences and 40.6% of chromatin-mark-selected natural sequences; synthetic sequences also had higher measured target/off-target separation across K562, HepG2 and SK-N-SH cells.
- Still unknown: The in-vitro work used three transformed human cell lines and reporter assays. In-vivo testing covered a small selected set: two of three liver-directed and two of three neuron-directed elements were active in zebrafish; mouse validation focused on neuronal synN1, with n=3 synN1 mice for postnatal cell-type analysis and n=6 for the 5-week cortical reporter assay. These are transgenic reporter results, not therapeutic delivery or evidence of clinical utility. Earlier synthetic enhancer studies existed; this paper’s counted result is its CODA/Malinois comparison against natural sequences plus selected vertebrate validation, not the general idea of AI-designed enhancers.
- AI role: Malinois predicted MPRA activity of candidate DNA sequences in K562, HepG2 and SK-N-SH cells; CODA iteratively optimized sequences for specified cell-type specificity. Researchers synthesized and experimentally measured the designs by MPRA, then tested selected CREs with reporter constructs in zebrafish and mice. The paper attributes the novel sequences and their measured expression specificity to this model-guided design loop.
Evidence sources: https://www.nature.com/articles/s41586-024-08070-z
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 180 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
The SMDET image-level neural network identified 118 previously uncatalogued high-proper-motion ultracool-dwarf candidates in WISE time-series images. Follow-up near-infrared spectra for two selected objects yielded T5 and T3 classifications, confirming their estimated spectral types within one subtype.
Technology: SMDET, a convolutional neural network for detecting faint moving sources in WISE time-series coadds
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: The SMDET image-level neural network identified 118 previously uncatalogued high-proper-motion ultracool-dwarf candidates in WISE time-series images.
- Still unknown: Most of the 118 objects remain candidates, and the study spectroscopically checks only two. Some candidates were also identified by citizen scientists, so the AI was part of a human/algorithm discovery process. The counted finding is limited to the two spectra and the candidate search result as one paper-level report.
- AI role: Scanned WISE image sequences and prioritized faint, fast-moving source candidates; researchers cross-checked catalogs, estimated photometric types and distances, and acquired spectra for two objects to verify their T-dwarf classifications.
Evidence sources: https://iopscience.iop.org/article/10.3847/1538-3881/ad77d2
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: at percentile 65 of works of the same type, year and subfield (OpenAlex work metrics, 8 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 2/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
The study used CRISPR knockouts of 84 genes in primary human CD4+ T cells and a new linear latent causal Bayes (LLCB) network-inference method to estimate 211 regulatory connections. Integrating the inferred network with immune GWAS analyses identified an interconnected module in which KMT2A was inferred upstream of JAK-STAT and IL-2 signaling regulators including IRF4, STAT5A/B and IL2RA, linking an epigenetic regulator to immune-disease-associated pathways.
Technology: Linear latent causal Bayes (LLCB), a Bayesian extension of linear latent causal network inference, applied to CRISPR knockout and RNA-sequencing data
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: The study used CRISPR knockouts of 84 genes in primary human CD4+ T cells and a new linear latent causal Bayes (LLCB) network-inference method to estimate 211 regulatory connections.
- Still unknown: The study perturbed 84 selected genes in primary CD4+ T cells from three donors, so this is not a complete T-cell regulatory map. LLCB “direct effects” are statistical effects adjusted for other perturbed genes and are not evidence of physical binding; effects from unperturbed genes can be attributed to modeled edges. The GWAS integration identifies pathway links, not proof that the inferred cascade causes clinical disease. A prior bioRxiv version (10.1101/2023.09.17.557749) is the same study and is not counted separately.
- AI role: LLCB inferred direct and indirect regulatory relationships from perturbation-associated expression effects across 84 targeted genes. The model-derived network and gene-program analysis, combined with immune GWAS statistics, supported the specific KMT2A-to-JAK-STAT/IL-2 regulatory finding. CRISPR perturbation supplied the experimental perturbation data; the network algorithm converted those measurements into the reported regulatory map.
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC11605694/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: at percentile 90 of works of the same type, year and subfield (OpenAlex work metrics, 21 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
The eDENTIFI gradient-boosted model estimated the probability of untreated infection during the next six hours from demographics, vital signs and laboratory results. Trained on 3,357 adult hospitalizations at two Illinois health systems and validated in 1,632 patients at a third, it achieved validation AUC 0.80 versus 0.64 for SIRS criteria. At the reported threshold it detected infection a median two hours before antimicrobial administration (47.6% sensitivity; 93.6% negative predictive value).
Technology: eDENTIFI, a gradient-boosted machine model using longitudinal discrete-time predictions
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: Trained on 3,357 adult hospitalizations at two Illinois health systems and validated in 1,632 patients at a third, it achieved validation AUC 0.80 versus 0.64 for SIRS criteria.
- Still unknown: Retrospective derivation and validation used historical hospital records from 2006–2018, and the infection reference standard was clinician chart review. The reported 10.8% absolute/46.4% relative reduction in unnecessary IV antibiotics is a counterfactual model comparison, not observed intervention evidence. The authors state prospective validation is needed; no patient benefit or implementation outcome was established.
- AI role: The model used routinely available demographics, vital signs and laboratory results to predict clinician-chart-reviewed untreated infection over the next six hours. The study tested temporal detection performance in a separate third health system; model estimates were not prospectively used to direct care.
Evidence sources: https://pubmed.ncbi.nlm.nih.gov/39392375/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: at percentile 49 of works of the same type, year and subfield (OpenAlex work metrics, 1 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 2/5 · Novelty 2/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A SimSiam self-supervised representation model with t-SNE and DBSCAN clustering explored two months of continuous seismological records from near the Fani Maoré submarine volcano. Researchers identified two previously undocumented multi-day sequences in October–November 2019, each beginning with increased volcano-tectonic events and followed by tremor-like, hydroacoustic or mixed signals. The paper describes the sequences as potentially associated with ongoing eruptive activity.
Technology: SimSiam self-supervised learning with t-SNE dimensionality reduction and DBSCAN clustering
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: A SimSiam self-supervised representation model with t-SNE and DBSCAN clustering explored two months of continuous seismological records from near the Fani Maoré submarine volcano.
- Still unknown: The event data are from two months in 2019 and a nearby ocean-bottom seismometer; the study does not establish exact event locations or independently prove that each sequence was eruptive. Authors state that instrumental noise cannot be ruled out for some high-energy signals, and describe the volcanic association as potential. It is a peer-reviewed 2025 issue paper published online 2024-10-10.
- AI role: The self-supervised model constructed signal representations from full continuous seismograms; clustering surfaced two coherent, previously undocumented temporal event sequences that the researchers then interpreted using expert signal review. This avoids treating the result as a detector-only tool release.
Evidence sources: https://academic.oup.com/gji/article/240/1/1/7817994
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: at percentile 78 of works of the same type, year and subfield (OpenAlex work metrics, 11 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 2/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A robot and an XGBoost classifier guided crystallization experiments that produced nine reported polyoxometalate-scaffolding metal-organic frameworks. Researchers defined the chemistry and performed structural and electrochemical characterization. The measured electron-transfer behavior does not establish practical energy-storage deployment.
Technology: XGBoost with uncertainty-feedback retraining
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Nine frameworks were synthesized with reported repeatability.
- Still unknown: Independent replication, exhaustive structural prior art and device performance were not assessed.
- AI role: XGBoost with uncertainty-feedback retraining supported the stated study; experimental and interpretive work remained with researchers.
Evidence sources: https://eprints.gla.ac.uk/336574/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review date: 16 Nov 2026. Review scope: Institutional publication record and primary abstract; full paper and supplements not reread..
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Random-forest models using chemistry and geometry descriptors were trained on 240 prior amidase datasets and used to guide substrate and variant selection. The primary article reports that variant I198F/W328F achieved an enantioselectivity E-value of 303.5, compared with 5.7 for wild type, for the reported substrate assay. Most experimental substrate predictions agreed, with identified exceptions. This bounded laboratory result does not establish broad performance across substrates or industrial application.
Technology: Random-forest classification models and descriptor-guided variant design
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Primary paper reports the stated assay results.
- Still unknown: Independent replication and exhaustive prior-art review not established.
- AI role: Random-forest classification models and descriptor-guided variant design
Evidence sources: https://www.nature.com/articles/s41467-024-53048-0
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 16 Nov 2026. Review scope: Primary publisher full text and bibliographic metadata reviewed for identity, AI role, dates, affiliations and bounded results..
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Across six kidney-expression datasets, researchers identified 314 differentially expressed genes in diabetic nephropathy and used LASSO, SVM-RFE and random forest to prioritize seven candidate genes. Mendelian-randomization integration further prioritized carbonic anhydrase II (CA2); its ROC AUC exceeded 0.878 across the analyzed datasets. The authors also reported qPCR validation of CA2 in clinical patient and control samples and an association with inferred immune-cell infiltration.
Technology: LASSO, support-vector-machine recursive feature elimination and random-forest feature selection
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: The authors also reported qPCR validation of CA2 in clinical patient and control samples and an association with inferred immune-cell infiltration.
- Still unknown: This is a candidate biomarker study, not evidence of a clinically deployed test or patient benefit. Reported AUC values are retrospective dataset performance; qPCR validation supports differential expression but does not establish CA2 as causal or disease-specific. The accessible abstract does not report the clinical validation sample size, so none is inferred.
- AI role: Three machine-learning approaches integrated candidate-gene selection from kidney-expression datasets; Mendelian randomization with expression-QTL data then prioritized CA2, which was evaluated as a diagnostic biomarker and measured by qPCR in clinical samples. The models supported gene prioritization and discrimination, while qPCR supplied the wet-lab follow-up.
Evidence sources: https://pubmed.ncbi.nlm.nih.gov/39370621/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 10% of works of the same type, year and subfield (OpenAlex work metrics, 12 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 2/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
An uncertainty-aware Gaussian process regression workflow iteratively selected 23 synthetic α/β-peptide analogues for synthesis and measurements of fungal inhibition and human-red-blood-cell hemolysis. One lead, peptide 6-4, was a measured variant of an existing aurein-derived template and achieved a selectivity index of 57.1 (60.9 using the study’s finer MIC interval) versus 1.2 for aurein 1.2. Its high selectivity was driven by low hemolysis; the result is a specific peptide-design and assay contribution, not a new peptide scaffold or evidence of therapeutic efficacy. The ChemRxiv v1 result was posted on 2024-10-04; the peer-reviewed journal article followed in 2025.
Technology: Iterative uncertainty-aware Gaussian process regression using molecular descriptors and measured antifungal/hemolysis labels
Publication: Preprint · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Preprint
- Test type: direct test
- Tested: Across six iterative rounds, uncertainty-aware Gaussian process regression selected 23 peptide variants for synthesis and antifungal/hemolysis assays. Lead 6-4 had a selectivity index of 57.1 (60.9 using the study’s finer MIC interval), versus 1.2 for aurein 1.2; the authors attribute the selectivity to reduced hemolysis.
- Still unknown: Peptide 6-4 modifies a known aurein-derived template. The result is a measured sequence and assay outcome, not evidence of a new scaffold or therapeutic efficacy. The version-specific first public result is the 2024 ChemRxiv preprint; the journal article followed in 2025.
- AI role: An uncertainty-aware Gaussian process regression model selected peptide variants for synthesis and assays, and measured outcomes were fed back into subsequent selection rounds.
Evidence sources: https://chemrxiv.org/engage/chemrxiv/article-details/67a7e4386dde43c9081bdfc1 · https://pubs.rsc.org/en/content/articlepdf/2025/sc/d4sc06689h · https://pubmed.ncbi.nlm.nih.gov/40028619/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review date: 17 Nov 2026. Review scope: Primary ChemRxiv version-history record and v2 manuscript; primary RSC indexed full article PDF, including sequence tables, Results, Methods and publication metadata; PubMed primary abstract/affiliation metadata..
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A noise-aware Kriging model with active-learning acquisition selected candidate compositions from a million-composition NiTi-based alloy design space. After four experiments, researchers reported Ti₀.₂₆₆Ni₀.₄₈₄Hf₀.₂₅ with a measured latent heat of −36.08 J g⁻¹, 9.2% greater in magnitude than the best value in the initial training data (−33.04 J g⁻¹). The alloy also had an austenite-finish temperature of 481.71 °C and relatively small transformation hysteresis, making it a candidate for high-temperature thermal energy storage.
Technology: Noise-aware Kriging surrogate model with an active-learning acquisition strategy incorporating experimental-data noise
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: A noise-aware Kriging model with active-learning acquisition selected candidate compositions from a million-composition NiTi-based alloy design space.
- Still unknown: The reported latent heat retains the source’s negative sign; the 9.2% improvement is a comparison of absolute magnitudes against the initial dataset’s best value. This is one material result from an active-learning search, not exhaustive exploration of the million-composition space. The reported temperature/hysteresis properties are laboratory characterizations; the paper does not demonstrate a deployed thermal-storage module or independent reproduction.
- AI role: The model estimated the measurement-noise level from noisy experimental data, incorporated that level into Kriging predictions, and used a noise-augmented acquisition strategy to choose the next NiTi-based compositions. Four experiments in the closed loop identified Ti0.266Ni0.484Hf0.25 as a promising high-latent-heat alloy; researchers synthesized and measured the material properties.
Evidence sources: https://doi.org/10.1002/advs.202406216
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: at percentile 78 of works of the same type, year and subfield (OpenAlex work metrics, 15 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 2/5 · Validation 3/5 · Novelty 2/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Connectome of the adult Drosophila brain: nearly 140,000 neurons, over 15 million synaptic connections, and over 8,000 cell types. The page says deep neural networks segmented images and extracted neurons, but automated segmentation had millions of errors and required about 33 years of human correction across over 50 labs and citizen scientists.
Technology: Deep neural networks for segmentation
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: The connectome was checked against an earlier partial map, and connectomes were found to be highly consistent between animals.
- Still unknown: About 0.5% of neurons have developmental wiring defects, connection signs were predicted rather than measured, and the dataset was not built for a specific medical application.
- AI role: Deep neural networks segmented electron microscopy images and extracted neurons; the first automated pass had millions of errors and was corrected by humans.
Evidence sources: https://mrclmb.ac.uk/?p=4875
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: press release. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 758 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 5/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Minerva combines Gaussian-process optimization with automated high-throughput experiments to search reaction conditions. In the nickel-catalyzed Suzuki campaign, the preprint reports a condition with 76 area-percent yield and 92% selectivity, while two chemist-designed screening plates did not find successful conditions. Area percent is a chromatographic assay measure, not isolated yield. The preprint and expanded journal report are one study; this does not establish invention of Suzuki coupling or clinical utility.
Technology: Minerva Gaussian-process surrogate; Bayesian acquisition; automated HTE
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Experimental 96-well optimization reports 76 AP yield and 92% selectivity; two expert-designed screening plates did not find successful conditions.
- Still unknown: Product assay area-percent is not isolated yield; pharmaceutical structures and independent replication remain unreviewed.
- AI role: A GP surrogate and Bayesian acquisition choose batches for automated experiments and use measured outcomes to update condition search.
Evidence sources: https://chemrxiv.org/engage/api-gateway/chemrxiv/assets/orp/resource/item/66f976b5cec5d6c142981e4e/original/highly-parallel-optimisation-of-nickel-catalysed-suzuki-reactions-through-automation-and-machine-intelligence.pdf
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review date: 16 Nov 2026. Review scope: Primary ChemRxiv v1 full-text PDF and primary PubMed/journal metadata and abstract checked; journal full PMC direct open CAPTCHA challenged. Exact first-version content for later API case studies is excluded from draft..
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
An analysis of four months of Axial Seamount seismicity combined supervised event discrimination with unsupervised spectral clustering and found a mixed-frequency earthquake class whose counts rose rapidly about 15 hours before the 2015 eruption. The events migrated along mapped eruptive fissures; the authors interpret them as brittle crack opening associated with magma or volatile influx. This is a retrospective single-eruption finding, not an operationally validated forecast.
Technology: Unsupervised machine-learning spectral clustering, with supervised event discrimination
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: An analysis of four months of Axial Seamount seismicity combined supervised event discrimination with unsupervised spectral clustering and found a mixed-frequency earthquake class whose counts rose rapidly about 15 hours before the 2015 eruption.
- Still unknown: The result is retrospective and concerns a single 2015 eruption at one well-instrumented submarine volcano. The proposed crack-opening/magma-or-volatile explanation is the authors’ interpretation; applicability to future eruptions or other volcanoes remains untested.
- AI role: Discriminated volcanic earthquakes from other signal sources and clustered waveform spectral features, revealing a previously unrecognized mixed-frequency event pattern and its timing and migration before eruption.
Evidence sources: https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2024GL108631
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: at percentile 69 of works of the same type, year and subfield (OpenAlex work metrics, 7 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 2/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Deep-learning segmentation and spatial annotation supported cerebellar atlases of macaques, marmosets and mice, revealing primate-specific cell subtypes and regional gene-expression differences. Transcriptomic gradients corresponded with awake functional-connectivity patterns. This represents the September 27, 2024 journal version; an institutional announcement gives September 26, and the observed associations do not establish causal functions.
Technology: Deep-learning cell segmentation; STAGATE; DestVI
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: New spatial tissue measurements, biological replicates and comparisons with awake fMRI supported the reported anatomical associations within this study.
- Still unknown: Not assessed: exhaustive supplements, earlier-preprint chronology and independent confirmation. Functional mechanisms are not established by spatial correlations.
- AI role: Deep-learning models segment cells; STAGATE clusters spatial domains and DestVI assigns cell-type probabilities.
Evidence sources: https://pure.eur.nl/ws/files/169697439/Cross-species_single-cell_spatial_transcriptomic_atlases_of_the_cerebellar_cortex.pdf
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review scope and date are not recorded; source checked does not establish full-text review.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Two classifiers previously prioritized approved drugs from a screened set; ten candidates were purchased for assays against L. major and L. infantum. Five showed dose-dependent effects; intracellular testing and macrophage cytotoxicity were also reported. Domperidone and dibucaine had prior antiparasitic evidence, so the bounded result concerns the study’s remaining tested drugs and assays, not ten novel therapies.
Technology: Random forest and support-vector machine classifiers
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Primary full text describes ten tested predicted candidates, three Leishmania strains, intracellular testing and host-cell counter-assays.
- Still unknown: Earlier drug-specific literature, concentrations, achievable exposure and clinical efficacy require separate assessment.
- AI role: Previously trained activity models selected FDA-approved drugs for experimental parasite and host-cell testing.
Evidence sources: https://doi.org/10.3389/fcimb.2024.1403589
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 16 Nov 2026. Review scope: publisher primary full text. Material checked for the specific claims stated in the record..
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
The model covers about 8,000 drugs across 17,080 diseases and is nearly 50% better than competitors at identifying candidates and 35% more accurate on contraindications. It was tested on 1.2 million medical records and on three rare diseases excluded from training. Computational predictions only, with no laboratory or clinical confirmation.
Technology: TxGNN (graph neural network)
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: The model was checked against about 8,000 drugs and 17,080 diseases, validated on 1.2 million medical records, and asked to find drugs for three rare diseases excluded from training; it was compared with leading drug-repurposing models.
- Still unknown: Identified therapies would require further evaluation for dosing and delivery, and there is no laboratory or clinical confirmation.
- AI role: Proposed existing drugs for thousands of diseases, predicted contraindications and side effects, and gave a rationale for each prediction.
Evidence sources: https://hms.harvard.edu/news/researchers-harness-ai-repurpose-existing-drugs-treatment-rare-diseases
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: press release. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 303 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
An IBM model flagged 1,309 candidates; about a quarter were ground-checked over roughly six months, finding 303 new figurative geoglyphs among 430 already known and increasing the discovery rate 16-fold. Another 968 candidates remain to be checked; the study had been accepted by PNAS at the time of the announcement.
Technology: Neural network for aerial images
Publication: Announcement only · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Announcement only
- Test type: direct test
- Tested: Field teams ground-checked about a quarter of the 1,309 AI-flagged candidates on foot over roughly six months, finding 303 new figurative geoglyphs.
- Still unknown: 968 candidates remain unsurveyed, training data was limited, and the study was accepted but not yet published in PNAS at the time of the announcement.
- AI role: An IBM model that works with few training samples analyzed aerial geospatial imagery to flag likely geoglyph areas.
Evidence sources: https://www.yamagata-u.ac.jp/en/information/info/20240924/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: press release. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 24 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 4/5 · Validation 3/5 · Novelty 4/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
PNNL and Argonne identified solvent mixtures that dissolve three times more than a compound proposed for flow batteries, exploring less than 10% of about 2,000 combinations. The model receives experimental data and guides subsequent experiments; the result concerns solubility, without demonstrating performance of a complete battery.
Technology: Machine learning in a loop with high-throughput experiments
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Active learning guided solubility measurements of 2,1,3-benzothiadiazole in binary solvents: 218 measurements from 2,101 candidates found 18 mixtures above 6.20 M; best 6.50 M. On a 98-solvent set, Bayesian optimization needed 17 ± 11 trials versus about 50 ± 27 random.
- Still unknown: Tested one solute, binary solvents only; solubility is not the only property and viscosity, conductivity and stability were not optimized. Model accuracy on 40 test solvents was R2 = 0.63. Real battery performance not reported.
- AI role: A Gaussian process model with an expected-improvement rule chose which solvents to measure; a robotic platform prepared samples and NMR measured solubility; humans selected features and candidate lists.
Evidence sources: https://doi.org/10.1038/s41467-024-47070-5
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: press release. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 102 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A moment-tensor machine-learning interatomic potential enabled large-scale simulations of CrCoNi processing. The authors report that thermomechanical deformation and solidification can generate steady-state chemical short-range order outside the equilibrium-annealing spectrum, and propose an ordering-bias mechanism. This is a simulation-based computational finding, not direct experimental validation of the predicted states in manufactured alloys. The result and AI role were already present in arXiv v1; journal publication followed in 2025.
Technology: Moment-tensor machine-learning interatomic potential; E(3)-equivariant graph neural network for local chemical-motif identification
Publication: Preprint · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Preprint
- Test type: direct test
- Tested: The primary v1 reports molecular-dynamics and nonequilibrium Monte Carlo simulations of CrCoNi thermomechanical deformation and solidification; these simulations yield steady-state SRO states outside the equilibrium-annealing spectrum.
- Still unknown: This is a simulation-based result, not direct experimental confirmation of every predicted state or industrial deployment. Broad prior-art screening is bounded; independent reproduction is not established.
- AI role: A moment-tensor machine-learning interatomic potential (TS-f) supplies energies and forces for simulations; an E(3)-equivariant graph neural network separately identifies local motifs and supports SRO quantification.
Evidence sources: https://arxiv.org/pdf/2409.15474v1 · https://arxiv.org/abs/2409.15474 · https://www.nature.com/articles/s41467-025-64733-z · https://pubmed.ncbi.nlm.nih.gov/41062492/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review date: 17 Nov 2026. Review scope: Primary arXiv v1 full text and submission history; indexed Nature Communications full text including Methods/contributions and affiliation block; PubMed metadata/abstract..
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A GPU-based transit search paired with a convolutional neural network identified five sub-Earth-size ultra-short-period transit signals in Kepler light curves: Kepler-158d, Kepler-963c, Kepler-879c, Kepler-1489c, and KOI-4978.02. The authors applied transit vetting, false-alarm tests, and statistical false-positive probabilities below 1% to each, and statistically validated the five signals as planets. The validation is photometric/statistical rather than independent radial-velocity or spectroscopic confirmation; KOI-4978.02 remains designated as a KOI object in the source.
Technology: GPFC algorithm: GPU phase folding with convolutional neural network transit scoring
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: A GPU-based transit search paired with a convolutional neural network identified five sub-Earth-size ultra-short-period transit signals in Kepler light curves: Kepler-158d, Kepler-963c, Kepler-879c, Kepler-1489c, and KOI-4978.02.
- Still unknown: The paper calls the five objects statistically validated, not all independently confirmed by spectroscopy or radial velocities. Validation relies on Kepler photometry, transit vetting, estimated false-positive probabilities and false-alarm tests. KOI-4978.02 retains its KOI designation. One paper-level result; individual planets are not split into separate entries.
- AI role: GPU phase folding searched Kepler light curves and a CNN scored/identified transit-like signals; follow-up statistical and transit-shape tests vetted the model-discovered signals. Authors report FPP below 1% for all five.
Evidence sources: https://academic.oup.com/mnras/article/534/3/1913/7762975
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: at percentile 50 of works of the same type, year and subfield (OpenAlex work metrics, 4 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A learned drug-target connectivity signature selected 23 predicted hits and 25 negatives for testing in four glioblastoma patient samples. Ex-vivo measurements agreed with predictions at a reported mean ROC AUC of 0.94. The small validation establishes a bounded drug-activity prediction result, not clinical response. Vortioxetine was found by the initial experimental screen and is not attributed to COSTAR nomination.
Technology: COSTAR logistic LASSO drug-target model
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Newly selected compounds were tested ex vivo across four patient samples.
- Still unknown: Patient benefit, generalization and clinical safety remain unestablished.
- AI role: COSTAR logistic LASSO drug-target model supported the stated study; experimental and interpretive work remained with researchers.
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC11564103/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review date: 16 Nov 2026. Review scope: Primary full-text model Methods and selected experimental Results, publication metadata and affiliations; supplements not reviewed..
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A RoseTTAFold2-Lite pipeline combined residue coevolution, deep-learning complex prediction and AlphaFold modelling to screen protein pairs across 19 human bacterial pathogens. In a validation set selected using the deep-learning pipeline, co-immunoprecipitation detected four of six predicted interactions. Three pairs lacked prior direct interaction evidence: E. coli UbiE–YcaR, P. aeruginosa PA4105–PA4106, and P. aeruginosa WbpG–HisF2. The PA4105–PA4106 structural similarity to proteins in 3-thiaglutamate biosynthesis and the WbpG–HisF2 result provide functional hypotheses, not established native-pathogen mechanisms.
Technology: RoseTTAFold2-Lite integrated with direct coupling analysis and AlphaFold complex modelling
Family AlphaFold: profile · models page
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: In a validation set selected using the deep-learning pipeline, co-immunoprecipitation detected four of six predicted interactions.
- Still unknown: The positive interaction evidence comes from heterologous co-immunoprecipitation/pull-down assays, with two independent experiments per positive pair, rather than native-pathogen systems. The validation subset was small (four positives among six AI-selected candidates); most of the thousands of predicted interactions remain untested, and false positives and false negatives are expected. Proposed roles for PA4105–PA4106 and UbiE–YcaR remain hypotheses; this paper does not establish their physiological function or a therapeutic target.
- AI role: RoseTTAFold2-Lite used residue–residue coevolution and predicted complex structures to rank candidate protein pairs in a proteome-wide screen. The authors selected six pairs from the final deep-learning pipeline for co-immunoprecipitation tests; four yielded positive interaction signals. Three were described as lacking previously reported direct interaction evidence, including UbiE–YcaR, PA4105–PA4106 and WbpG–HisF2. The models enabled the experimental target selection; they did not themselves establish physical interactions.
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC11445079/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 10% of works of the same type, year and subfield (OpenAlex work metrics, 52 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Physically informed feature selection derived an ARSC descriptor from calculated catalyst properties to guide oxygen-reaction candidate design. The study reports synthesis and alkaline testing of two carbon-supported samples: ORR half-wave potentials of 0.941 and 0.937 V and OER overpotentials of 330 and 340 mV at 10 mA cm−2. These are laboratory measurements; structural analyses support but do not prove every proposed active-site configuration.
Technology: ARSC descriptor; feature selection
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Publisher PDF reports synthesized Co–Co and Co–Ir samples with stated ORR and OER tests.
- Still unknown: Prior composition overlap and independent replication remain open; specific active-site configuration is not directly resolved for every site.
- AI role: Calculated dual-atom catalyst descriptors selected candidates followed by synthesis and electrochemical testing.
Evidence sources: https://www.nature.com/articles/s41467-024-52519-8
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 16 Nov 2026. Review scope: Primary source checked for identity, publication date, AI workflow and the bounded experimental result stated..
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
An experiment-driven active-learning workflow selected pentapeptide sequences for testing of beta-sheet self-assembly. Among 268 synthesized and tested peptides, 96 formed beta-sheet assemblies, including sequences such as ILFSM, LMISI, MITIY, MISIW and WKIYI that conventional propensity rules did not predict. The result concerns measured assembly under the study conditions; it does not establish a functional material application or therapeutic effect. The outcome was already reported in the September 2024 preprint and is dated accordingly.
Technology: AI-driven active learning with machine-learning models that selected peptide sequences for iterative synthesis and testing
Publication: Preprint · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Preprint
- Test type: direct test
- Tested: The v1 reports synthesis/testing of 268 pentapeptides and 96 beta-sheet-forming assemblies, including several unconventional sequences.
- Still unknown: Assembly findings are bounded to the reported experimental conditions; no application-level performance, therapeutic efficacy or independent replication is established. The specific result was present in the preprint before journal publication.
- AI role: AI-driven active learning prioritized sequences for iterative synthesis and experimental characterization, focusing in part on cases where model predictions differed from conventional propensity tables.
Evidence sources: https://assets-eu.researchsquare.com/files/rs-5100605/v1_covered_2002f639-5284-4380-a401-67cb063dd68a.pdf · https://www.nature.com/articles/sciadv.adt9466
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review date: 17 Nov 2026. Review scope: Primary PMC full article for active-learning workflow and measured assemblies; primary Research Square v1 manuscript for September 2024 version date and named affiliations; primary AAAS article for the peer-reviewed version and reported results..
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Across experiments with more than 2,000 participants, three dialogues with GPT-4 reduced belief in the targeted conspiracy theory by about 20%, and the effect persisted for two months and spilled over to unrelated conspiracies. A professional fact-checker rated 99.2% of 128 sampled chatbot claims as true. Caveats: lab setting, the chatbot did not reduce belief in true conspiracies, and most participants still held the hoax belief.
Technology: GPT-4 (DebunkBot dialogue system)
Family GPT-4: profile · models page
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: More than 2,000 conspiracy believers held three personalized, evidence-based dialogues with GPT-4 Turbo, and belief change was measured, including persistence at two months and spillover to unrelated conspiracies.
- Still unknown: The abstract does not describe a control group and states no limitations; the page metadata lists a 2026 expression of concern and erratum.
- AI role: GPT-4 Turbo delivered the personalized counterevidence in the dialogues.
Evidence sources: https://www.ebi.ac.uk/europepmc/webservices/rest/search?query=DOI:10.1126/science.adq1814&resultType=core&format=json
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 529 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 4/5 · Validation 3/5 · Novelty 4/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A ResNet-based DASEventNet model applied to continuous distributed acoustic sensing from the April 2022 Utah FORGE Well 16A (78)-32 stimulation manually verified 5,749 microearthquakes absent from the Silixa STA/LTA catalog; 45 of 5,794 model-positive windows were false detections on manual review. The study reports events down to Mw −1.80 versus −1.14 for STA/LTA and a broader Stage 3 seismicity distribution consistent with fracture geometry seen in an independent deep-borehole geophone catalog. This is a local event-catalog and stimulation-geometry result, not a direct DAS-based event-location measurement.
Technology: DASEventNet (ResNet-50 convolutional neural network)
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: A ResNet-based DASEventNet model applied to continuous distributed acoustic sensing from the April 2022 Utah FORGE Well 16A (78)-32 stimulation manually verified 5,749 microearthquakes absent from the Silixa STA/LTA catalog; 45 of 5,794 model-positive windows were false detections on manual review.
- Still unknown: Original JGR: Solid Earth paper reports manual review of model-positive events and comparison with conventional Silixa detections. Model predicts event presence but does not estimate event locations or magnitudes from DAS alone; timing association and location/magnitude comparison use an independent deep-borehole geophone catalog with an assumed 0.2 s matching window. Fiber placement limits direct geometry inference. Detection of cataloged events is not independent proof of every physical interpretation.
- AI role: Trained on manually checked DAS event and noise windows, the network detected weak microearthquake waveforms in continuous seven-day DAS recordings. It expanded the catalog and enabled a denser spatiotemporal characterization of the hydraulic stimulation seismicity.
Evidence sources: https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2024JB029102
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 10% of works of the same type, year and subfield (OpenAlex work metrics, 29 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 2/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A generative chemical language model was fine-tuned on known ligands for paired targets. Prospective synthesis and in-vitro assays found seven of twelve tested designs modulated both intended targets, with confirmed dual activity across FXR/sEH, FXR/THRβ and PPARδ/sEH. Activities varied substantially and were often unbalanced; the strongest reported sEH inhibitor, compound 7, had IC50 17 nM with micromolar PPARδ agonism.
Technology: Target-pair-fine-tuned chemical language model (four-layer LSTM recurrent neural network trained on SMILES, pretrained on 365,000 ChEMBL molecules)
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Prospective synthesis and in-vitro assays found seven of twelve tested designs modulated both intended targets, with confirmed dual activity across FXR/sEH, FXR/THRβ and PPARδ/sEH.
- Still unknown: This is in-vitro activity of a small set of synthesized molecules, not evidence of therapeutic efficacy or synergistic clinical benefit. Potencies were often unbalanced; several selected candidates were active on only one intended target, and the authors caution the success rate may not generalize beyond target pairs with compatible ligand chemistry. Lead optimization remains necessary.
- AI role: The pooled fine-tuning strategy combined known ligands of each target pair to bias de-novo SMILES generation toward both ligand classes. Researchers used model sampling frequency, similarity and docking to select molecules for synthesis; assays then tested receptor agonism and soluble epoxide hydrolase inhibition. The model proposed chemical candidates, while biological activity was measured experimentally.
Evidence sources: https://www.nature.com/articles/s41467-024-52060-8
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 10% of works of the same type, year and subfield (OpenAlex work metrics, 62 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
MILTON uses machine learning on routine clinical biomarkers, quantitative traits and plasma proteomics to predict disease phenotypes and augment case labels in UK Biobank. In a phenome-wide analysis of 484,230 genome-sequenced participants, the authors report 182 putative novel gene–disease associations that were not genome-wide significant in baseline cohorts or explained by biomarker associations; 38.9% of the 167 mappable signals were nominally supported with the same direction in FinnGen.
Technology: MILTON (Machine Learning with Phenotype Associations), an ensemble machine-learning framework for disease prediction from clinical and molecular biomarkers
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: In a phenome-wide analysis of 484,230 genome-sequenced participants, the authors report 182 putative novel gene–disease associations that were not genome-wide significant in baseline cohorts or explained by biomarker associations; 38.9% of the 167 mappable signals were nominally supported with the same direction in FinnGen.
- Still unknown: The authors classify the 182 associations as putative novel; they are computational associations, not experimentally established causal genes. FinnGen nominal same-direction support applied to 38.92% (65/167) of the mappable putative novel PheWAS associations, while the paper also reports other aggregate exome-wide results. A separate example, APOB–thyrotoxicosis, showed FinnGen support. These signals require further biological validation.
- AI role: MILTON predicted cases from clinical biomarkers and other quantitative traits, augmenting phenotype labels for genetic association analyses. The AI-enabled cohort enrichment yielded putative novel gene–disease signals; the study compared them with baseline analyses, orthogonal prioritization tools and FinnGen.
Evidence sources: https://www.nature.com/articles/s41588-024-01898-1.pdf
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 137 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Machine-learning pH-optimum predictors screened metagenomic sequences, and two predicted Lepista nuda laccases were produced and profiled. The authors report LnLccA and LnLccB optimum activity at pH 10 and 9 for substrate 2,6-DMP, respectively, while both had pH 4 optima for ABTS. The substrate-dependent contrast is retained; it is not a general alkaline optimum or industrial demonstration.
Technology: Supervised, semi-supervised and unsupervised pH-optimum prediction models
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Open primary full article reports recombinant LnLccA and LnLccB pH profiles, substrate-specific optima and validation of ML predictions.
- Still unknown: Prior characterization and novelty of sequence/activity remain open; industrial applicability not demonstrated.
- AI role: Models trained on curated laccase pH data predicted candidates from metagenomic sequences that were expressed and profiled.
Evidence sources: https://link.springer.com/article/10.1186/s13068-024-02566-6
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 16 Nov 2026. Review scope: Primary source checked for identity, date, model role and bounded reported result..
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A human-in-the-loop model selected nickel-cobalt catalysts and reaction settings for a 48-test laboratory campaign. Further selected experiments improved methane space-time yield by nearly 50% over the previous maximum in that dataset. Extrapolation remained harder than interpolation. This is a bounded catalyst-and-condition result, not evidence of industrial deployment or lifecycle climate benefit.
Technology: Gaussian-process active learning with batch acquisition
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Selected catalysts and conditions were experimentally tested; model-suggested activity trends received follow-up.
- Still unknown: Generality, industrial durability and lifecycle carbon benefit were not established.
- AI role: Gaussian-process active learning with batch acquisition supported the stated study; experimental and interpretive work remained with researchers.
Evidence sources: https://pubs.rsc.org/en/content/articlehtml/2024/cy/d4cy00873a
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review date: 16 Nov 2026. Review scope: Indexed primary publisher abstract, model Methods, selected Results, affiliations and first-online date; supplementary material not reviewed..
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Unsupervised clustering of 25 echocardiographic and biomarker variables in 1,199 ARIC participants with diabetes identified three phenogroups; the group with the greatest cardiac remodelling and neurohormonal-stress abnormalities had 12.1% five-year heart-failure incidence, versus 4.6% and 3.1% in the other groups. A deep-neural-network classifier trained to recognize this phenotype identified corresponding high-risk groups in the external CHS cohort (n=802) and UT Southwestern EHR cohort (n=5,071); phenotype classification was associated with incident heart failure in both.
Technology: Unsupervised hierarchical clustering and a deep neural network (DeepNN) classifier, with random forest variable-importance analysis
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: Unsupervised clustering of 25 echocardiographic and biomarker variables in 1,199 ARIC participants with diabetes identified three phenogroups; the group with the greatest cardiac remodelling and neurohormonal-stress abnormalities had 12.1% five-year heart-failure incidence, versus 4.6% and 3.1% in the other groups.
- Still unknown: The cohorts were observational; association with future heart failure does not establish that model-guided intervention improves outcomes. The phenotype is an operational study definition, not a universally accepted diagnostic threshold. The authors describe potential preventive use, but it was not tested. The cohorts were external validations rather than prospective deployment trials.
- AI role: The authors used unsupervised hierarchical clustering of echocardiography and cardiac biomarkers to define outcome-associated diabetic-cardiomyopathy phenogroups, then trained and externally validated a DeepNN classifier to recognize the high-risk phenotype. Random forest permutation importance prioritized NT-proBNP, left-ventricular mass, Doppler measurements, left-atrial dimension and LV end-diastolic volume.
Evidence sources: https://onlinelibrary.wiley.com/doi/10.1002/ejhf.3443
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: at percentile 88 of works of the same type, year and subfield (OpenAlex work metrics, 8 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A system for designing high-affinity protein binders, tested against seven targets, achieved 88% laboratory success on BHRF1 and bindings 3 to 300 times stronger than the best existing methods. It produced no valid binders for TNFα; strong binding is only a first step.
Technology: AlphaProteo
Family AlphaProteo: profile · models page
Publication: Preprint · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Preprint
- Test type: direct test
- Tested: Binders were designed against seven target proteins and tested in the lab, with 3 to 300 times stronger binding and higher experimental success rates than the best existing methods.
- Still unknown: Per-target success rates are not given on the abstract page and no limitations are listed.
- AI role: A family of machine learning models for protein design that generates the binders.
Evidence sources: https://arxiv.org/abs/2409.08022
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: blog. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
The OpenVS platform combined target-specific neural-network active learning with RosettaVS physics-based docking to search billion-scale compound libraries. The study reported seven low-micromolar KLHDC2 degron-pocket binders and four Naᵥ1.7 VSD4 binders. KLHDC2 hit C29 competed with a degron peptide at IC50 about 2.9 μM and its predicted pose was supported by a 2.0 Å crystal structure; the strongest Naᵥ1.7 hit, Z8739902234, had IC50 1.3 μM in whole-cell electrophysiology.
Technology: OpenVS active-learning platform with target-specific feed-forward neural-network triage, coupled to RosettaVS physics-based docking and RosettaGenFF-VS scoring
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: The study reported seven low-micromolar KLHDC2 degron-pocket binders and four Naᵥ1.7 VSD4 binders.
- Still unknown: Early ligand-hit discovery only: binding/inhibition was measured in biochemical competition, biolayer interferometry and cell electrophysiology assays, not in animals or patients. The paper states further optimization is needed to improve KLHDC2 compounds toward nanomolar potency. KLHDC2 C29’s crystal structure is experimental support for one predicted binding pose; target modulation, therapeutic efficacy and safety were not established. The reported target hit rates apply to manually filtered, synthesized subsets, not the full billion-molecule library.
- AI role: The target-specific neural network used docking-labeled data to guide which molecules were prioritized for additional expensive docking over iterative multi-billion-compound screens; RosettaVS/RosettaGenFF-VS performed physics-based docking and pose/affinity scoring. The AI-guided screen nominated molecules that were synthesized and experimentally tested. Experiments confirmed KLHDC2 degron displacement and a crystal structure matching the predicted pose, plus Naᵥ1.7 binding/inhibition in patch clamp. The paper attributes discovery to the combined AI-accelerated and physics-based workflow, not to a generative model alone.
Evidence sources: https://www.nature.com/articles/s41467-024-52061-7.pdf
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 290 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
DeepBiosphere combines a residual CNN trained on citizen-science observations with aerial imagery to map California distributions for 2,221 plant species. The authors report median AUC 0.95 versus 0.88 for climate-only models and use repeat imagery to map plant-community change associated with the 2013 Rim Fire. These are California-specific ecological maps and observational associations; the study does not establish causal effects of fire or a new species discovery.
Technology: DeepBiosphere residual convolutional neural network for species-distribution mapping
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: The primary article reports held-out mapping median AUC 0.95 versus 0.88 for climate-only models and uses repeat imagery to map community change following the 2013 Rim Fire.
- Still unknown: The fire-severity relationship is observational and does not establish causation. Mapping performance is California-specific and weaker for rare species.
- AI role: A residual CNN trained with citizen-science observations and aerial imagery produced fine-scale species-distribution maps that were then used to estimate rapid plant-community change.
Evidence sources: https://www.pnas.org/doi/10.1073/pnas.2318296121 · https://pmc.ncbi.nlm.nih.gov/articles/PMC11406280/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review date: 17 Nov 2026. Review scope: Primary PNAS indexed article and PMC full-text indexed content, including abstract, Methods, results and plant/fire case studies; PubMed primary metadata/affiliations and first-online date..
Editorial assessment: Impact 3/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Trained on 15 million patches and 60,000 slides and tested on more than 19,400 slides from 32 datasets and 24 hospitals: it detects cancer with about 94% accuracy across 15 datasets, predicts mutations in 54 genes with over 70% accuracy, and improves survival prediction by 8% across 17 institutions. Retrospective validation; not trained on rare diseases or premalignant tissue.
Technology: CHIEF (pathology foundation model)
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: Retrospective validation on more than 19,400 slides from 32 datasets at 24 hospitals; cancer detection was about 94% accurate across datasets.
- Still unknown: Validation beyond these datasets is still needed, rare and non-cancerous conditions and pre-malignant tissue were not tested, and gene prediction accuracy varies by gene and cancer type.
- AI role: Read digital tumor slides to detect cancer, predict gene mutations and survival, and generate heat maps of tissue features for pathologists.
Evidence sources: https://hms.harvard.edu/news/new-artificial-intelligence-tool-cancer
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: press release. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 685 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A conditional variational autoencoder generated candidate alloy compositions and an artificial neural network ranked eutectic-forming likelihood. The team experimentally validated previously unreported eutectic compositionally complex alloys across quaternary to senary systems. Reported examples include Al₁₆Cr₁₈Fe₂₄Ni₄₂ with a dual-phase FCC/BCC eutectic lamellar structure and Al₁₁.₃Nb₁₃.₇Co₃₆.₆Cr₈.₂Fe₇.₉Ni₂₂.₃ with a tri-phase FCC/BCC/Laves eutectic structure, supported by XRD, SEM and DSC.
Technology: Conditional variational autoencoder (CVAE) and artificial neural network (ANN)
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: A conditional variational autoencoder generated candidate alloy compositions and an artificial neural network ranked eutectic-forming likelihood.
- Still unknown: The authors report five newly identified ECCA compositions across the studied systems; this is one paper-level result, not five records. The entry names two representative experimentally examined compositions. Validation is limited to the reported arc-melted samples and XRD/SEM/DSC characterization; no long-term mechanical performance or application testing is established. The training data were sparse and imbalanced, and dataset filtering was part of the framework.
- AI role: A CVAE generated candidate compositions conditioned on eutectic/non-eutectic microstructure labels; an ANN trained on composition and thermodynamics-derived descriptors estimated eutectic-forming likelihood and selected generated candidates for experimental validation. Dataset preprocessing with PCA and K-means addressed class imbalance. The ML pipeline directly proposed the compositions whose eutectic structures were examined experimentally.
Evidence sources: https://www.nature.com/articles/s41524-024-01385-5
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 10% of works of the same type, year and subfield (OpenAlex work metrics, 51 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Semi-supervised co-training with kernel ridge regression and support-vector regression expanded the labeled phosphor dataset and selected Lu3Y2Ga3O12:Cr3+. The study reports emission near 750 nm with internal and external quantum efficiencies of 97.1% and 38.8%, respectively, and a fabricated laboratory phosphor-converted LED. The abstract’s values are author-reported laboratory measurements, not commercial product performance.
Technology: Semi-supervised KRR/SVR co-training
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Primary publisher article reports synthesized candidate, emission/efficiencies and laboratory pc-LED.
- Still unknown: Prior composition-level disclosures and independent device durability are not established.
- AI role: Co-training predicts emission peak and screens candidate garnet compositions for experimental synthesis.
Evidence sources: https://www.nature.com/articles/s41524-024-01395-3
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 16 Nov 2026. Review scope: Primary source checked for identity, date, model role and bounded reported result..
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Interpreting a microfluidic droplet experiment movie with a Fourier neural operator and sparse regression yielded a KdV–Burgers-type continuum model containing a dispersive term absent from the prior Burgers description. The authors’ simulations attribute the nonlinear-wave propagation and stabilization to nonreciprocal hydrodynamic interactions. The work analyzes existing experimental data and does not report a new experiment or a general proof for all nonequilibrium crystals.
Technology: Fourier neural operator, interpretable neural-network pipeline and sparse regression
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: The authors trained/interpreted a neural-operator network on a microfluidic droplet experiment movie and used sparse regression to derive a KdV–Burgers-type model. The new dispersive term improved the shock-front description relative to Burgers’ equation; a microscopic model attributes nonlinear-wave stabilization/propagation to nonreciprocal hydrodynamic interactions. The source says the model was trained on an existing experimental movie.
- Still unknown: This is a model-derived interpretation of a specific microfluidic experiment and supporting simulations, not a new experiment or a general proof for all nonequilibrium crystals.
- AI role: A Fourier neural operator learned latent experimental dynamics from the microfluidic movie; sparse regression extracted the governing equation, and the learned term guided the interpretation of nonreciprocal hydrodynamic interactions.
Evidence sources: https://journals.aps.org/prl/abstract/10.1103/PhysRevLett.133.107301 · https://link.aps.org/accepted/10.1103/PhysRevLett.133.107301
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review date: 17 Nov 2026. Review scope: Primary APS abstract and accepted manuscript for the neural-operator and sparse-regression analysis of microfluidic wave data, affiliation block, source data and publication chronology..
Editorial assessment: Impact 2/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Researchers at Hunan University and collaborators used the peptide language model deepAMP to optimize antimicrobial peptides and experimentally tested 29 candidates. More than 90% inhibited Gram-positive and Gram-negative bacteria better than the reference peptide penetratin; three candidates showed reduced resistance emergence in serial-passage assays and activity in a mouse Pseudomonas wound-infection model.
Technology: deepAMP peptide language model (deepAMP-general, deepAMP-AOM, deepAMP-POM and deepAMP-predict)
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Researchers at Hunan University and collaborators used the peptide language model deepAMP to optimize antimicrobial peptides and experimentally tested 29 candidates.
- Still unknown: This is a separate discovery study from the AMPSphere microbiome-mining paper already in the tracker: deepAMP optimizes peptide sequences, whereas AMPSphere mined microbiomes. Testing involved 29 candidates and one female mouse wound model; this is preclinical evidence, not clinical efficacy. The article itself calls for more animal-model validation.
- AI role: The generative and ranking models optimized candidate peptide sequences for antimicrobial potency and membrane-disrupting activity. Researchers synthesized and tested selected sequences in vitro and in a mouse wound infection experiment.
Evidence sources: https://doi.org/10.1038/s41467-024-51933-2
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 159 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Classifiers identified receiver-related structure in calls from ten captive marmosets. Playback experiments also found stronger responses to calls originally directed at the listener. The authors interpret these patterns as vocal labels; a later commentary disputes whether the evidence establishes name-like reference, and the authors defend that interpretation. The bounded finding is receiver-specific acoustic and behavioral differentiation, not decoded human-like names.
Technology: Random-forest acoustic classifiers
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Recorded calls and playback responses were analyzed.
- Still unknown: Name-like reference and alternative explanations remain debated; the captive sample is small.
- AI role: Random-forest acoustic classifiers supported the stated study; experimental and interpretive work remained with researchers.
Evidence sources: https://pubmed.ncbi.nlm.nih.gov/39208084/ · https://zoology.ubc.ca/edg/pdfs/Oren2024.pdf · https://pubmed.ncbi.nlm.nih.gov/39747802/ · https://pubmed.ncbi.nlm.nih.gov/40866605/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review date: 16 Nov 2026. Review scope: Indexed primary published-paper Methods and abstract, PubMed metadata, critical commentary abstract and authors response abstract..
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A framework combining single-cell ATAC-seq, motif analysis, topic modeling, machine learning and genetic variation identified SMARCC1 as a previously unrecognized controller of regulatory T-cell chromatin programs. CRISPR-Cas9 deletion of Smarcc1 in Tregs produced the predicted loss of accessibility in topic-3 regulatory regions, supporting the computationally inferred connection.
Technology: Integrated machine-learning framework using topic modeling and high-density genetic variation over single-cell ATAC-seq profiles
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: A framework combining single-cell ATAC-seq, motif analysis, topic modeling, machine learning and genetic variation identified SMARCC1 as a previously unrecognized controller of regulatory T-cell chromatin programs.
- Still unknown: CRISPR validation supported the predicted SMARCC1 effect for topic 3; topic 10 was not represented and the authors note possible redundancy with SMARCC2. The result concerns regulatory chromatin programs in Tregs, not clinical or therapeutic outcomes.
- AI role: The framework organized co-varying open chromatin regions into Treg regulatory modules, inferred transcription-factor effects on those programs, and nominated SMARCC1 as a regulator of topic-3/10 accessibility. The team tested the prediction by CRISPR-Cas9 editing Tregs and profiling chromatin accessibility.
Evidence sources: https://cbdm.hms.harvard.edu/assets/Publications/2024pub/chowdharyPNAS.pdf
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 10% of works of the same type, year and subfield (OpenAlex work metrics, 32 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A closed-loop synthesis and learning campaign found that solution photostability across a donor–bridge–acceptor molecular series correlated with the high-energy portion of the triplet excited-state manifold, beyond the commonly considered lowest triplet state. Researchers tested the model-derived hypothesis on an experimental set and changed solvent triplet energy to test the proposed Dexter triplet-energy-transfer degradation pathway. The study used Bayesian optimization and interpretable supervised learning to extract the chemical hypothesis from measured photostability data.
Technology: Closed-loop transfer (CLT), combining Gryffin Bayesian optimization with physics-feature supervised learning
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Researchers tested the model-derived hypothesis on an experimental set and changed solvent triplet energy to test the proposed Dexter triplet-energy-transfer degradation pathway.
- Still unknown: The finding concerns solution photostability in the studied donor–bridge–acceptor chemical space; it does not by itself establish a universal law for all organic molecules or performance in processed films/devices. The work used human expertise in the loop and experimentally tested the model-generated hypothesis. Device deployment was not demonstrated.
- AI role: Bayesian optimization recommended molecules for automated modular synthesis and photostability assays; an interpretable supervised model using physics-based features identified a triplet-state hypothesis, which researchers tested with experimental test sets and solvent-energy perturbation. AI guided experimental selection and surfaced a specific chemical relationship; the follow-up measurements tested it.
Evidence sources: https://schroeder.cbe.princeton.edu/sites/g/files/toruqf7951/files/documents/Nature-2024.pdf
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 99 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A random-forest method using statistical features from earthquake catalogs identified abnormal low-magnitude seismicity spreading across roughly 15–25% of Southern California and Southcentral Alaska up to about three months before the 2019 Ridgecrest sequence and 2018 Anchorage M7.1 earthquake. Finite-element mechanics models were used to propose a pore-fluid-pressure explanation. This is a retrospective analysis of two target events, not a demonstrated reliable warning system.
Technology: Random-forest machine-learning algorithm with earthquake-catalog statistical features
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: A random-forest method using statistical features from earthquake catalogs identified abnormal low-magnitude seismicity spreading across roughly 15–25% of Southern California and Southcentral Alaska up to about three months before the 2019 Ridgecrest sequence and 2018 Anchorage M7.1 earthquake.
- Still unknown: The study analyzes two retrospective target events and explicitly notes that earthquake precursors remain ambiguous and reliable alert systems are not established. The pore-pressure mechanism is proposed from modeling rather than directly measured, and the finding is not prospective validation.
- AI role: Detected regional nonlinear patterns in low-magnitude seismic catalogs that the authors report as precursory unrest before two large earthquake sequences; the AI analysis identified the statistical pattern, while finite-element models supported a proposed mechanism.
Evidence sources: https://www.nature.com/articles/s41467-024-51596-z
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 10% of works of the same type, year and subfield (OpenAlex work metrics, 22 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A convolutional neural network trained on idealized Earth-system-model simulations classified eruption occurrence and broad latitude band from global surface-temperature anomalies in the first post-eruption summer. Applied to reanalysis and observational climate records, the method identified major historic events and assigned a northern-extratropical footprint to the unidentified 1809 eruption. The same analysis found that some smaller events leave weaker, less distinguishable temperature signals.
Technology: Convolutional neural network with explainable-AI relevance mapping
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: A convolutional neural network trained on idealized Earth-system-model simulations classified eruption occurrence and broad latitude band from global surface-temperature anomalies in the first post-eruption summer.
- Still unknown: The model is trained on idealized simulations with only three broad eruption latitude classes and four sulfur-injection strengths. The 1809 geographic inference is an AI classification of a climate fingerprint, not direct observation of the eruption source; predictions show sensitivity to normalization and differing observational datasets. It should be interpreted as a clue, not a definitive eruption-location determination.
- AI role: The CNN classified volcanic presence/location using post-eruption temperature patterns; explainability analysis localized climate regions driving those classifications and helped infer the 1809 event’s northern-extratropical footprint from observational/reanalysis data.
Evidence sources: https://www.nature.com/articles/s43247-024-01617-y
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: at percentile 41 of works of the same type, year and subfield (OpenAlex work metrics, 2 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 2/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A deep-learning system called MitoReID classified drug mechanisms from time-resolved mitochondrial morphology and membrane-potential images. Among its natural-compound predictions, it nominated epicatechin as a COX-2 inhibitor; the authors report in-vitro follow-up including COX-2 activity, cellular thermal-shift and surface-plasmon-resonance assays. This is a target-mechanism finding, not evidence of clinical efficacy.
Technology: MitoReID, deep-learning re-identification model using an Inflated 3D ResNet backbone on temporal mitochondrial images
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Among its natural-compound predictions, it nominated epicatechin as a COX-2 inhibitor; the authors report in-vitro follow-up including COX-2 activity, cellular thermal-shift and surface-plasmon-resonance assays.
- Still unknown: Evidence is biochemical/in-vitro and does not establish therapeutic efficacy or safety. MitoReID's six unseen-drug mechanism predictions are a separate task; the epicatechin COX-2 result is the specific new mechanistic finding recorded here. Full publisher page was cookie-blocked, but the primary PubMed record and publisher-indexed article text/extended-data caption were read.
- AI role: Ranked mechanisms of action for compounds by matching learned time-dependent mitochondrial phenotypes against reference drugs; the model nominated COX-2 inhibition for epicatechin, which was tested in vitro.
Evidence sources: https://pubmed.ncbi.nlm.nih.gov/39169261/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 10% of works of the same type, year and subfield (OpenAlex work metrics, 63 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 2/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A Gaussian naive-Bayes classifier using ECFP6 fingerprints ranked commercial compounds from a 13-million-compound library, followed by docking, property filtering and manual selection. Researchers purchased and assayed 25 candidates; three showed micromolar ULK1 inhibition, with reported IC50 values of 37.11 ± 4.41, 7.67 ± 2.82 and 3.15 ± 0.77 μM. The molecules were purchased rather than newly synthesized, and this entry does not claim absolute chemical novelty. The study’s comparator MRT68921 was more potent; the results do not establish cellular activity or clinical utility.
Technology: Gaussian naive-Bayes classifier on ECFP6 molecular fingerprints (NB-ECFP6)
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: The paper reports 25 purchased candidates tested in a biochemical ULK1 kinase assay; three had measured micromolar inhibition with IC50 values 37.11 ± 4.41, 7.67 ± 2.82 and 3.15 ± 0.77 μM.
- Still unknown: The candidates were purchased, not synthesized. The reported assays do not establish cell activity or clinical effect; the reference MRT68921 was more potent in the paper’s comparison. Absolute structure novelty is not claimed here.
- AI role: The NB-ECFP6 model screened a large commercial compound set and prioritized candidates for downstream docking and biochemical testing.
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC11485869/ · https://pmc.ncbi.nlm.nih.gov/articles/PMC11485869/pdf/IFMC_16_2385288.pdf
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review date: 17 Nov 2026. Review scope: Primary PMC full article and linked publisher-version PDF, including methods, kinase assay results, date and author affiliation block..
Editorial assessment: Impact 2/5 · Validation 3/5 · Novelty 2/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A man with ALS used a brain-computer interface with machine learning to turn attempted speech into text and synthetic voice. After further training, the system maintained 97.5% accuracy; the study reports more than 248 hours of spontaneous conversation over 32 weeks. Result from one participant with an invasive implant; still experimental.
Technology: Neural speech decoding with machine learning
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: One participant with ALS in the BrainGate2 trial completed 84 sessions over 32 weeks with over 248 hours of self-paced use; word accuracy was measured per session, with 97.5% maintained after continued data collection.
- Still unknown: The page does not state the study's limitations, and the result comes from one participant with an invasive implant.
- AI role: Machine learning decoded neural activity into phonemes and then words in real time, with continuous updates, and software generated the voice.
Evidence sources: https://www.ucdavis.edu/news/brain-computer-interface-allows-man-als-speak
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: press release. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 282 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 4/5 · Validation 3/5 · Novelty 4/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
High-throughput calculations and symbolic regression identified mechanical descriptors that classify positive- versus negative-thermal-expansion 2D crystals, then used those descriptors to predict monolayers with near-zero thermal expansion (2D Invar) and unusually large positive or negative expansion (anti-Invar). The paper reports multiple specific predicted crystals and explains the phonon/deformation-mode basis. The materials are computational predictions rather than experimentally synthesized monolayers.
Technology: High-throughput calculations combined with symbolic regression machine learning
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: High-throughput calculations and symbolic regression identified mechanical descriptors that classify positive- versus negative-thermal-expansion 2D crystals, then used those descriptors to predict monolayers with near-zero thermal expansion (2D Invar) and unusually large positive or negative expansion (anti-Invar).
- Still unknown: The identified materials and thermal-expansion properties are from computational/high-throughput predictions and were not experimentally validated in this paper. This is a new 2D monolayer property result, distinct from previously cataloged bulk high-entropy Invar alloys.
- AI role: Learned interpretable relationships between in-plane tensile stiffness, out-of-plane bending stiffness and thermal-expansion class, then used them to screen 2D-crystal databases for predicted Invar and anti-Invar materials.
Evidence sources: https://www.nature.com/articles/s41467-024-51379-6
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: at percentile 81 of works of the same type, year and subfield (OpenAlex work metrics, 17 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Tencent AI Lab and collaborators developed PALM-H3, a pretrained antibody language model that generates heavy-chain CDRH3 sequences, and A2binder to rank antigen-antibody binding. For SARS-CoV-2 XBB, which the paper says was not present during training, the selected generated antibody (Artificial 1) bound spike and neutralized pseudovirus in vitro, with reported KD 0.13 nM and IC50 0.00301 μg/mL. The study also assayed generated antibodies against wild-type, Alpha and Delta; performance varied by antibody and variant.
Technology: PALM-H3 pretrained antibody generative language model and A2binder antigen-antibody affinity predictor
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: For SARS-CoV-2 XBB, which the paper says was not present during training, the selected generated antibody (Artificial 1) bound spike and neutralized pseudovirus in vitro, with reported KD 0.13 nM and IC50 0.00301 μg/mL.
- Still unknown: The functional evidence is in vitro, using SARS-CoV-2 pseudovirus rather than live-virus or clinical outcomes. The study tested selected antibody candidates; several generated candidates performed substantially worse, so the reported top result is not representative of all generated sequences. The stated novelty is experimental neutralization against the XBB variant by a generated antibody, not proof of a treatment or protection in people.
- AI role: PALM-H3 generated CDRH3 sequences conditioned on antigen specificity using a RoFormer-based antibody model; A2binder scored antibody-antigen pairs to select candidates for binding, surface-plasmon-resonance and pseudovirus neutralization tests. The paper reports it generated and selected an XBB spike-binding antibody despite XBB being absent from its training set.
Evidence sources: https://www.nature.com/articles/s41467-024-50903-y
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 80 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A MeerKAT survey of its first four Large Magellanic Cloud pointings reported seven new radio pulsars. The search pipeline applied the Pulsar Image-based Classification System (PICS), a deep-neural image classifier, to folded candidates and retained candidates above a PICS score threshold before human visual classification. The AI narrowed the candidate set; volunteers performed the final candidate classification and follow-up observations supplied additional checks. Six pulsars were detected again in follow-up; the incoherent-beam pulsar was not. A later paper reports 12 discoveries from additional pointings and timing follow-up of some Paper I pulsars, which are not additional discoveries in this entry.
Technology: Pulsar Image-based Classification System (PICS), a deep-neural image classifier for folded candidate plots
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: The survey reports seven new LMC radio pulsars. Folded candidates were screened using PICS scores and then visually classified; six of the seven were detected again in follow-up, while the incoherent-beam pulsar was not confirmed in follow-up.
- Still unknown: PICS filtered candidates but humans classified the candidates as new pulsars. The one incoherent-beam pulsar lacked a follow-up detection in this paper. Later timing work on Paper I objects is not a separate discovery result.
- AI role: PICS applied a pulsar-like score to folded candidate plots and filtered the survey candidate set before human visual classification; the paper does not attribute final discovery confirmation to the model alone.
Evidence sources: https://academic.oup.com/mnras/article/533/3/2570/7731149 · https://arxiv.org/html/2405.12029v1 · https://academic.oup.com/mnras/article/545/4/staf2193/8376644
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review date: 17 Nov 2026. Review scope: Primary OUP full-text indexed article for candidate selection workflow, score threshold, seven-pulsar result, follow-up status, publication date and first-author affiliation; primary arXiv v1 and OUP Paper II for version and overlap context..
Editorial assessment: Impact 2/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
In an acute-coronary-syndrome derivation cohort matched for plaque rupture (PR) versus intact fibrous cap (IFC), quantitative mass spectrometry measured 1,121 plasma proteins. Five machine-learning algorithms screened candidate proteins; APOC3, RAB39A and KNG1 were then significantly different between PR and IFC in an independent 85-patient cohort using ELISA. The three proteins discriminated plaque rupture better than the study’s conventional biomarkers and risk factors, suggesting a candidate biomarker panel for identifying PR among ACS patients.
Technology: Unbiased machine-learning feature screening of quantitative mass-spectrometry plasma proteomics using five algorithms
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: In an acute-coronary-syndrome derivation cohort matched for plaque rupture (PR) versus intact fibrous cap (IFC), quantitative mass spectrometry measured 1,121 plasma proteins.
- Still unknown: This was a small matched derivation cohort (110 ACS patients) and an 85-patient validation cohort, not prospective outcome prediction or a clinical-impact trial. The proteins are candidate diagnostic markers; the study does not show that using them improves management or outcomes, nor establish causal roles in plaque rupture. The publisher reports electronic publication 2024-08-06 and issue date 2024-09-15; the record uses online-first date.
- AI role: The ML algorithms selected nine candidate proteins from the proteomic measurements; the study subsequently tested candidates in an independent ACS cohort and confirmed three analytes—APOC3, RAB39A and KNG1—as significantly different in plaque-rupture versus intact-cap groups. ELISA provided the independent protein validation. The AI-enabled result is this prioritized, externally tested three-protein panel for a specific plaque phenotype.
Evidence sources: https://www.sciencedirect.com/science/article/pii/S0009898124021570
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: at percentile 51 of works of the same type, year and subfield (OpenAlex work metrics, 1 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 2/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Machine-learning screening and expert down-selection identified Ba–Sr cerate oxide compositions for synthesis. The researchers used an automated rapid-sintering and dielectric-characterization platform and reported X-ray diffraction evidence of single-phase solid-solution specimens with refined orthorhombic Pnma structures. The finding is an experimental validation of selected ML-predicted compositions and a process-structure result; it does not establish device performance or broad predictive success.
Technology: Gradient-boosting perovskite classifier; VAE and composition-embedding screening
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Machine-learning screening and expert down-selection identified Ba–Sr cerate oxide compositions for synthesis.
- Still unknown: The paper experimentally reports selected ML-predicted Ba–Sr cerate compositions as single-phase by XRD and refines their structures, while additional tested compositions showed multiple phases; ML prediction alone did not guarantee synthesis. The text is inconsistent about the Sr fraction for one reported specimen (abstract/table/figure versus one results sentence), so this entry avoids assigning that composition. The paper validates selected members of a prior predicted set rather than claiming a universal synthesis predictor. Dielectric analysis is reported, but no application/device performance is established.
- AI role: A gradient-boosting classifier trained on experimental perovskite/non-perovskite structures screened roughly 0.6 million possible compositions, assigning high perovskite likelihood to a candidate set. Variational-autoencoder and element-composition embeddings, plus human literature and synthesis-feasibility review, down-selected compositions that were then synthesized and characterized.
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC11297172/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 10% of works of the same type, year and subfield (OpenAlex work metrics, 41 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 2/5 · Validation 3/5 · Novelty 2/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Researchers applied MODIFY, an ensemble model combining protein language models and sequence-density models, to design a diverse high-fitness library for Rhodothermus marinus cytochrome c. Experimental screening identified variant MLYPPT, which catalyzed both non-natural borylation and silylation: reported total turnover numbers were 2,880 and 2,740, with enantiomeric ratios of 94:6 and 97:3, respectively. The same library yielded distinct high-performing variants for each reaction.
Technology: MODIFY (ML-optimized library design with improved fitness and diversity)
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Experimental screening identified variant MLYPPT, which catalyzed both non-natural borylation and silylation: reported total turnover numbers were 2,880 and 2,740, with enantiomeric ratios of 94:6 and 97:3, respectively.
- Still unknown: The C–B and C–Si reactions had previously been demonstrated with separately evolved cytochrome c variants; this finding is the discovery of additional, structurally distinct variants and a generalist MLYPPT catalyst. The reported yields and enantiomeric ratios are from the study's assays; industrial scale-up was not shown.
- AI role: MODIFY used evolutionary sequence models and protein language models for zero-shot fitness estimates, then co-optimized predicted fitness and residue-level diversity to propose a 1,000-variant library. Researchers synthesized and assayed sampled library members; the model-guided library contained experimentally measured new cytochrome c variants, including MLYPPT with activity in both reactions.
Evidence sources: https://par.nsf.gov/servlets/purl/10529357
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 99 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A pretrained model, ensemble learning and active learning screened 16,050 compositions. The paper reports synthesis of 36 predicted oxides, including 13 pure perovskite structures. Two selected Pr–Sr–Co–Fe(-Mn) materials showed overpotentials of 327 and 315 mV at 10 mA/cm² under the authors’ alkaline three-electrode protocol. These values are assay-specific, not device performance.
Technology: Pretrained model; ensemble learning; active learning
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: The primary article reports the study-specific material and measurement summarized here.
- Still unknown: Independent replication and commercial deployment are not established.
- AI role: Transfer learning and active-learning selection guide candidate synthesis; authors measured OER activity and characterized the catalysts.
Evidence sources: https://www.nature.com/articles/s41467-024-50605-5
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 16 Nov 2026. Review scope: Primary publisher indexed full text/abstract and version metadata reviewed..
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
The two systems solved 4 of 6 problems at IMO 2024, scoring 28 of 42 points, one point below the gold threshold. Problems were manually translated into formal language, and some required up to three days of computation instead of the official 4.5 hours.
Technology: AlphaProof (Lean) and AlphaGeometry 2
Family AlphaGeometry: profile · models page
Publication: Peer-reviewed paper · Test context: prospective or independent check. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: prospective or independent check
- Tested: The systems solved official IMO 2024 problems, graded under IMO rules by Prof. Timothy Gowers and Dr Joseph Myers, reaching 28 of 42 points; problems were manually translated into formal language first.
- Still unknown: Two combinatorics problems remained unsolved, translation was manual, and some solutions took up to three days.
- AI role: AlphaProof solved three problems with formal proofs, and AlphaGeometry 2 proved the geometry problem within 19 seconds of formalization.
Evidence sources: https://deepmind.google/blog/ai-solves-imo-problems-at-silver-medal-level/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: blog. Paper: linked separately.
Review date: 16 Nov 2026. Review scope: Original institutional announcement checked for the specific results, AI contribution and limitations cited in the antibiotic dossier or mathematical follow-up pilot. Full-paper reanalysis, external replication and an exhaustive follow-up search were not performed..
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 146 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 4/5 · Novelty 4/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
- publication · 12 Nov 2025: The updated original announcement dates publication of the AlphaProof methodology in Nature to 12 November 2025. This journal transition is not an independent replication and does not apply to every component of the combined system. Source. Recorded 16 Nov 2026.
Follow-up checked 16 Nov 2026 within scope: Bounded check of the original announcements cited by the events recorded here, including explicit links to later program work. This is not an exhaustive literature search or ongoing monitoring commitment..
Directed relationships from this entry; they do not merge catalog counts.
- 2024-01-alphageometry · extends: AlphaGeometry 2 is explicitly described as an improved AlphaGeometry version; this link concerns the geometry component, not attribution of all four IMO solutions. Source
The IsAb2.0 workflow, combining AlphaFold-Multimer complex modeling with FlexddG mutation scoring, proposed five substitutions to improve the binding of humanized J3 nanobody HuJ3 to HIV-1 gp120. Of these, E44R experimentally improved binding (ELISA EC50 1.696 nM) and neutralization relative to HuJ3; the other proposed substitutions did not consistently improve binding, and T99R failed to express. The gain did not restore neutralization to the parental llama J3 level.
Technology: IsAb2.0 antibody-design protocol integrating AlphaFold-Multimer 2.3/3.0, SnugDock, alanine scanning and FlexddG
Family AlphaFold: profile · models page
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Of these, E44R experimentally improved binding (ELISA EC50 1.696 nM) and neutralization relative to HuJ3; the other proposed substitutions did not consistently improve binding, and T99R failed to express.
- Still unknown: This is a single, previously known J3 nanobody humanization/optimization case, not de novo antibody discovery, animal efficacy or clinical evidence. Only one of five proposed mutants improved binding; several predictions failed and the authors note limitations in prediction accuracy and manual steps. E44R remained less potent than parental J3.
- AI role: The protocol modeled the HuJ3–gp120 complex, nominated interface hotspots and computationally ranked five single-residue mutations. Researchers synthesized/tested variants in binding and HIV-1 neutralization assays; the measured E44R improvement is the experimental outcome, while most predictions did not improve performance.
Evidence sources: https://academic.oup.com/bib/article/25/5/bbae445/7758758
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: at percentile 81 of works of the same type, year and subfield (OpenAlex work metrics, 11 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 2/5 · Validation 3/5 · Novelty 2/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
ESM-IF1 scored substitutions in antibody–antigen complexes, and researchers experimentally screened variants of two clinical antibodies. The journal article reports up to 25-fold neutralization improvement against BQ.1.1 and up to 37-fold affinity improvement against XBB.1.5; these are distinct assays/variants. The preprint abstract reports a different neutralization figure, so this record uses journal values and does not backdate them to preprint v1.
Technology: ESM-IF1 inverse-folding protein model
Family ESM / ESMFold: profile · models page
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: PMC-hosted primary article reports experimental screening and improved neutralization/affinity results for antibody variants.
- Still unknown: Exact v1 result content and sequence-level overlap with earlier group studies not completely established; no clinical efficacy.
- AI role: Structure-informed inverse-folding scores guided substitutions followed by antibody binding and neutralization tests.
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC11616794/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 16 Nov 2026. Review scope: Primary source checked for identity, date, model role and bounded reported result..
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A fused machine-learning and virtual-screening workflow narrowed compounds from three libraries to five candidate FtsZ inhibitors for bioactivity evaluation. The compound T3995 inhibited Staphylococcus aureus ATCC 6538 and Bacillus subtilis ATCC 9732 in reported assays, with MIC values of 32 and 2 μg/mL, respectively. Morphology and FtsZ-polymerization assays supported FtsZ-related antibacterial activity.
Technology: ROC-guided machine-learning selection combined with virtual screening, clustering and binding-model analysis
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: The compound T3995 inhibited Staphylococcus aureus ATCC 6538 and Bacillus subtilis ATCC 9732 in reported assays, with MIC values of 32 and 2 μg/mL, respectively.
- Still unknown: Evidence is early in-vitro antibacterial activity and indirect support for an FtsZ-related mechanism; the abstract does not report direct target-binding measurements or in-vivo efficacy. Hemolysis was reported as safe at or below 16 μg/mL, while the MIC against S. aureus was 32 μg/mL, so this does not establish a therapeutic safety margin. The paper’s 2025 issue date follows its 2024 online publication date; the source-publication date here is the documented first online publication (23 July 2024).
- AI role: ROC-guided ML filtered candidates from three libraries; virtual screening of acquired compounds followed, then clustering and binding-model analysis prioritized five hits. Researchers performed growth, time-kill, cell-morphology, FtsZ-polymerization and hemolysis assays, including testing T3995.
Evidence sources: https://pubmed.ncbi.nlm.nih.gov/39043879/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: at percentile 58 of works of the same type, year and subfield (OpenAlex work metrics, 4 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 2/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
At 1.4° resolution, the model outperformed ECMWF-ENS in 95% of cases for forecasts from 2 to 15 days; at 2.8°, its mean 1980–2020 temperature error was 0.25 °C versus 0.75 °C for atmosphere-only models. It is more than 3,500 times faster than X-SHiELD on a TPU. It models only the atmosphere and is not yet a complete climate model.
Technology: NeuralGCM (differentiable dynamical core and neural networks)
Family NeuralGCM: profile · models page
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: Evaluated with the WeatherBench 2 benchmark against ECMWF-ENS (better in 95% of 2-15 day cases at 1.4 degrees) and against AMIP climate baselines for 1980-2020 temperature (0.25 vs 0.75 degrees C error).
- Still unknown: The model covers only the atmosphere and is not yet a complete climate model; oceans and the carbon cycle are future goals.
- AI role: A neural network learns small-scale physics such as cloud formation, replacing hand-built parameterizations, coupled to a physics-based fluid dynamics solver.
Evidence sources: https://research.google/blog/fast-accurate-climate-modeling-with-neuralgcm/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: blog. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 780 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 4/5 · Validation 2/5 · Novelty 4/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Using Virchow pathology-model embeddings, a pan-cancer detector achieved a specimen-level AUC of 0.95 across nine common and seven rare cancer types. The held-out evaluation included 6,142 specimens from Memorial Sloan Kettering Cancer Center and external institutions. Separate comparisons approached the performance of tissue-specific clinical products and exceeded them on some rare variants. This is diagnostic validation on existing specimens, not evidence that the Virchow detector itself was deployed or improved patient outcomes.
Technology: Virchow (vision transformer foundation model for pathology)
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: A pan-cancer detector built on Virchow embeddings was evaluated on a held-out set of 6,142 specimens from Memorial Sloan Kettering and external institutions, and compared with tissue-specific clinical products.
- Still unknown: The abstract does not establish deployment or patient-outcome benefit; no explicit limitations are stated.
- AI role: The Virchow pathology foundation model provided embeddings for the pan-cancer detector.
Evidence sources: https://www.ebi.ac.uk/europepmc/webservices/rest/search?query=DOI:10.1038/s41591-024-03141-0&resultType=core&format=json
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 898 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 4/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Researchers used regularized machine-learning models to select sparse plasma-protein signatures for 10-year incidence across 218 diseases in 41,931 UK Biobank Pharma Proteomics Project participants. Five to twenty proteins improved held-out risk prediction over basic clinical information for 67 diseases and over clinical information plus routine assays for 52. The study identified TNFRSF17 as a specific predictor for future multiple myeloma and TNFRSF13B for monoclonal gammopathy of undetermined significance. In bone-marrow single-cell data from 11 newly diagnosed myeloma patients and three controls, four of five myeloma predictor proteins were most highly expressed by plasma cells; FCRLB and QPCT elevation was attributable to malignant plasma cells. Proteomic signatures also replicated improved prediction for five of six diseases tested in EPIC-Norfolk.
Technology: LASSO feature selection and regularized Cox machine-learning models integrating plasma proteomics and clinical covariates
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: Five to twenty proteins improved held-out risk prediction over basic clinical information for 67 diseases and over clinical information plus routine assays for 52.
- Still unknown: The disease predictions are observational risk associations, not causal mechanisms or clinical screening outcomes. Most protein signatures were not externally replicated; EPIC-Norfolk tested six diseases, with improved prediction replicated for five. The myeloma single-cell analysis involved 11 patients and three healthy controls and supports cellular expression attribution, not prospective validation of an assay. Absolute quantification with clinical-grade assays and further validation would be needed before clinical use.
- AI role: The authors applied a three-stage machine-learning framework—feature selection, hyperparameter tuning and held-out validation—to ~2,923 proteins in UK Biobank. It selected parsimonious protein signatures for 218 future disease outcomes and identified disease-specific predictors and shared predictors. For multiple myeloma, the model-selected five-protein signature and follow-up single-cell expression analysis linked four predictors to marrow plasma cells and showed malignant-cell-associated FCRLB and QPCT expression.
Evidence sources: https://www.nature.com/articles/s41591-024-03142-z
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 273 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
The polyVERSE workflow used reaction-aware polymer generation and machine-learning property screens to identify PONB-2Me5Cl, which the researchers synthesized and cast into films. Experimental characterization measured a discharged energy density of 8.3 J cm⁻³ at 200 °C and a breakdown field of 750 MV m⁻¹; the paper compares this energy density with commercial polymer dielectrics at the same temperature. This is a laboratory polymer-film result, not a packaged capacitor or aerospace-ready device.
Technology: polyVERSE with four polymer graph neural-network property predictors
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: The polyVERSE workflow used reaction-aware polymer generation and machine-learning property screens to identify PONB-2Me5Cl, which the researchers synthesized and cast into films.
- Still unknown: The reported 8.3 J cm⁻³ is measured on synthesized polymer films under the article’s laboratory breakdown and dielectric characterization protocol. The comparison with commercial polymers is the authors’ reported comparison; the work does not demonstrate a packaged capacitor, operational lifetime, aerospace use, or scale-up. Other suggested polyimides in the paper remain predicted and are not counted here. The reported polymerization used non-green solvents.
- AI role: polyVERSE generated polymers from commercially available monomers using encoded reaction templates, then machine-learning models predicted glass-transition temperature, band gap and dielectric properties for screening. AI down-selected polymer candidates for synthesis; PONB-2Me5Cl was subsequently synthesized and its dielectric properties were measured.
Evidence sources: https://www.nature.com/articles/s41467-024-50413-x.pdf
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 120 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Researchers at the University of Greifswald resolved how a post-translational modification regulates Bacillus subtilis acetyl-CoA synthetase, a central metabolic enzyme. The study used AI-based structural prediction to explain the molecular mechanism at near-atomic detail.
Technology: AlphaFold2 and AI-assisted structural analysis
Family AlphaFold: profile · models page
Publication: Peer-reviewed paper · Test context: Test context unspecified in annotation. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: not reported
- Tested: Researchers at the University of Greifswald resolved how a post-translational modification regulates Bacillus subtilis acetyl-CoA synthetase, a central metabolic enzyme.
- Still unknown: The institutional announcement says the regulatory mechanism is evolutionarily conserved and may matter in higher organisms; that implication is not itself a demonstration in those organisms.
- AI role: AlphaFold2 structural predictions supported the researchers' mechanistic interpretation of how the enzyme is regulated; the source describes a collaboration with biochemical and structural experiments, not an AI-only result.
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC11255334/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 10% of works of the same type, year and subfield (OpenAlex work metrics, 27 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 1/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
In rheo-XPCS measurements of a sheared silica-nanoparticle glass, an unsupervised convolutional autoencoder and K-means clustering grouped two-time correlation patterns by relative relaxation rate. Changes in the frequencies of slow-dynamics classes tracked non-monotonic shear-stress relaxation, while some later local slow-dynamics episodes appeared without a corresponding change in the bulk stress signal. This is a measured microscopic-to-macroscopic relationship in the study’s model colloidal system.
Technology: AI-NERD: convolutional autoencoder for XPCS correlation images with K-means clustering in the learned latent space
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: In rheo-XPCS measurements of a sheared silica-nanoparticle glass, an unsupervised convolutional autoencoder and K-means clustering grouped two-time correlation patterns by relative relaxation rate.
- Still unknown: The findings come from rheo-XPCS of a model concentrated silica-nanoparticle suspension under the reported shear/relaxation protocol. Cluster labels encode relative relaxation rates and latent-space distances are not quantitative physical measures. The authors describe the stress/dynamics relationship qualitatively and say further work is needed to extract quantitative physical information. A publisher correction dated 2024-09-09 corrected corresponding-author information; it did not report a change to the results.
- AI role: The unsupervised neural autoencoder compressed experimental two-time X-ray photon correlation functions into a latent representation, then K-means grouped frames by dynamical similarity. Class-transition frequencies over experimental time let researchers compare local relaxation patterns with simultaneous rheometer stress measurements, revealing where local slow dynamics coincided with stress minima and where they did not.
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC11251071/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: at percentile 85 of works of the same type, year and subfield (OpenAlex work metrics, 23 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 2/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Machine-learning analysis of untargeted serum metabolomic and lipidomic profiles from clinical pig cohorts identified lysophosphatidic acid (LPA) as a PRRSV-associated biomarker. Follow-up infection experiments reported increased LPA with autotaxin upregulation; in porcine alveolar macrophages, LPA manipulation altered RIG-I/type-I-interferon responses and viral replication. In infected pigs, targeting LPA was reported to protect against infection and associated disease outcomes.
Technology: Random-forest analysis of serum metabolomic and lipidomic profiles
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Machine-learning analysis of untargeted serum metabolomic and lipidomic profiles from clinical pig cohorts identified lysophosphatidic acid (LPA) as a PRRSV-associated biomarker.
- Still unknown: The clinical classifier and biomarker work concerns pigs, not human diagnosis. The infection challenge experiments were small (figure descriptions report five piglets per group), and cell assays were performed in vitro. The authors report that the mechanism by which LPA suppresses RIG-I remains unclear; the findings do not establish a treatment for people or broad field effectiveness in livestock.
- AI role: Random-forest analysis ranked metabolite and lipid features from clinical PRRSV-positive and negative pig cohorts and contributed to identification of LPA as the candidate biomarker. The authors then tested LPA in infection models, including cultured porcine alveolar macrophages and infected pigs, to examine host-response and disease outcomes.
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC11425916/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 10% of works of the same type, year and subfield (OpenAlex work metrics, 14 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A machine-learning classifier trained on trace-element fingerprints identified detrital zircon signatures from sediment-derived S-type granites as old as 4.24 billion years in the Jack Hills record. Across global detrital zircon data, the authors report long-term cyclic changes in S-type granite proportions and infer that exposed, weathered continental crust and subduction operated by about 4.2 Ga. The result bears on early Earth geodynamics, not a direct fossil or biological discovery.
Technology: Trace-element Support Vector Machine (TSVM) classifier
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: A machine-learning classifier trained on trace-element fingerprints identified detrital zircon signatures from sediment-derived S-type granites as old as 4.24 billion years in the Jack Hills record.
- Still unknown: The early-subduction interpretation is inferred from zircon trace elements and remains one interpretation of sparse Hadean geological evidence. The article reports that only the combination of early subaerial weathering and subduction explains the prevalence of the inferred S-type granites; it does not directly observe tectonics at 4.2 Ga.
- AI role: Classified zircon as S-type or non-S-type from multielement geochemical fingerprints, allowing the authors to identify the oldest sediment-derived granite signal and quantify its temporal distribution across detrital-zircon records.
Evidence sources: https://www.pnas.org/doi/10.1073/pnas.2405160121
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 10% of works of the same type, year and subfield (OpenAlex work metrics, 25 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A machine-learning-guided peptide optimization workflow selected GUB021794, reported at 0.018 nM GLP-1R potency in the paper assay and dose-dependent weight loss in diet-induced obese mice. These preclinical results do not establish human efficacy, safety or a dosing regimen.
Technology: streaMLine machine-learning QSAR analysis integrated with peptide synthesis and screening
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Primary abstract reports the named candidate and mouse outcomes; no human clinical result is established.
- Still unknown: Independent replication and exhaustive prior-art review are not established.
- AI role: streaMLine machine-learning QSAR analysis integrated with peptide synthesis and screening
Evidence sources: https://pubmed.ncbi.nlm.nih.gov/38977267/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 16 Nov 2026. Review scope: Primary source checked for study identity, AI operation, dates, affiliations and the bounded result. PubMed affiliation block names Gubra..
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A multimodal AI model analyzed demographics, personal and family history, medications, cognitive and functional testing, and neuroimaging across nine datasets totaling 51,269 participants. It assigned probabilities to ten primary or contributing dementia etiologies and addressed mixed cases. The paper reports microaveraged AUROC 0.96 for dementia-etiology differentiation; model probabilities aligned with etiology-specific biomarkers and postmortem pathology. In a review of 100 cases, AI-augmented neurologist assessment outperformed neurologist-only ratings.
Technology: Multimodal machine-learning model combining clinical, neuropsychological, functional and neuroimaging data
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: A multimodal AI model analyzed demographics, personal and family history, medications, cognitive and functional testing, and neuroimaging across nine datasets totaling 51,269 participants.
- Still unknown: The nine-dataset analysis is retrospective and includes in-house as well as external cohorts. The 100-case clinician comparison is small, and the study did not test prospective changes in care or patient outcomes. The authors call for prospective studies before claiming improved patient care. Model probabilities aligned with available biomarkers and postmortem evidence but are not histopathologic confirmation for every individual prediction.
- AI role: The model integrated routinely collected clinical variables and multimodal neuroimaging to estimate cognitive status and probabilities for ten dementia etiologies, including mixed etiologies and cases with incomplete features. It was evaluated across independent cohorts and compared with neurologist assessments; its output is the etiology-specific classification and probability result.
Evidence sources: https://pubmed.ncbi.nlm.nih.gov/38965435/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 209 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 2/5 · Novelty 2/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
FRASE-bot used a neural-network nativeness score to triage protein–ligand fragments, then used the retained structural features to seed virtual screening for Calcium and Integrin Binding Protein 1 (CIB1). Of 56 purchased virtual-screen hits, 29 showed dose-dependent CIB1 displacement; compound UNC10245380 (UNC5380) had a TR-FRET displacement IC50 of 8.0 ± 2.5 μM and selectively killed CIB1-depletion-sensitive triple-negative breast cancer cells in the reported assays, with no detected killing in an insensitive line.
Technology: FRASE-bot: fragment-in-structural-environment screening with a neural-network nativeness scoring model
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Of 56 purchased virtual-screen hits, 29 showed dose-dependent CIB1 displacement; compound UNC10245380 (UNC5380) had a TR-FRET displacement IC50 of 8.0 ± 2.5 μM and selectively killed CIB1-depletion-sensitive triple-negative breast cancer cells in the reported assays, with no detected killing in an insensitive line.
- Still unknown: The authors describe UNC10245380 as a starting chemical probe/lead, not a drug. Binding was inferred from peptide displacement in TR-FRET and a docking pose; the pose is not a solved small-molecule complex. Selective cell killing was in vitro at 30 μM, with no animal or clinical evidence. The earlier Research Square manuscript is the same underlying paper/result and is not counted separately.
- AI role: The neural network triaged fragments by predicted native-ligand pose likelihood; selected fragments informed pharmacophore construction and virtual screening. The resulting purchased compounds were then tested in TR-FRET binding-displacement and cell-death assays. The model did not perform the wet-lab validation.
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC11219766/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: at percentile 86 of works of the same type, year and subfield (OpenAlex work metrics, 14 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 2/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A multicenter observational serum multi-omics study used LASSO and Boruta to select candidate biomarkers for 28-day outcome in acute respiratory distress syndrome. Eight proteins (VCAM1, LDHB, MSN, FLG2, TAGLN2, LMNA, MBL2 and LBP) showed consistency in an independent cohort of 183 patients and were confirmed by targeted parallel-reaction-monitoring mass spectrometry. The protein model AUC was 0.802 in validation versus 0.738 for the clinical model. The discovery analysis also reported sphingosine-1-phosphate association with ARDS prognosis and an enriched sphingolipid pathway.
Technology: LASSO and Boruta feature selection; machine-learning prognostic model; targeted PRM validation
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Eight proteins (VCAM1, LDHB, MSN, FLG2, TAGLN2, LMNA, MBL2 and LBP) showed consistency in an independent cohort of 183 patients and were confirmed by targeted parallel-reaction-monitoring mass spectrometry.
- Still unknown: Observational prognostic associations do not show proteins cause outcomes or that biomarker-guided care improves survival. Discovery n=130 ARDS patients; validation n=183. PRM confirmation is not independent replication by another group; deployment and prospective benefit were not established.
- AI role: LASSO, fold-change screening and Boruta prioritized candidate 28-day ARDS prognostic proteins; candidates were tested by targeted PRM in an external cohort.
Evidence sources: https://pubmed.ncbi.nlm.nih.gov/38956604/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 41 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Learned models guided ink formulation and printing settings for known BiSbTe-based material. The resulting printed samples reached a reported room-temperature zT of 1.3. This is an experimental processing result, rather than discovery of a new chemical composition or demonstrated deployed-device efficiency. The result was public in the July 2024 preprint before the journal version.
Technology: Gaussian-process regression and constrained Bayesian optimization
Publication: Preprint · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Preprint
- Test type: direct test
- Tested: Printed samples were characterized after model-guided processing.
- Still unknown: Generality, industrial manufacturing and the contribution of each downstream processing step remain unestablished in this review.
- AI role: Gaussian-process regression and constrained Bayesian optimization supported the stated study; experimental and interpretive work remained with researchers.
Evidence sources: https://arxiv.org/abs/2407.01145 · https://arxiv.org/pdf/2407.01145
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review date: 16 Nov 2026. Review scope: Primary preprint abstract, submission history and selected PDF passages on model and affiliations..
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A model trained on simulated galaxy observations inferred accreted stellar fractions in nearby MaNGA galaxies. Predictions suggest in-situ formation dominates most of the studied mass range, while accreted fractions rise in the most massive systems and vary with galaxy characteristics. These are simulation-dependent inferences, not direct measurements of each star's formation history. The result preceded the July 2024 journal article.
Technology: MaNGIA/TNG50-trained stellar-mass inference
Publication: Preprint · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Preprint
- Test type: retrospective or benchmark
- Tested: Model inference was evaluated with simulated observations and applied to real survey data.
- Still unknown: Observational ground truth and transfer beyond the training simulation remain unresolved.
- AI role: MaNGIA/TNG50-trained stellar-mass inference supported the stated study; experimental and interpretive work remained with researchers.
Evidence sources: https://arxiv.org/abs/2407.00166v1 · https://arxiv.org/html/2407.00166v1
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review date: 16 Nov 2026. Review scope: Primary v1 abstract, history, affiliations and selected HTML Methods; supplements not reviewed..
Editorial assessment: Impact 3/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A convolutional neural network shortlisted cataclysmic-variable candidates from more than two million SDSS-V spectra, reducing the spectra requiring visual inspection to about 2% of the total. Human review and follow-up analysis of the resulting 504-object sample reported 61 previously unreported cataclysmic variables, spectroscopically confirmed 248 candidates and yielded 82 new or improved orbital periods; the article discusses selection biases and the CNN missed 27 spectra later found by human review.
Technology: Convolutional neural network classifier for SDSS-V/BOSS spectra
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: A convolutional neural network shortlisted cataclysmic-variable candidates from more than two million SDSS-V spectra, reducing the spectra requiring visual inspection to about 2% of the total.
- Still unknown: This is one grouped paper-level survey result, not 61 separate records. The CNN was a triage aid rather than an autonomous confirmation system; authors report completeness/selection limitations and 27 misclassified spectra discovered through human inspection.
- AI role: Ranked over two million spectra for likely cataclysmic variables, reducing the manual-inspection workload; human investigators visually inspected candidate spectra and established source classifications and periods.
Evidence sources: https://arxiv.org/abs/2406.19459
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: at percentile 77 of works of the same type, year and subfield (OpenAlex work metrics, 13 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
ESM3 designed esmGFP, a fluorescent protein of 229 amino acids with 96 mutations relative to the closest known protein (about 58% similarity). In a second round of 96 designs, some proteins glowed as brightly as natural GFPs. The comparison with ‘more than 500 million years of evolution’ is an analogy-based estimate.
Technology: ESM3
Family ESM / ESMFold: profile · models page
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: In two rounds of 96 designs, lab measurements of fluorescence found some proteins as bright as natural GFPs; esmGFP differs by 96 mutations and is about 58% similar to the closest natural protein.
- Still unknown: The page lists no formal limitations; the designed B8 was about 50 times dimmer and slower to mature than natural GFPs.
- AI role: ESM3 generated new GFP candidates prompted with structural residues from natural GFP.
Evidence sources: https://www.evolutionaryscale.ai/blog/esm3-release
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: blog. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 4/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Hybrid CALPHAD/ICME calculations and Bayesian optimization selected an Al-Er-Zr-Y-Yb-Ni composition that was manufactured as powder and printed by laser powder bed fusion without the cracking reported for Al 7075. The preprint reports 50% greater strength after 400 °C aging than the strongest benchmark printable aluminum alloy. This is a measured material result, not component-scale qualification.
Technology: Hybrid CALPHAD/ICME; neural-network regression; Gaussian-process Bayesian optimization
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: The source reports the study-specific result summarized here.
- Still unknown: Independent replication and broader deployment or clinical benefit are not established by this source review.
- AI role: Simulation-derived features train a forward model and Bayesian optimization proposes compositions; the selected alloy is manufactured and mechanically tested.
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC12810664/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review date: 16 Nov 2026. Review scope: Primary-source abstract and publication metadata reviewed; full text scope is specified in source_checks..
Editorial assessment: Impact 4/5 · Validation 3/5 · Novelty 4/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A deep-learning pipeline based on inversion of AlphaFold2 and ProteinMPNN designed soluble proteins with claudin, rhomboid-protease and GPCR-like folds. Experiments reported stable, folded soluble designs and structural agreement with target folds. The researchers grafted functional motifs, including a ghrelin-receptor epitope that bound its antibody and active/inactive A2A-receptor-like designs that differed in mini-Gs binding; a binding-site knockout served as a control.
Technology: AF2seq deep-learning backbone design with ProteinMPNN sequence design
Family ProteinMPNN: profile · models page
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Experiments reported stable, folded soluble designs and structural agreement with target folds.
- Still unknown: This is a protein-design proof of concept with in-vitro biophysical and binding assays; the analogues are not native membrane receptors and do not demonstrate signaling, drug activity or therapeutic efficacy. The GPCR analogues tested binding to mini-Gs in conformation-designed constructs. The separate live-catalog record on designed proteins modulating dopamine D1 receptor signaling reports a different binder/target and signaling result; this paper reports soluble receptor-fold analogues and binding assays, not D1 signaling modulation.
- AI role: The deep-learning design pipeline generated soluble analogues of complex membrane-protein folds and the scaffolds into which native functional motifs were grafted. The study tested solubility, thermal stability, structural accuracy and selected binding functions experimentally, including conformation-specific A2A-receptor-like constructs.
Evidence sources: https://www.nature.com/articles/s41586-024-07601-y
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 184 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Researchers developed a targeted multiplex mass-spectrometry assay for eight plasma proteins and applied machine learning to distinguish Parkinson disease from healthy controls and classify people with isolated REM sleep behavior disorder (iRBD). In the reported longitudinal iRBD cohorts, the model classified 79% of participants up to seven years before motor Parkinson disease or dementia with Lewy bodies onset. The work connects a targeted protein panel to a pre-motor disease signal in measured blood samples.
Technology: Machine-learning classifier using recursive feature elimination on targeted plasma proteomic measurements
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: Researchers developed a targeted multiplex mass-spectrometry assay for eight plasma proteins and applied machine learning to distinguish Parkinson disease from healthy controls and classify people with isolated REM sleep behavior disorder (iRBD).
- Still unknown: The discovery phase used only 10 Parkinson cases and 10 controls. Although the study includes larger targeted-assay cohorts, the longitudinal iRBD groups were small (18 and 54 participants), and the finding is not a validated screening test or proof that intervention changes outcomes. The model’s apparent early classification needs larger prospective cohorts.
- AI role: The classifier selected an eight-protein panel from targeted mass-spectrometry measurements, then classified motor Parkinson cases and iRBD participants; the longitudinal analysis linked panel scores with later clinical conversion.
Evidence sources: https://www.nature.com/articles/s41467-024-48961-3
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 139 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Untargeted serum metabolomics from 69 participants, including eight patients followed from HBV-related cirrhosis to hepatocellular carcinoma (HCC), was analyzed with LASSO and random forest to prioritize metabolites. 1-methylnicotinamide (1-MNAM) declined across healthy, cirrhosis and HCC groups and was lower at the HCC than cirrhosis stage in the eight progressors. Targeted mass spectrometry in a separate cohort of 25 healthy controls and 25 early-stage HCC patients confirmed lower 1-MNAM in HCC (AUC 0.954 for HCC versus healthy controls).
Technology: LASSO regression and random forest classifier applied to serum metabolomics
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Targeted mass spectrometry in a separate cohort of 25 healthy controls and 25 early-stage HCC patients confirmed lower 1-MNAM in HCC (AUC 0.954 for HCC versus healthy controls).
- Still unknown: The longitudinal transition signal comes from eight HBV-cirrhosis patients who progressed during follow-up. The independent targeted cohort compared 25 early-HCC patients with 25 healthy controls; it did not validate prediction of progression among cirrhosis patients. The authors describe the overall sample as small and call for larger prospective validation. 1-MNAM is a known endogenous metabolite; this record concerns the study-specific association and selection result, not discovery of a new compound, causal mechanism, or clinical test.
- AI role: The authors used LASSO and random forest to select ten candidate metabolites distinguishing cirrhosis from HCC, then identified 1-MNAM as the sole candidate with a consistent decline across healthy, cirrhosis and HCC groups. The algorithms prioritized the measured metabolite; targeted mass spectrometry in a separate cohort provided assay follow-up.
Evidence sources: https://link.springer.com/article/10.1007/s00432-024-05848-6
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: at percentile 63 of works of the same type, year and subfield (OpenAlex work metrics, 7 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 2/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
An automated machine-learning workflow screened public blood-methylation datasets and selected IGF2BP1, CENPI and PSME4 as candidate schizophrenia-associated methylation markers. The researchers then measured these genes by targeted qMSP in 30 first-episode, drug-naive schizophrenia patients and 30 controls: IGF2BP1 methylation was higher and PSME4 lower in cases, while CENPI did not differ. A five-feature model combining the three genes with age and sex classified cases versus controls with AUC 0.755 (95% CI 0.636–0.862).
Technology: JADBio AutoML platform, comparing classifiers including support-vector machines, random forests, ridge logistic regression and decision trees with feature selection and cross-validation
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: The researchers then measured these genes by targeted qMSP in 30 first-episode, drug-naive schizophrenia patients and 30 controls: IGF2BP1 methylation was higher and PSME4 lower in cases, while CENPI did not differ.
- Still unknown: The patient validation was a small case-control pilot (30 first-episode cases and 30 healthy controls); its AUC of 0.755 is modest and has a wide confidence interval. CENPI methylation did not differ in the clinical samples, so it should not be presented as independently validated. The study does not establish diagnostic performance in clinical differential-diagnosis populations, prospective utility or causality, and the authors describe the signature as potentially useful for future diagnostics rather than clinically established.
- AI role: AutoML selected gene-methylation features from genome-wide public schizophrenia methylome datasets and built predictive signatures. The authors then measured the selected markers using qMSP in a separate 30-case/30-control first-episode drug-naive cohort and trained a second model on methylation plus demographic variables.
Evidence sources: https://www.nature.com/articles/s41398-024-02946-4.pdf
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: at percentile 78 of works of the same type, year and subfield (OpenAlex work metrics, 12 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 2/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
In a reader study with 62 radiologists on 400 cases, the AI detected about 6.8% more clinically significant cancers at the same specificity and produced about 50% fewer false positives. On 1,000 routine-practice cases, however, non-inferiority was not confirmed (specificity 68.9% vs 69.0% at 96.1% sensitivity). The authors call for prospective validation, and the system is not available for clinical use.
Technology: PI-CAI challenge super-algorithm (ensemble of top five submitted deep learning models)
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: A multireader study with 62 radiologists on 400 paired cases compared AI and PI-RADS 2.1 reads, and 1,000 routine-practice cases were used to test non-inferiority against historical clinical readings.
- Still unknown: Non-inferiority to routine practice was not confirmed (specificity 68.9% vs 69.0%); the authors state prospective validation is needed and the system is not available for clinical use.
- AI role: A deep-learning system detected clinically significant prostate cancer on MRI.
Evidence sources: https://www.ebi.ac.uk/europepmc/webservices/rest/search?query=EXT_ID:38876123%20AND%20SRC:MED&resultType=core&format=json
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 418 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 4/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Gaussian-process models and chemist decisions guided synthesis and testing of organic photocatalysts for a specified cross-coupling reaction. A catalyst search reached 67% yield before reaction-condition optimization. This establishes a bounded formulation result for the studied reaction, not discovery of the reaction or general superiority to iridium catalysts.
Technology: Gaussian-process Bayesian optimization
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Candidate catalysts were synthesized and reaction yields measured in repeated experiments.
- Still unknown: General substrate scope, independent replication and exhaustive molecular prior art were not assessed.
- AI role: Gaussian-process Bayesian optimization supported the stated study; experimental and interpretive work remained with researchers.
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC11321994/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review date: 16 Nov 2026. Review scope: Primary article abstract, selected Results, computational Methods and publication metadata; supplements not reviewed..
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
About 470 calls from 101 callers to 117 recipients in Kenya: a model trained only on acoustic structure identifies the intended recipient, and playback experiments show elephants respond more to calls addressed to them. The model confirms a previously suspected pattern; more data are needed to isolate the ‘names’.
Technology: Acoustic classification model
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: About 470 calls from 101 callers to 117 receivers were analyzed, and playback experiments showed elephants respond more to calls addressed to them.
- Still unknown: More data are needed to isolate the names, it is unknown whether elephants name other things, and naming was not universal.
- AI role: A model trained only on acoustic structure identified the intended recipient of each call, supporting the name-like finding.
Evidence sources: https://newsmediarelations.colostate.edu/2024/06/10/elephants-have-names-for-each-other-like-people-do-new-study-shows/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: press release. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 77 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 4/5 · Validation 3/5 · Novelty 4/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A semi-supervised PhaseNet-TF workflow detected and relocated 13,616 events from a one-year Tonga OBS and land-station record, about 11 times the manually picked reference catalog. The denser catalog delineates the intermediate-depth double seismic zone to about 300 km and indicates that its lower plane is limited to roughly 19°–21°S, rather than continuous along the studied section. The study did not recover a deeper 350–460 km double zone; sparse events and location uncertainties limit inference at those depths.
Technology: PhaseNet-TF deep-learning phase picker with semi-supervised iterative retraining
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: A semi-supervised PhaseNet-TF workflow detected and relocated 13,616 events from a one-year Tonga OBS and land-station record, about 11 times the manually picked reference catalog.
- Still unknown: The catalog is based on one year of OBS and land data. About half of events are outside the array and have larger hypocenter uncertainty from azimuthal gaps; velocity-model choice adds absolute-location uncertainty. The authors say the absent 350–460 km double zone may reflect insufficient events, not definitive absence. Stress and hydration explanations are interpretations, and some rare events above the slab may be false.
- AI role: The time-frequency neural picker detected P/S arrivals in noisy ocean-bottom seismometer data; iterative semi-supervised retraining expanded picks, while phase association and double-difference relocation filtered unreliable events. This expanded catalog enabled the localized double-seismic-zone inference.
Evidence sources: https://academic.oup.com/gji/article/238/2/1073/7689220
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: at percentile 88 of works of the same type, year and subfield (OpenAlex work metrics, 19 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Machine learning identified 863,498 candidate sequences in microbial genomes and metagenomes. Of 100 synthesized peptides, 63 inhibited at least one of the eleven strains tested; some showed activity in animal models. The catalog contains candidates, not hundreds of thousands of clinically validated antibiotics.
Technology: Machine learning for antimicrobial peptides
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: 100 predicted peptides were synthesized and tested in vitro and in vivo against drug-resistant pathogens and human gut commensals; 63 inhibited pathogens.
- Still unknown: The abstract states no limitations and does not name the strains or animal models used.
- AI role: Machine learning predicted antimicrobial peptides across 63,410 metagenomes and 87,920 genomes, producing a catalog of 863,498 non-redundant candidates.
Evidence sources: https://www.ebi.ac.uk/europepmc/webservices/rest/search?query=DOI:10.1016/j.cell.2024.05.013&resultType=core&format=json
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: press release. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 477 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 4/5 · Validation 3/5 · Novelty 4/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A structure-based machine-learning mutation-ranking method, combined with molecular dynamics and human rational design, guided variant selection for a lysine dioxygenase. The authors report a five-substitution variant with total turnover number 218 ± 3, compared with 160 ± 2 for wild type in the reported assay. The improvement is specific to this enzyme and assay, not a claim of industrial or therapeutic performance. The source uses conflicting variant abbreviations: its abstract says LPNYI and its Results say LPYNI.
Technology: MutComputeX; molecular dynamics
Publication: Preprint · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Preprint
- Test type: direct test
- Tested: Primary abstract reports the designed pentamutant’s TTN and wild-type comparator.
- Still unknown: Independent replication and broader application are not established by this bounded review. The source uses conflicting abbreviations for the five-substitution variant.
- AI role: Structure-based ML prioritized amino-acid substitutions; MD and rational selection also contributed to the final combined variant.
Evidence sources: https://pubmed.ncbi.nlm.nih.gov/38895203/ · https://www.biorxiv.org/content/10.1101/2024.06.04.597480v1
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review date: 17 Nov 2026. Review scope: Primary preprint version and journal methods/results; the variant abbreviation differs between abstract and Results..
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A 1.3-billion-parameter model for high-resolution weather and air-pollution forecasting: it matches or exceeds CAMS on 74% of targets and matches or beats GraphCast on 94%. The roughly 5,000-fold speedup over the IFS system is an estimate; GraphCast remains slightly better up to five days.
Technology: Aurora (foundation model)
Family Aurora: profile · models page
Publication: Publication status unspecified in annotation · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Not reported
- Test type: retrospective or benchmark
- Tested: Forecasts were compared with CAMS for air pollution (Jun-Nov 2022, matched or beat 74% of targets) and with GraphCast (matched or beat 94% of targets) over 2022 evaluation periods.
- Still unknown: GraphCast was slightly better on specific humidity up to five days, and the page has no dedicated limitations section.
- AI role: A 1.3-billion-parameter foundation model, pretrained on over a million hours of simulations, forecasts weather and air pollution at 0.1 degrees.
Evidence sources: https://www.microsoft.com/en-us/research/blog/introducing-aurora-the-first-large-scale-foundation-model-of-the-atmosphere/
The legacy label “Not reported” does not distinguish information absent from reviewed material from information not assessed. It is not a scientific finding of absence.
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: blog. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Editorial assessment: Impact 3/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Researchers used AlphaFold2 to predict structures for 1,483 cytidine deaminases, clustered the predictions and experimentally characterized representatives. The screen identified previously uncharacterized enzymes with high editing efficiency, and CD0208^P52A showed little sequence-context preference across AC/GC/TC/CC motifs, a narrow editing window and relatively low off-target activity. In mammalian cells, its base editor introduced stop codons in single- and multicopy genes without DNA double-strand breaks.
Technology: AlphaFold2 protein-structure prediction, followed by structural clustering of candidate cytidine deaminases
Family AlphaFold: profile · models page
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Researchers used AlphaFold2 to predict structures for 1,483 cytidine deaminases, clustered the predictions and experimentally characterized representatives.
- Still unknown: The base-editing experiments were preclinical cell experiments, not animal or clinical efficacy tests. AlphaFold2 supplied predicted rather than experimentally solved structures. The paper reports multiple promising deaminases; CD0208^P52A is the specific context-independent editor highlighted here. No therapeutic safety or in-vivo delivery claim is implied.
- AI role: AlphaFold2-predicted structures at scale were used to mine and cluster 1,483 putative deaminases; representatives were then selected for experimental characterization in cytosine base editors. The model enabled structure-guided candidate selection; editing outcomes were measured experimentally in mammalian cells.
Evidence sources: https://www.nature.com/articles/s41551-024-01220-8.pdf
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 10% of works of the same type, year and subfield (OpenAlex work metrics, 37 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A deep-learning device analyzed routine H&E whole-slide images to estimate whether advanced urothelial tumors carried selected FGFR2/FGFR3 alterations used for erdafitinib eligibility. In a separate 350-slide retrospective validation set, the device achieved 88.7% sensitivity and 31.8% specificity; it was also prospectively integrated into a non-interventional study spanning 89 sites in nine countries. The paper projected a 28.7% reduction in molecular tests assuming 15% biomarker prevalence; that estimate was simulated, not a measured population-level reduction.
Technology: FGFR Device, a multi-instance-learning attention model with a convolutional neural-network backbone
Publication: Peer-reviewed paper · Test context: prospective or independent check. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: prospective or independent check
- Tested: In a separate 350-slide retrospective validation set, the device achieved 88.7% sensitivity and 31.8% specificity; it was also prospectively integrated into a non-interventional study spanning 89 sites in nine countries.
- Still unknown: The 350-slide retrospective validation set was deliberately enriched (150 FGFR-positive and 200 negative; 43% prevalence), and the device fell slightly short of the prespecified 90% sensitivity target. Prospective real-time workflow validation covered only 17 samples in one month (15 predictions; two quality-control failures); production deployment reported 24 samples, with only 20 receiving molecular testing. The 28.7% test reduction is a simulation assuming 15% prevalence, not an observed clinical outcome. This is prescreening support, not a stand-alone diagnosis or evidence of patient benefit.
- AI role: Trained on more than 3,000 urothelial-cancer slides with molecular test labels, the deep-learning model inferred selected FGFR alteration status from H&E images and returned a likelihood for prescreening before confirmatory molecular testing. The device was prospectively integrated into a non-interventional study workflow; molecular tests remained the reference standard.
Evidence sources: https://www.nature.com/articles/s41467-024-49153-9.pdf
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 10% of works of the same type, year and subfield (OpenAlex work metrics, 30 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 4/5 · Novelty 2/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A translational study combined a disease-related mouse gene-expression signature, human liver-biopsy transcriptomes and serum protein measurements to select IGFBP7, SSc5D and Sema4D as a three-protein panel for metabolic dysfunction-associated steatotic liver disease (MASLD). A LightGBM model classified biopsy-staged fibrosis, with AUCs 0.82 for F0/F1, 0.89 for F2 and 0.87 for F3/F4 in the development/testing data. In an independent Danish cohort of 156 patients, AUC was 0.84 for F0/F1 and F3/F4; performance for F2 was lower (AUC 0.62).
Technology: LightGBM classifier (TLM3) using the three serum proteins IGFBP7, SSc5D and Sema4D
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: A translational study combined a disease-related mouse gene-expression signature, human liver-biopsy transcriptomes and serum protein measurements to select IGFBP7, SSc5D and Sema4D as a three-protein panel for metabolic dysfunction-associated steatotic liver disease (MASLD).
- Still unknown: The model is an early translational biomarker result. Its independent cohort replicated better classification at F0/F1 and F3/F4 (AUC 0.84) but was modest for F2 (AUC 0.62). The authors say mechanistic claims could not be tested in the single-biopsy cohorts and call for prospective and intervention studies before clinical utility is established. TNO disclosed a patent application related to the biomarkers.
- AI role: The LightGBM model combined the three protein measurements selected from a translational mouse/human fibrosis signature to classify liver-biopsy fibrosis stages. The authors compared the model with Fib-4, APRI and FibroScan in held-out and independent cohorts.
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC11137090/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 70 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Reanalysis of single-cell profiles from four sepsis patients and two healthy controls identified two neutrophil clusters enriched in the sepsis samples and a set of cluster-associated gene modules. LASSO and seven additional machine-learning classifiers prioritized eight hub genes, including ALPL, CD177, S100A8, S100A9 and STXBP2. Expression analyses in separate public sepsis transcriptome cohorts supported several markers, and qRT-PCR in cecal-ligation/puncture and LPS mouse models found increased neutrophil expression of those five genes. The result links sepsis-associated neutrophil states to a small marker panel, with experimental expression validation in mice.
Technology: LASSO feature selection followed by seven machine-learning classifiers (k-nearest neighbors, linear discriminant analysis, logistic regression, naive Bayes, random forest, recursive partitioning and support-vector machine), combined with hdWGCNA and single-cell transcriptomic analysis
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Expression analyses in separate public sepsis transcriptome cohorts supported several markers, and qRT-PCR in cecal-ligation/puncture and LPS mouse models found increased neutrophil expression of those five genes.
- Still unknown: The single-cell discovery comparison was very small (four sepsis cases and two controls) and the two clusters are analysis-derived states in this dataset; the study does not establish that they are distinct universal neutrophil lineages. The public bulk cohorts validate expression/classification associations, not an independent single-cell subtype. qRT-PCR confirmed expression changes for five genes in mouse models; it did not test whether the genes cause sepsis phenotypes or are therapeutic targets. The article itself notes earlier reports of other sepsis-related neutrophil subtypes, so this record is bounded to its specific clusters/markers and assays.
- AI role: Machine-learning feature selection prioritized eight hub genes from expression modules associated with the two sepsis-enriched neutrophil clusters. The models were assessed on separate public training and validation transcriptome cohorts; the resulting marker set guided qRT-PCR validation of five genes in two sepsis mouse models.
Evidence sources: https://www.frontiersin.org/journals/immunology/articles/10.3389/fimmu.2024.1377817/full
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 10% of works of the same type, year and subfield (OpenAlex work metrics, 19 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 2/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Machine-learning-guided synthesis and testing identified a lead-free metal-halide perovskite for oxidative alkene cleavage. The publisher reports an 80% benzoic-acid yield under the study’s conditions, compared with 6% for the initial Cs2AgBiBr6 benchmark. This is a formulation- and reaction-specific laboratory result, not universal photocatalytic performance or industrial deployment.
Technology: Machine-learning-assisted materials selection and iterative synthesis
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Publisher article reports an experimentally tested ML-selected perovskite yielding 80% benzoic acid versus 6% for the initial benchmark in the specified reaction.
- Still unknown: Independent replication and broader application are not established by this bounded review.
- AI role: ML ranks perovskite candidates using experimental feedback across synthesis rounds; researchers synthesize and test the selected photocatalysts.
Evidence sources: https://advanced.onlinelibrary.wiley.com/doi/abs/10.1002/advs.202309714
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review date: 16 Nov 2026. Review scope: Publisher indexed primary full text and metadata; specific formulation and selected result reviewed. Earlier disclosure not established in this pass..
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
An iterative active-learning campaign engineered a streptavidin artificial metalloenzyme for the known-to-the-authors gold-catalyzed hydroamination of 2-ethynylaniline to indole. Gaussian-process models guided screening across five residues in a 3.2-million-variant sequence space. The final-round library contained 38% variants more active than the prior reference, compared with 3% in the initial library. The leading variant, Sav CTNLV, showed 18-fold higher cell-specific activity than wild-type Sav and threefold higher activity than the previous reference variant in replicate testing.
Technology: Active-learning Gaussian-process regression model with optimized sequence descriptors and experimental-noise estimates
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: The leading variant, Sav CTNLV, showed 18-fold higher cell-specific activity than wild-type Sav and threefold higher activity than the previous reference variant in replicate testing.
- Still unknown: The hydroamination reaction and streptavidin–gold artificial metalloenzyme were established in earlier work by the same group; this record is specifically the later AI-guided identification of a higher-activity five-residue variant, not discovery of a new reaction. Activity was measured in an E. coli periplasmic whole-cell assay; purified variants did not retain the same rank order in vitro. Industrial utility or catalytic scale-up was not demonstrated.
- AI role: The model used measured sequence-activity data from thousands of variants to select exploratory and exploitation libraries across five residues, then prioritized variants for experimental testing. It enabled identification of a high-activity sequence in a large combinatorial space; the paper reports considering experimental uncertainty in its rankings.
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC11273458/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 10% of works of the same type, year and subfield (OpenAlex work metrics, 37 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 2/5 · Validation 3/5 · Novelty 2/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A human-supervised self-driving laboratory selected and physically tested more than 25,000 additively manufactured polymer shell designs across an 11-dimensional parameter space. The campaign found a structure with 75.2% energy-absorbing efficiency, exceeding the paper’s previously reported maximum of 71.8% for balsa wood, and the authors derived transferable design heuristics from the resulting dataset. The result is a structural design for specified compression thresholds, not a universal material efficiency ranking.
Technology: MAMA BEAR self-driving laboratory with Bayesian optimization and automated additive manufacturing/testing
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: A human-supervised self-driving laboratory selected and physically tested more than 25,000 additively manufactured polymer shell designs across an 11-dimensional parameter space.
- Still unknown: The result concerns generalized cylindrical shells made from a limited set of polymers and a defined nondimensional metric, whose value depends on threshold stress. The optimization was human-in-the-loop, and transfer to other geometries, materials or operational settings is not established.
- AI role: Bayesian optimization selected designs from over a trillion possible parameter combinations; the laboratory manufactured and mechanically tested candidates, while human researchers monitored and periodically modified the campaign.
Evidence sources: https://www.nature.com/articles/s41467-024-48534-4
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 10% of works of the same type, year and subfield (OpenAlex work metrics, 60 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A MeerKAT survey of the Small Magellanic Cloud reported seven new radio pulsars. Its pipeline partially classified folded candidates with PICS, a machine-learning system, and applied score and signal-to-noise thresholds before volunteer visual review and follow-up. The paper does not identify each pulsar’s PICS score, so this is an AI-assisted survey-level result rather than a claim that the model independently discovered all seven sources. The paper also reports pulsar–nebula associations; the later timing paper follows this cohort and is not counted again.
Technology: Pulsar Image-based Classification System (PICS), a machine-learning classifier for folded pulsar candidates
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: The survey reports seven new SMC radio pulsars; candidate triage included PICS filtering followed by human review and observational follow-up.
- Still unknown: The paper does not report per-pulsar PICS scores; it does not establish model-only detection or that PICS was essential for each individual pulsar.
- AI role: PICS partially classified folded candidates and filtered the candidate set before human visual classification.
Evidence sources: https://arxiv.org/html/2405.12029v1 · https://pure.mpg.de/pubman/item/item_3597707_4/component/file_3597709/stae1310.pdf
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review date: 17 Nov 2026. Review scope: Primary arXiv v1 methods, results and author front matter; peer-reviewed author-deposited journal PDF for publication date, author affiliations and survey result..
Editorial assessment: Impact 2/5 · Validation 3/5 · Novelty 2/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Bayesian optimization guided experiments on a five-component self-assembled photocatalytic micelle system for CO₂-to-CO reduction. The authors report that optimized conditions increased measured CO yield from 1.48 ± 0.24 to 5.72 ± 0.12 μmol h⁻¹, with improved turnover and quantum-yield metrics while maintaining CO selectivity. The article’s narrative fold-change (4.2×) is inconsistent with the displayed means (about 3.9×), so this record gives the means and omits the fold claim. The findings apply to the tested aqueous micelle system and conditions.
Technology: Bayesian optimization for experimental selection; random-forest models and SHAP for later analysis
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: The article reports CO yield means of 5.72±0.12 and 1.48±0.24 μmol h−1 for optimized vs initial conditions, along with turnover and quantum-yield metrics and maintained selectivity. The narrative states a 4.2-fold yield improvement, inconsistent with these displayed means; the record does not repeat the fold value.
- Still unknown: This demonstrates a study-specific photocatalytic system optimization, not industrially scaled carbon conversion or independent replication. The source’s stated yield fold change conflicts with its displayed means.
- AI role: Bayesian optimization guided the experimental parameter search and selected test conditions in a five-component micelle system; random-forest models and SHAP then analyzed parameter effects.
Evidence sources: https://api.repository.cam.ac.uk/server/api/core/bitstreams/9bb1fb68-8d73-4b07-b47b-376146e130f8/content · https://pubs.acs.org/doi/pdf/10.1021/jacs.4c01305?download=true
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 17 Nov 2026. Review scope: Primary ACS publisher metadata and University of Cambridge repository copy of the article full text checked for article identity, affiliations, date, optimization method, results and source discrepancy..
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
An ensemble of multitask convolutional neural networks trained on 125 peptide analogues guided optimization of 15 synthesized variants. The study reports three potent dual agonists; P3 improved potency at both glucagon and GLP-1 receptors compared with the best dual agonist in the training set. These are cell-based receptor assays only. The paper reports weak performance for designing GCGR-selective peptides and establishes no clinical safety or therapeutic efficacy.
Technology: Ensemble multitask convolutional neural networks for peptide sequence optimization
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Primary paper reports the stated assay results.
- Still unknown: Independent replication and exhaustive prior-art review not established.
- AI role: Ensemble multitask convolutional neural networks for peptide sequence optimization
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC11374683/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 16 Nov 2026. Review scope: Primary publisher/PMC article and journal metadata reviewed; author affiliation confirmed from publisher article PDF. The PMC full-text open displayed a reCAPTCHA..
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A vision transformer applied to multispectral Sentinel-2 data detects methane plumes as small as 0.01 km2, corresponding to sources of about 200 to 300 kg CH4 per hour, and is reported to improve on the prior state of the art by an order of magnitude. Detections were compared with airborne methane measurement campaigns; the claims rest on that comparison and the abstract gives no formal detection-rate metrics.
Technology: Vision transformer for multispectral methane plume detection (Sentinel-2)
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: Detections were compared with airborne methane measurement campaigns, and the method is reported to detect plumes down to 0.01 km2 (about 200 to 300 kg CH4 per hour); the authors describe this as supporting evidence rather than a definitive validation.
- Still unknown: Not reported; the page lists no explicit limitations and gives no formal detection-rate metrics.
- AI role: Vision transformer applied to multispectral Sentinel-2 data to detect methane plumes, reported to improve on prior methods by an order of magnitude.
Evidence sources: https://www.ebi.ac.uk/europepmc/webservices/rest/search?query=DOI:10.1038/s41467-024-47754-y&resultType=core&format=json
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 115 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 4/5 · Validation 2/5 · Novelty 4/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Researchers combined a genetic chemical-proximity platform and experimental screening of 72 Mycobacterium smegmatis protein candidates with machine learning. A Lasso model trained on measured degradation data identified N-terminal disorder propensity as associated with degradation potential and predicted candidate targets among 348 essential proteins conserved between M. smegmatis and M. tuberculosis. Experimentally induced degradation of selected essential M. smegmatis proteins inhibited bacterial growth and, for some targets, potentiated antimicrobial compounds.
Technology: Lasso regression model trained on sequence and structural protein descriptors, including AlphaFold2 confidence and predicted disorder propensity
Family AlphaFold: profile · models page
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Researchers combined a genetic chemical-proximity platform and experimental screening of 72 Mycobacterium smegmatis protein candidates with machine learning.
- Still unknown: Most mechanistic and growth evidence was generated in M. smegmatis, a non-pathogenic model, not in clinical treatment or an animal efficacy study. The machine-learning target predictions for the larger set of 348 essential proteins require orthogonal experimental validation. The study reports that targeted degradation of some proteins inhibits growth and potentiates antimicrobials; this does not establish a deployable drug or clinical benefit. Primary paper source checked; no independent replication assessed.
- AI role: Machine learning connected measured degradation constants from experimentally screened protein substrates to protein sequence/structure features, identifying N-terminal disorder propensity as a predictor of degradation potential; the trained model then ranked candidate targets among essential proteins conserved between M. smegmatis and M. tuberculosis. Experiments—not the model alone—showed degradation and antibacterial phenotypes for selected targets.
Evidence sources: https://www.nature.com/articles/s41467-024-48506-8
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 10% of works of the same type, year and subfield (OpenAlex work metrics, 28 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A six-model ensemble classified 1.6 million commercial molecules and reduced them to a 1,672-compound CDK8-focused screening library. Docking and visual inspection selected 25 candidates for enzymatic assays; six showed CDK8 inhibition. The most potent hit, T479-0984, had a reported IC50 of 99.9 nM. The paper reports low structural similarity (Tanimoto <0.300) to a comparison set of 30 diverse known CDK8 inhibitors. These are assay hits and starting points, not validated drugs or evidence of therapeutic efficacy.
Technology: Six-model ensemble classifier (logistic regression, k-nearest neighbors, support vector classification, random forest, XGBoost and multilayer perceptron)
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Twenty-five candidates from the focused library underwent CDK8 enzymatic testing; six showed inhibition. T479-0984 was reported with IC50 99.9 nM. The paper compared hits against 30 diverse known CDK8 inhibitors and reports Tanimoto similarity below 0.300.
- Still unknown: The assay establishes in-vitro enzyme inhibition only, not cellular target engagement, animal efficacy, clinical utility, or exhaustive prior patent/publication status for the compounds.
- AI role: The six-model ensemble filtered a 1.6-million-molecule commercial library into a 1,672-molecule CDK8-focused library; docking and visual inspection selected candidates that were then tested enzymatically.
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC11081523/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review date: 17 Nov 2026. Review scope: Primary PMC full text, article metadata, methods/results, author affiliations and compound-level structural comparison reviewed..
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 2/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
It predicts joint structures of proteins, nucleic acids, small molecules and ions. DeepMind and Isomorphic Labs report at least 50% greater accuracy than existing methods for protein interactions with other molecules. Free for non-commercial research through the AlphaFold Server.
Technology: AlphaFold 3
Family AlphaFold 3: profile · models page
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: Higher accuracy than docking tools for protein–ligand interactions, than nucleic-acid-specific predictors for protein–nucleic acid interactions, and than AlphaFold-Multimer v2.3 for antibody–antigen prediction.
- Still unknown: Accuracy on new experimental targets beyond the benchmarks is not described in the abstract.
- AI role: A diffusion-based architecture predicting joint structures of proteins, nucleic acids, small molecules, ions and modified residues.
Evidence sources: https://doi.org/10.1038/s41586-024-07487-w
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: blog. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 17949 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 5/5 · Validation 2/5 · Novelty 4/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
An agent with 18 tools and GPT-4 planned and carried out the synthesis of DEET and three thiourea organocatalysts on a cloud platform, and guided discovery of a chromophore absorbing at 336 nm against a 369 nm target; evaluated on 14 cases by four expert chemists. Hallucinations, limited tasks and dual-use risks; 12 of 18 tools were released.
Technology: ChemCrow (GPT-4 with tools)
Family GPT-4: profile · models page
Publication: Preprint · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Preprint
- Test type: direct test
- Tested: Expert and LLM judgments were used to evaluate the agent, which planned and executed syntheses of an insect repellent and three organocatalysts and guided discovery of a chromophore.
- Still unknown: The abstract says LLMs struggle with chemistry and lack external knowledge, and GPT-4 as evaluator could not distinguish clearly wrong completions from ChemCrow's performance.
- AI role: An LLM agent with 18 expert-designed chemistry tools planned and executed syntheses and guided molecular discovery; the page does not name the underlying model.
Evidence sources: https://arxiv.org/abs/2304.05376
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 1741 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Analysis of nearly 9,000 codas from the eastern Caribbean clan, collected over more than ten years, proposes a combinatorial structure involving rhythm, tempo, rubato and ornamentation. Published in Nature Communications; communicative function remains unknown, data cover one clan, and the page does not specify machine-learning methods.
Technology: Machine-learning analysis of whale calls
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: Nearly 9,000 codas from the eastern Caribbean clan, collected over more than ten years, were analyzed for context-sensitive features (rubato, ornamentation) and categorical rhythm and tempo.
- Still unknown: The communicative function of many codas is unknown, whether the features encode intent needs future field research, and the duality-of-patterning link is only a possibility.
- AI role: Not reported; the page does not specify what machine learning did in the analysis.
Evidence sources: https://www.projectceti.org/blog-posts/sperm-whale-phonetic-alphabet-proposed-for-the-first-time
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: press release. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 96 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 4/5 · Validation 2/5 · Novelty 4/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Targeted serum metabolomics compared 89 patients with systemic lupus erythematosus (SLE) and 44 healthy participants, identifying 146 differential metabolites. The authors combined statistical screening with seven machine-learning models to prioritize eight diagnostic biomarker candidates (reported AUCs above 0.95). Two additional metabolites, N-methyl-L-glutamic acid and L-2-aminobutyric acid, were evaluated by ELISA in a 90-patient SLE validation cohort and associated with SLE Disease Activity Index scores.
Technology: Seven machine-learning classification models applied to targeted serum metabolomics
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: The authors combined statistical screening with seven machine-learning models to prioritize eight diagnostic biomarker candidates (reported AUCs above 0.95).
- Still unknown: This retrospective case-control study has modest cohorts and does not establish a clinically deployed test, causal SLE mechanism, or patient benefit. Reported diagnostic AUCs and SLEDAI associations require independent prospective replication; the validation cohort described in the abstract consists of SLE patients and does not by itself establish performance against a new external healthy/control population. The authors’ pathway interpretation is hypothesis-generating.
- AI role: The study iteratively compared seven machine-learning models to select candidate metabolites for SLE classification and combined the selected features with statistical screening. N-methyl-L-glutamic acid and L-2-aminobutyric acid were subsequently assessed by ELISA in a patient validation cohort; the machine-learning analysis nominated candidates, while the assays supplied biological measurement follow-up.
Evidence sources: https://pubmed.ncbi.nlm.nih.gov/38710348/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 10% of works of the same type, year and subfield (OpenAlex work metrics, 19 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 2/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
HydrogelFinder used a generative model and learned prediction to prioritize 111 candidate peptides; 17 were synthesized and nine self-assembled under tested conditions. Cell assays covered two selected materials and do not establish general safety or clinical suitability.
Technology: HydrogelFinder generative language model and learned gelation prediction
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Primary abstract reports 111 candidates, 17 synthesized and nine experimentally self-assembling; cell compatibility tests were limited to two selected materials.
- Still unknown: Independent replication and exhaustive prior-art review are not established.
- AI role: HydrogelFinder generative language model and learned gelation prediction
Evidence sources: https://pubmed.ncbi.nlm.nih.gov/38704695/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 16 Nov 2026. Review scope: Primary source checked for study identity, AI operation, dates, affiliations and the bounded result..
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
The LAMBDAZERO generative active-learning workflow searched for synthesizable soluble epoxide hydrolase (sEH/EPHX2) inhibitors and produced a quinazoline scaffold absent from the authors' comparison set of known sEH inhibitors. A synthesized analog, UM0152893, inhibited sEH in vitro with sub-micromolar potency.
Technology: LAMBDAZERO generative active learning with deep reinforcement learning and docking
Publication: Preprint · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Preprint
- Test type: direct test
- Tested: The LAMBDAZERO generative active-learning workflow searched for synthesizable soluble epoxide hydrolase (sEH/EPHX2) inhibitors and produced a quinazoline scaffold absent from the authors' comparison set of known sEH inhibitors.
- Still unknown: This is a workshop paper/preprint rather than a clinical study. The reported lead is a biochemical inhibitor of purified sEH; no cellular, animal, safety or therapeutic outcome is established. The authors report 24/35 synthesized analogs active in vitro with potency spanning roughly 90 μM to 0.4 μM; one best lead is sub-micromolar.
- AI role: The generative model searched molecular space using docking, synthesizability and drug-likeness constraints and proposed the scaffold; researchers synthesized analogs and measured enzyme inhibition in biochemical assays.
Evidence sources: https://arxiv.org/abs/2405.01616
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Editorial assessment: Impact 2/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A study combined GPT-3.5-generated gene-function context with a genome-wide Drosophila RNAi behavioral screen. It identified MRE11 knockdown as reducing sleep in group-housed flies, a result the authors report as previously unrecognized in this social context. GPT-3.5 reasoning proposed a link from MRE11 to Dop1R1/dopaminergic signaling; treatment with a Dop1R1 antagonist partially rescued sleep, locomotor and social-activity phenotypes in MRE11-RNAi flies.
Technology: GPT-3.5, prompted for genome-wide Drosophila gene-function interpretation and chain-of-thought relationship hypotheses
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: It identified MRE11 knockdown as reducing sleep in group-housed flies, a result the authors report as previously unrecognized in this social context.
- Still unknown: The experiments are in Drosophila and concern sleep in the presence of conspecifics. Dop1R1 antagonist treatment only partially rescued phenotypes; this does not establish direct MRE11 regulation of the receptor or dopamine pathway. GPT-3.5 had high false-negative rates in the authors’ known-gene evaluation and contributed knowledge summarization/hypothesis generation, while the broad RNAi screen itself was experimental. No mammalian or human result is implied.
- AI role: GPT-3.5 was queried about fly gene functions and used to construct hypothesized relationships among behavioral-screen candidates; it proposed a MRE11–Dop1R1 link that researchers tested pharmacologically. The RNAi screen and behavioral measurements were experimental, while direct molecular regulation was not demonstrated.
Evidence sources: https://www.nature.com/articles/s41467-024-48005-w
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 10% of works of the same type, year and subfield (OpenAlex work metrics, 26 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 2/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
In a two-month campaign, more than 1,000 candidates were synthesized and tested at six laboratories across five sites on three continents, identifying at least 21 high-performance candidates as gain materials from a space of more than 150,000 experiments. These are gain candidates, not confirmed lasers; the page gives no laser metrics; source is a consortium press release.
Technology: Closed-loop optimization with autonomous laboratories
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Over 1,000 candidates were synthesized and tested in a two-month distributed campaign across six teams in five labs on three continents, identifying at least 21 high-performance gain candidates.
- Still unknown: The commercialization speed goal was not met, the candidates are gain materials rather than confirmed lasers, and the page gives no laser metrics.
- AI role: A closed-loop AI reviewed data, selected molecules to synthesize and test, and predicted the next best experiment; a cloud platform coordinated the labs.
Evidence sources: https://acceleration.utoronto.ca/news/published-in-the-journal-science-new-paper-delocalized-asynchronous-closed-loop-discovery-of-organic-laser-emitters-demonstrates-the-acs-global-leadership-in-self-driving-labs
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: press release. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 188 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 4/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A mixed workflow using deep reinforcement learning, pharmacophore and shape filters, docking and hit expansion reports three SARS-CoV-2 main-protease inhibitor series with biochemical IC50 values of 1.3–2.3 μM. The paper combines generated and commercially sourced hits; not every final inhibitor is claimed as directly model-generated. PDB entry 8R11 was released before the article and already disclosed a bound compound structure, so no priority claim is made for that structure. These biochemical findings do not establish antiviral efficacy or clinical safety.
Technology: Deep reinforcement learning; pharmacophore/shape filtering; docking
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Publisher abstract reports three experimentally validated Mpro inhibitor series with IC50 1.3–2.3 μM; PDB 8R11 pre-dates the paper.
- Still unknown: Exact assignment of each final inhibitor to generated versus commercially sourced material and prior chemotype overlap not fully audited; no antiviral or clinical claims.
- AI role: Reinforcement learning generates candidate structures; computational filters and hit expansion guide follow-up, but commercially sourced hits also contributed.
Evidence sources: https://pubs.rsc.org/en/content/articlepdf/2024/md/d4md00106k
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review date: 16 Nov 2026. Review scope: Primary RSC article landing page and full-text PDF abstract; first-publication date verified. PDB identity/release was reported in input ledger and should be retained as earlier structure disclosure..
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
SpliceCore’s XGBoost models used splicing-factor RNA-binding and spliceosome information to nominate NEDD4L exon 13 (NEDD4Le13) and the SSO-0205 binding site. In triple-negative breast-cancer MDA-MB-231 cells, SSO-0205 increased exon inclusion (67% versus 15% control), reduced proliferation and decreased migration by 40–50%; the effects were not observed in the non-tumorigenic MCF10A comparator under the reported assays.
Technology: SpliceCore XGBoost models with SHAP and out-of-bag feature analysis
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: In triple-negative breast-cancer MDA-MB-231 cells, SSO-0205 increased exon inclusion (67% versus 15% control), reduced proliferation and decreased migration by 40–50%; the effects were not observed in the non-tumorigenic MCF10A comparator under the reported assays.
- Still unknown: Experiments were in vitro, including three breast-cell lines and limited biological replicates; the migration assay used two to three replicates. These results do not establish tumor control in animals, patient benefit, safety or clinical efficacy. The source reports the target as novel in the studied context; this is not evidence that the SSO is a treatment.
- AI role: The models prospectively nominated the NEDD4L exon-13 splice event, predicted functional SSO binding sites, and prioritized implicated splicing factors. Researchers synthesized selected oligonucleotides and tested splicing, proliferation and migration in cell lines.
Evidence sources: https://link.springer.com/article/10.1038/s44320-024-00034-9
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: at percentile 78 of works of the same type, year and subfield (OpenAlex work metrics, 12 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 2/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A generative model designed molecules from the three-dimensional surfaces of PPAR proteins: synthesized in the laboratory, they increased activity and were tested by Roche. They demonstrate the method and are not intended for commercial development; the tests do not establish safety or efficacy in humans.
Technology: Protein-surface-guided generative AI
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: ETH chemists synthesized the AI-designed molecules in the lab, and Roche ran a variety of tests showing they are stable and non-toxic.
- Still unknown: The article does not state limitations, the molecules were not developed toward market, and the tests do not establish safety or efficacy in humans.
- AI role: A generative model designed molecules from scratch matched to a protein's three-dimensional surface, without human intervention.
Evidence sources: https://ethz.ch/en/news-and-events/eth-news/news/2024/04/ai-designs-new-drugs-based-on-protein-structures.html
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: press release. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 121 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Profluent built an atlas of 5.1 million CRISPR-Cas proteins and generated a Cas9-like editor with 403 mutations relative to SpCas9. Of 48 generated sequences, the best (OpenCRISPR-1) achieved 55.7% on-target editing versus 48.3% for SpCas9, and 0.32% off-target editing versus 6.1%. Human cells in the laboratory only; company data; the page does not indicate peer review.
Technology: Protein language models
Family OpenCRISPR: profile · models page
Publication: Publication status unspecified in annotation · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Not reported
- Test type: direct test
- Tested: Of 48 selected generated sequences, the best (OpenCRISPR-1) was tested in human cells: 55.7% on-target editing vs 48.3% for SpCas9, and 0.32% off-target vs 6.1%.
- Still unknown: The company article has no explicit limitations section and does not state whether a preprint or peer review exists.
- AI role: A language model pretrained on protein data and steered to CRISPR families using a 5.1-million-protein atlas generated Cas9-like nucleases and compatible sgRNAs.
Evidence sources: https://www.profluent.bio/media/editing-the-human-genome-with-ai
The legacy label “Not reported” does not distinguish information absent from reviewed material from information not assessed. It is not a scientific finding of absence.
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: press release. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Editorial assessment: Impact 4/5 · Validation 3/5 · Novelty 5/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A machine-learning and multi-omics framework screened 1.09 million pairs across 408 human GPCRs and 335 gut microbial metabolites. It prioritized orphan receptor GPR84 and the metabolite triacanthine, which activated GPR84 in a cell assay. In Alzheimer’s patient-derived iPSC neurons, agmatine reduced C3AR and tau markers, while phenethylamine reduced phosphorylated and total tau in the reported experiments.
Technology: Extra Trees machine-learning model integrated with molecular docking, multi-omics and Mendelian-randomization analyses
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: It prioritized orphan receptor GPR84 and the metabolite triacanthine, which activated GPR84 in a cell assay.
- Still unknown: The metabolite–receptor screen is computational and many pairs remain predictions. Receptor activation was tested in a cell assay and tau-marker effects in patient-derived neurons; these experiments do not demonstrate clinical benefit, in-vivo efficacy or that altered tau markers translate to disease modification. Mendelian-randomization results and model scores are not proof of causality.
- AI role: The ML model ranked predicted metabolite–GPCR associations over a large search space; multi-omics and genetic analyses prioritized Alzheimer’s-related receptors and metabolites. These predictions guided follow-up assays, including triacanthine activation of GPR84 and testing agmatine and phenethylamine in patient-derived neurons.
Evidence sources: https://www.sciencedirect.com/science/article/pii/S221112472400456X
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 10% of works of the same type, year and subfield (OpenAlex work metrics, 45 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A structure-based iterative-learning workflow combined prior α-synuclein fibril-binding hits, a junction-tree variational autoencoder, a random-forest regressor and Gaussian-process residual model to rank compounds for repeated testing. Prospective iterations produced multiple molecules inhibiting seeded α-synuclein aggregation in vitro; lead I4.05 had a reported kinetic inhibitory constant (KIC50) of 0.52 μM (95% CI 0.45–0.59 μM), compared with extrapolated values of 18.2 μM for its parent and 36.4 μM for Anle-138b in the authors’ assay. Leads also showed activity in lipid-induced and spontaneous aggregation assays.
Technology: Iterative structure-based learning workflow using a junction-tree variational autoencoder, random-forest regressor and Gaussian-process residual model
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: A structure-based iterative-learning workflow combined prior α-synuclein fibril-binding hits, a junction-tree variational autoencoder, a random-forest regressor and Gaussian-process residual model to rank compounds for repeated testing.
- Still unknown: Evidence is biochemical/in-vitro inhibition of α-synuclein aggregation, including seeded, lipid-induced and spontaneous assays. The paper does not establish animal efficacy, neuronal protection or clinical benefit. The KIC50 and comparison values are assay-specific; no random-molecule comparator was tested, and model QSAR R² was described as modest. This is related to, but not the same paper/result as, the earlier docking-screen study that supplied starting compounds.
- AI role: The machine-learning model ranked a ZINC-derived evaluation set for potency using earlier assay measurements, then researchers tested and fed results from successive rounds back into the model. It proposed compounds for experimental testing; aggregation assays measured outcomes. Initial fibril binders came from an earlier docking-and-screening study, so this paper reports iterative optimization rather than an entirely independent start.
Evidence sources: https://www.nature.com/articles/s41589-024-01580-x.pdf
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 10% of works of the same type, year and subfield (OpenAlex work metrics, 60 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 2/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A graph-based machine-learning model ranked polymer candidates; two were synthesized and measured above selected empirical permeability–selectivity upper bounds for O2/N2, H2/CH4 and H2/N2. The bounds are historical empirical comparisons, not physical limits; the study does not demonstrate industrial mixed-gas performance. The claim here is bounded to the primary arXiv v1, which reports the synthesized candidates.
Technology: Explainable graph machine learning
Publication: Preprint · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Preprint
- Test type: direct test
- Tested: Two synthesized candidates reported above selected empirical upper bounds in specified gas pairs.
- Still unknown: Exact journal/preprint result correspondence and prior structure-level art not exhaustively checked.
- AI role: Graph ML predicts permeability/selectivity from polymer structure and prioritizes candidates for synthesis and testing.
Evidence sources: https://arxiv.org/abs/2404.10903
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 16 Nov 2026. Review scope: Primary source checked for identity, publication date, AI workflow and the bounded experimental result stated..
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
The OSMES workflow screened AlphaFold-modeled vitamin-B6-dependent enzymes by docking the substrate into active sites and ranking catalytically favorable conformations. It identified human SHMT1/2 as enzymes with hydroxytrimethyllysine-aldolase activity, validated in vitro; mouse Tha1 showed much higher catalytic efficiency, supporting a species difference in the pathway. The authors infer that humans may have lost a dedicated gene, with SHMT activity partly compensating.
Technology: OSMES (One Substrate–Many Enzymes Screening), automated structure-based docking/ranking workflow using AlphaFold models
Family AlphaFold: profile · models page
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: It identified human SHMT1/2 as enzymes with hydroxytrimethyllysine-aldolase activity, validated in vitro; mouse Tha1 showed much higher catalytic efficiency, supporting a species difference in the pathway.
- Still unknown: Enzyme activity was validated in vitro. The claim about loss of a dedicated human gene is an evolutionary inference, not a direct gene-knockout demonstration; the authors note unresolved pathway questions.
- AI role: Ranked candidate enzymes for a specified metabolic reaction using docked substrate conformations and catalytically favorable active-site geometry; guided biochemical testing that verified SHMT1/2 activity.
Evidence sources: https://www.nature.com/articles/s41467-024-47466-3.pdf
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: at percentile 87 of works of the same type, year and subfield (OpenAlex work metrics, 13 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Using satellite observations of four Kīlauea volcanic outgassing episodes in 2008 and 2018, an ML counterfactual analysis found that volcanic aerosols increased cloud cover and reflected sunlight over tropical marine shallow-convective clouds, even as rainfall strengthened. Cloud-cover adjustment contributed about 65% of aerosol–cloud cooling under mixed meteorological conditions and 84–87% in humid conditions; under humid, stable conditions, a 30% increase in droplet number could correspond to up to a 50% relative increase in cloud cover. The result identifies a tropical regime in which aerosol cloud expansion is strong and its response varies with humidity and stability.
Technology: Random-forest machine-learning surrogates of MODIS cloud properties, radiative fluxes, and precipitation, trained on 20 years of satellite observations and meteorology while excluding volcanic years 2008 and 2018
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: Using satellite observations of four Kīlauea volcanic outgassing episodes in 2008 and 2018, an ML counterfactual analysis found that volcanic aerosols increased cloud cover and reflected sunlight over tropical marine shallow-convective clouds, even as rainfall strengthened.
- Still unknown: This is an observational ML-counterfactual analysis of four Kīlauea plume months over tropical marine shallow-convective clouds, not a controlled aerosol intervention or a direct test of deliberate cloud seeding. The source reports strong cloud-cover responses and radiative estimates, especially under humid/stable conditions, but results depend on satellite retrievals and the learned unperturbed baseline. The authors’ proposed mechanism for increased cloud cover alongside stronger rainfall is explicitly a hypothesis requiring further testing; do not treat the marine-cloud-brightening scenario as experimentally demonstrated.
- AI role: The random-forest surrogates estimated unperturbed cloud properties and radiative fluxes for given meteorological conditions. Comparing those counterfactuals with observations in four volcanic plume episodes isolated aerosol-associated signals from meteorological co-variability and enabled estimates of cloud-cover, rainfall, and radiative-forcing responses across distinct tropical meteorological regimes.
Evidence sources: https://www.nature.com/articles/s41561-024-01427-z
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 10% of works of the same type, year and subfield (OpenAlex work metrics, 63 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Baker’s group showed RFdiffusion can be adapted to generate antibodies: in laboratory tests, designs bound influenza virus and other targets as intended. Proof-of-concept study, first circulated as a preprint in March 2024; experts quoted by Nature call it promising but an early step; no affinity values reported.
Technology: RFdiffusion
Family RFdiffusion: profile · models page
Publication: Preprint · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Preprint
- Test type: direct test
- Tested: Laboratory tests confirmed that the designed antibodies bind influenza virus and other targets as intended; no affinity values are reported.
- Still unknown: The work is described as a proof-of-concept study, and no specific limitations are listed.
- AI role: RFdiffusion was tuned to generate antibodies from scratch.
Evidence sources: https://www.ipd.uw.edu/2024/04/nature-a-landmark-moment-scientists-use-ai-to-design-antibodies-from-scratch/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: press release. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 4/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A machine-learning system selected semapimod, a known anti-inflammatory drug, for activity against dormant Gram-negative bacteria. MIT reports laboratory activity against E. coli and A. baumannii and an effect that sensitizes these bacteria to other antibiotics; the result is preclinical and does not demonstrate efficacy in patients.
Technology: Machine learning for compound screening
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Lab-only time-lapse microscopy of E. coli treated with semapimod showed activity against stationary-phase E. coli and A. baumannii and sensitization to other antibiotics.
- Still unknown: The article states no limitations, no animal or patient testing is mentioned, and no drug combinations are named.
- AI role: AI-based high-throughput screening of known drug compounds identified semapimod for antibiotic activity.
Evidence sources: https://news.mit.edu/2024/mit-scientists-are-using-artificial-intelligence-target-sleeper-bacteria-0408
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: press release. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 10% of works of the same type, year and subfield (OpenAlex work metrics, 59 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
An earthquake catalog built with the PhaseNet machine-learning phase detector and cross-correlation/template matching revealed a statistically significant (over 99%) increase in small earthquakes as the second Baige landslide-dammed lake approached peak water level; rates returned toward background after breach. Coulomb-stress modeling found about 90% of the earthquakes in regions of positive modeled stress change, supporting likely triggering through direct water loading and pore-pressure diffusion. The paper presents this as evidence that earthquake–surface-hazard triggering can be bidirectional.
Technology: PhaseNet machine-learning phase detector, combined with PyOcto association and cross-correlation/template-matching methods
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: An earthquake catalog built with the PhaseNet machine-learning phase detector and cross-correlation/template matching revealed a statistically significant (over 99%) increase in small earthquakes as the second Baige landslide-dammed lake approached peak water level; rates returned toward background after breach.
- Still unknown: This is a case study of one pair of 2018 landslide-dammed lakes and nearby faults. Event-location uncertainties are several kilometers; fault geometry and fluid-diffusion parameters require assumptions. The authors conclude the lakes likely triggered the sequence, rather than directly observing causality, and the global implication is modeled potential.
- AI role: PhaseNet detected P- and S-wave arrivals in continuous seismic records; these detections supported a new 2014–2023 earthquake catalog used to identify and statistically test the local seismicity increase. Coulomb-stress modeling then assessed a possible physical mechanism.
Evidence sources: https://www.nature.com/articles/s41467-024-47130-w
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 10% of works of the same type, year and subfield (OpenAlex work metrics, 39 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 2/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A custom machine-learning approach suggested constructions that authors formalized as two six-colorings. The primary source reports an off-diagonal distance range of 0.354 to 0.657, extending the previously known range. This concerns a specified off-diagonal variant and does not solve the original Hadwiger–Nelson problem.
Technology: Custom machine-learning approach; neural networks; differentiable optimization
Publication: Preprint · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Preprint
- Test type: direct test
- Tested: The source reports the specific mathematical or algorithmic result summarized here.
- Still unknown: The result is attributed to the paper; no independent proof review beyond the authors’ source text is claimed.
- AI role: Neural optimization suggests geometric configurations; authors formalize valid colorings and prove the construction.
Evidence sources: https://arxiv.org/abs/2404.05509 · https://proceedings.mlr.press/v267/mundinger25a.html
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 16 Nov 2026. Review scope: Primary preprint full text or abstract reviewed; evidence scope is specified in source_checks..
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Progeni combines a probabilistic knowledge graph with graph neural networks to rank disease-associated molecular targets. The authors selected six candidates lacking literature support and reported that knockdown of HSP90AB1 and MME reduced proliferation in a murine melanoma line, while knockdown of ADCY5, ADRA2A and EEF2 reduced proliferation in a murine colorectal-cancer line.
Technology: Progeni probabilistic knowledge graph and graph neural network
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: The authors selected six candidates lacking literature support and reported that knockdown of HSP90AB1 and MME reduced proliferation in a murine melanoma line, while knockdown of ADCY5, ADRA2A and EEF2 reduced proliferation in a murine colorectal-cancer line.
- Still unknown: The functional validations were in B16F10 and MC38 murine cell lines, not human tumor samples or clinical settings. The paper evaluates a model and validates selected candidates; it does not establish these as clinically effective therapeutic targets.
- AI role: The graph neural network ranked new target hypotheses from heterogeneous biological networks and literature co-occurrence evidence. Researchers then tested selected predictions by shRNA knockdown in representative murine cancer cell lines.
Evidence sources: https://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1011945
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: at percentile 81 of works of the same type, year and subfield (OpenAlex work metrics, 11 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 2/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A generative model and protein-language-model-based temperature selector were combined in an iterative design procedure. The article reports eight heat-tolerant glyceraldehyde-3-phosphate dehydrogenase variants among 30 generated and tested sequences. This is an experimentally screened enzyme set, not proof of industrial deployment.
Technology: DeepEvo generative model; protein language model temperature selector
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Publisher article reports eight heat-tolerant variants among 30 generated sequences tested.
- Still unknown: Exact assay definition, earlier disclosure and sequence-level prior overlap remain open.
- AI role: Generative model proposes functional sequences and PLM-based selector ranks high-temperature candidates in iterative design/test loop.
Evidence sources: https://spj.science.org/doi/10.34133/bdr.0031
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 16 Nov 2026. Review scope: Primary source checked for identity, date, model role and bounded reported result..
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
The researchers combined personalized gut-microbiome community metabolic models from human iHMP data with tree-based machine-learning feature selection to identify microbial strains and metabolites associated with ulcerative-colitis samples. Parabacteroides merdae ATCC 43,184 and N-acetyl-D-mannosamine were among the most predictive features; the strain was also the main modeled microbial contributor to the metabolite. In a DSS-induced mouse colitis experiment, oral administration of the bacterium or metabolite improved reported colitis measures.
Technology: Tree-based supervised classifiers including random forest, gradient-boosted regression trees, XGBoost, XGBoost-random forest and LightGBM, combined with individualized genome-scale microbiome community models
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: In a DSS-induced mouse colitis experiment, oral administration of the bacterium or metabolite improved reported colitis measures.
- Still unknown: The human data are observational microbiome samples, while intervention was tested in a DSS-induced mouse model, not in people. The animal groups contained eight mice each; findings do not establish treatment efficacy or safety in human ulcerative colitis. Metabolite-production associations came from community-model simulations and were not direct evidence of individual strain production in each person. The model provides candidate prioritization, not a causal human mechanism.
- AI role: Personalized community metabolic models estimated strain-level contributions to secreted metabolites in human iHMP samples. Tree-based classifiers selected discriminative strain and metabolite features; XGBRF feature attribution ranked P. merdae ATCC 43,184 and N-acetyl-D-mannosamine highly, motivating the mouse validation. The AI analysis nominated these targets; it did not establish human therapeutic efficacy.
Evidence sources: https://www.tandfonline.com/doi/full/10.1080/19490976.2024.2336877
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: at percentile 88 of works of the same type, year and subfield (OpenAlex work metrics, 19 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Using a sample selected by unsupervised machine learning on optical and mid-infrared photometry, the team obtained 178 MMT spectra and found the brighter class was 85% Type I active galactic nuclei (AGN), with <3% non-AGN contamination, while a fainter class was 65% Type II AGN with 15% contamination, mainly star-forming galaxies. The follow-up showed that the machine-selected classes recover redder and more obscured AGN populations than typical optical spectroscopic samples.
Technology: Unsupervised UMAP and DBSCAN photometric AGN selection pipeline
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Using a sample selected by unsupervised machine learning on optical and mid-infrared photometry, the team obtained 178 MMT spectra and found the brighter class was 85% Type I active galactic nuclei (AGN), with <3% non-AGN contamination, while a fainter class was 65% Type II AGN with 15% contamination, mainly star-forming galaxies.
- Still unknown: This is a follow-up validation of two machine-selected candidate classes, not a claim that every cataloged object is confirmed. Spectroscopy recovered redshifts for 139 of 178 spectra; the fainter Type II class was limited to the brightest 75% of candidates and had 15% contamination, primarily from star-forming galaxies. The unsupervised pipeline is described in related work listed as in preparation in this paper; findings here are bounded to the spectroscopically analyzed sample.
- AI role: The selection pipeline applied UMAP dimensionality reduction and DBSCAN clustering to optical and mid-infrared colors and morphology, with matched spectroscopy and X-ray information used to identify AGN candidate classes. Those classes defined the targets for Hectospec follow-up; the paper then used optical spectra, emission-line diagnostics, and spectral stacking to test their AGN content. The AI selected the candidate populations, while spectroscopic diagnostics established the reported classifications.
Evidence sources: https://iopscience.iop.org/article/10.3847/1538-3881/ad28b4
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: at percentile 73 of works of the same type, year and subfield (OpenAlex work metrics, 11 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 2/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Expert review of 455 previously observed transients changed 16 normal labels to anomalous; the random-forest model flagged 11 of these. Spectra, light curves and host context supported the reclassifications. This is a retrospective study-level result rather than a count of newly observed events or newly established physical classes. Human review and similarity searches also contributed, so the entire retrospective sample is not attributed to the classifier.
Technology: LAISS random-forest anomaly classifier and similarity search
Publication: Preprint · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Preprint
- Test type: retrospective or benchmark
- Tested: Existing photometry and classification spectra were re-examined to support revised labels.
- Still unknown: Anomaly categories depend on expert judgments; universal first-discovery priority and new physical mechanisms are not established.
- AI role: LAISS random-forest anomaly classifier and similarity search supported the stated study; experimental and interpretive work remained with researchers.
Evidence sources: https://arxiv.org/abs/2404.01235v1 · https://arxiv.org/html/2404.01235v1
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review date: 16 Nov 2026. Review scope: Primary arXiv v1 HTML abstract, selected classifier/reclassification Results and affiliations; version history checked..
Editorial assessment: Impact 3/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
The model produces 256 members at 2° resolution every three minutes on a TPU and matches or exceeds the operational GEFS ensemble (31 members) on rank histograms, RMSE and CRPS. For the Lisbon heatwave (July 2022), with 16,384 members, none of GEFS’s 31 members was as warm as observed. One extreme case; depends on physics-based forecasts as seeds; published in Science Advances.
Technology: SEEDS (diffusion model)
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: The ensemble was compared with the operational GEFS (31 members) using rank histograms, RMSE and CRPS, and with a 16,384-member ensemble for the Lisbon heatwave of July 2022.
- Still unknown: The heatwave comparisons are single examples, the model depends on physics-based seed forecasts, and the page has no explicit limitations section.
- AI role: A denoising diffusion model generates large weather-forecast ensembles from one or two operational forecasts.
Evidence sources: https://research.google/blog/generative-ai-to-quantify-uncertainty-in-weather-forecasting/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: blog. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 172 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
For the 2020 Mw 5.1 Sparta, North Carolina sequence, an EQTransformer-based catalog combined with matched-filter detections and differential-time relocations contains 1,761 high-precision events, about 5.8 times the original NEIC count. The denser locations resolve several shallow clusters and support the authors’ interpretation that the mainshock nucleated near the intersection of a blind strike-slip fault and a reverse fault associated with the surface rupture; moment-tensor inversion places its centroid at about 1.3 km depth.
Technology: Earthquake Transformer (EQTransformer) deep-learning phase/event detector, followed by matched-filter detection and differential-time relocation
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: For the 2020 Mw 5.1 Sparta, North Carolina sequence, an EQTransformer-based catalog combined with matched-filter detections and differential-time relocations contains 1,761 high-precision events, about 5.8 times the original NEIC count.
- Still unknown: This is a retrospective study of one 2020 earthquake sequence, using a local temporary-network deployment. The authors describe the inferred fault intersection/flower-structure interpretation as possible; the fit for the smallest cluster is poor and its geometry uncertain. The result is not a forecast or general fault model. The AI contribution was event detection within a hybrid workflow, not independent geological validation.
- AI role: EQTransformer detections expanded the template catalog, which seeded matched-filter detection; the resulting event set was relocated to improve temporal and spatial resolution used to infer the sequence’s shallow fault geometry. The fault interpretation is author analysis from the denser catalog plus waveform and moment-tensor methods, not a model-generated classification alone.
Evidence sources: https://www.nature.com/articles/s43247-024-01316-8
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 10% of works of the same type, year and subfield (OpenAlex work metrics, 25 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 2/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
The generative model proposed about 25,000 candidates in under nine hours from more than 130,000 synthesis building blocks; of 58 synthesized, six killed a drug-resistant A. baumannii strain and two were not toxic in mice. Mechanism unknown, efficacy in infected animals untested, and four of the six compounds were not water-soluble.
Technology: SyntheMol (generative model)
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Of 58 synthesized candidates, six killed a drug-resistant A. baumannii strain in the lab, and two of those were not toxic in mice.
- Still unknown: Four of the six compounds were not water-soluble so were not tested in mice, efficacy in infected animals is untested, and the mechanism is unknown.
- AI role: A generative model proposed about 25,000 candidate antibiotics and synthesis recipes in under nine hours from over 130,000 building blocks.
Evidence sources: https://med.stanford.edu/news/all-news/2024/03/ai-drug-development.html
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: press release. Paper: linked separately.
Review date: 16 Nov 2026. Review scope: Original institutional announcement checked for the specific results, AI contribution and limitations cited in the antibiotic dossier or mathematical follow-up pilot. Full-paper reanalysis, external replication and an exhaustive follow-up search were not performed..
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 221 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 4/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
In 2,772 adults with diabetes from Singapore’s population-based SEED study, LASSO and gradient-boosting models screened 220 circulating metabolites and 19 clinical factors for prevalent diabetic kidney disease (DKD) and retinopathy (DR), then were externally assessed in 5,843 UK Biobank participants. Both outcomes shared diabetes duration, insulin use, age and tyrosine as important model features. DKD additionally selected cardiovascular disease history, antihypertensive use, lactate, citrate and the intermediate-density-lipoprotein cholesterol-ester/total-lipid ratio; DR additionally selected HbA1c, glucose, pulse pressure and alanine. External AUCs were 0.791 for DKD and 0.778 for DR, versus 0.691 and 0.760 for conventional adjusted logistic-regression comparators.
Technology: Least absolute shrinkage and selection operator logistic regression and gradient-boosting decision trees applied to metabolomics and clinical risk factors
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: In 2,772 adults with diabetes from Singapore’s population-based SEED study, LASSO and gradient-boosting models screened 220 circulating metabolites and 19 clinical factors for prevalent diabetic kidney disease (DKD) and retinopathy (DR), then were externally assessed in 5,843 UK Biobank participants.
- Still unknown: The analysis is retrospective and cross-sectional; it classifies prevalent complications and cannot establish temporal prediction or causal effects of tyrosine, lactate, citrate, alanine or lipid features. The Singapore cohort was recruited in 2004–2011 and was multiethnic Asian; about 95% had type 2 diabetes, but the study did not separate diabetes types. UK Biobank definitions and available eye-severity measures differed, and validation of moderate-or-worse DR was not feasible. Performance is modest, particularly external AUC 0.791 DKD/0.778 DR, and does not establish clinical utility.
- AI role: ML feature selection prioritized a shared set of clinical/metabolic factors and distinct DKD- and DR-associated metabolites, then built detection models externally evaluated in UK Biobank. These models outperformed traditional adjusted logistic-regression comparators in the reported internal and external validation. The scientific result is the disease-specific feature pattern and its validated detection performance, not merely an algorithm benchmark.
Evidence sources: https://www.jmir.org/2024/1/e41065/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 10% of works of the same type, year and subfield (OpenAlex work metrics, 30 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
In two multicenter longitudinal cohorts of early multiple sclerosis (discovery n=309; validation n=232), unsupervised clustering of high-dimensional flow-cytometry and serum-proteomics profiles identified three peripheral-blood immune endophenotypes. Follow-up associated the groups with different trajectories of inflammation versus early structural damage. In the endophenotype-3 subgroup, patients subsequently treated with interferon beta had higher four-year progression and MRI activity than patients receiving other therapies.
Technology: Unsupervised clustering of high-dimensional flow-cytometry and serum-proteomics profiles (the reviewed abstract does not name the clustering algorithm)
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: In two multicenter longitudinal cohorts of early multiple sclerosis (discovery n=309; validation n=232), unsupervised clustering of high-dimensional flow-cytometry and serum-proteomics profiles identified three peripheral-blood immune endophenotypes.
- Still unknown: The source reports cohort associations, not randomized treatment effects. The higher progression and MRI activity among endophenotype-3 patients treated with interferon beta may be affected by treatment selection and other confounding; it does not establish that interferon beta caused worse outcomes or that the signature can guide care. The validation cohort is an independent patient cohort within the study, not independent replication by another research group. The reviewed abstract does not provide the clustering algorithm name or establish prospective clinical utility.
- AI role: The researchers combined flow-cytometry immune measurements and serum-protein profiles, then used unsupervised clustering to define three immune endophenotypes in the discovery cohort and assess them in an independent multicenter validation cohort. The clustering generated the patient groups whose longitudinal clinical and paraclinical trajectories were compared; the repository abstract does not specify the algorithm name.
Evidence sources: https://edoc.mdc-berlin.de/id/eprint/24233/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 55 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
DIP-MS combines deep-learning-based signal processing with native-gel fractionation and mass spectrometry to resolve complex isoforms. In human prefoldin-like-complex experiments, repeated coelution and affinity-purification data supported POLR2E as a constitutive PFDL component, updating the view that it was merely an associated protein.
Technology: Deep Interactome Profiling by Mass Spectrometry (DIP-MS), including deep-learning-based signal processing
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: In human prefoldin-like-complex experiments, repeated coelution and affinity-purification data supported POLR2E as a constitutive PFDL component, updating the view that it was merely an associated protein.
- Still unknown: The paper proposes POLR2E as a constitutive PFDL subunit based on coelution and affinity-purification in the tested human-cell system; it does not establish that every cellular PFDL complex contains POLR2E or fully characterize its functional contribution. DIP-MS also recapitulated known ASDURF association, which is not counted here as novel because ASDURF’s PAQosome/PFDL role had been reported earlier.
- AI role: Deep-learning signal processing helped resolve coexisting protein-complex isoforms from AP-MS/native PAGE/DIA-MS data. DIP-MS on UXT and PFDN2 baits repeatedly detected PFDL coelution with POLR2E; affinity purification of four tagged PFDL-like subunits also recovered POLR2E, supporting the proposed constitutive association.
Evidence sources: https://www.nature.com/articles/s41592-024-02211-y
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 10% of works of the same type, year and subfield (OpenAlex work metrics, 32 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 2/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A self-supervised contrastive-learning search of Dark Energy Survey imaging isolated a group of 11 sources that included seven known z≈6 quasars. Follow-up Gemini-South and archival NTT spectroscopy confirmed three additional quasars at redshifts 5.94–6.07. One source, J0109−5424, shows evidence of a foreground object and is a lensed-quasar candidate, but its multiple-image status requires high-resolution imaging.
Technology: Self-supervised contrastive-learning neural network trained on Dark Energy Survey imaging data
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: A self-supervised contrastive-learning search of Dark Energy Survey imaging isolated a group of 11 sources that included seven known z≈6 quasars.
- Still unknown: The three quasar identifications are spectroscopically confirmed, but the possible gravitational lens interpretation of J0109−5424 remains unconfirmed pending high-resolution imaging. The reported 91% efficiency refers to ten quasars among the eleven-source island, including seven previously known objects; three of the four previously unclassified targets were confirmed quasars. It is not a universal completeness estimate.
- AI role: Grouped sources into a learned embedding-space island that prioritized four unclassified candidates; spectroscopy confirmed three as quasars, including one source missed by conventional color selection because foreground light contaminated the photometry.
Evidence sources: https://arxiv.org/abs/2403.17903
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: at percentile 73 of works of the same type, year and subfield (OpenAlex work metrics, 11 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 2/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
LSCNN ranked a combinatorial Ugi library after model development on focused activity data. The paper reports six active products among ten selected combinations against MRSA; purified H4–H6 had MIC and MBC of 12 μM. Different screening phases and cutoffs mean these figures are not a randomized general hit-rate comparison.
Technology: LSCNN feature-fusion neural network
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Primary article reports 6/10 selected combinations active against MRSA and purified H4–H6 MIC/MBC results.
- Still unknown: Prior chemical-structure overlap, full independent replication, and clinical efficacy are not established.
- AI role: Model ranks products from combinatorial Ugi space for synthesis and testing.
Evidence sources: https://pubs.rsc.org/en/content/articlepdf/2024/sc/d3sc06441g · https://pmc.ncbi.nlm.nih.gov/articles/PMC11041243/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 16 Nov 2026. Review scope: Primary source checked for identity, date, model role and bounded reported result..
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Researchers used consensus machine-learning classifiers trained on known P2Y6 agonists to select compounds for pharmacological testing. The anticancer drug ABBV-744 showed non-surmountable inhibition of UDP-induced P2Y6 receptor activation and enhanced P2Y1 receptor activation in cellular assays, revealing previously unreported purinergic-receptor activity for this molecule.
Technology: Consensus of deep learning, AdaBoost, naive Bayes, k-nearest neighbors, logistic regression, random forest, SVM and XGBoost classifiers
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Researchers used consensus machine-learning classifiers trained on known P2Y6 agonists to select compounds for pharmacological testing.
- Still unknown: The paper found no novel P2Y6 agonists, which was the initial aim. ABBV-744 inhibited P2Y6 activation at relatively high micromolar concentration (reported IC50 about 75.7 µM); its precise inhibitory mechanism was not determined and requires further assays. P2Y1 enhancement and weak P2Y14 binding effects are assay-specific observations, not clinical pharmacology or therapeutic benefit.
- AI role: Models trained on curated P2Y6 agonist activity data prioritized molecules from compound libraries for experimental screening. Functional calcium-transient assays in receptor-expressing cells and fluorescent binding assays were used to test predictions; ABBV-744 was the principal confirmed hit with unanticipated activity across P2Y receptor subtypes.
Evidence sources: https://link.springer.com/article/10.1007/s11302-024-10003-4
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: at percentile 88 of works of the same type, year and subfield (OpenAlex work metrics, 10 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 2/5 · Validation 3/5 · Novelty 2/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A machine-learning model trained on 71 previously reported nucleoside derivatives was applied to an external set, from which the authors selected 24 candidates for hydrogel testing. They report gel formation for 20 of the 24 tested molecules and characterize cation-independent hydrogel behavior for 8-aminoguanosine and 8-hydroxyguanosine. A 2023 Sichuan University patent by the same team had already disclosed the model, 12 selected candidates and 10 gelation positives, including these two derivatives. This entry represents one study/campaign and does not claim that the 2024 article was the first public disclosure or that the compounds were novel.
Technology: Machine-learning classifier using molecular descriptors to predict nucleoside derivative hydrogel formation
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: The paper reports that 20 of 24 externally selected nucleoside derivatives formed hydrogels in experimental validation. It also reports cation-independent hydrogel behavior for 8-aminoguanosine and 8-hydroxyguanosine.
- Still unknown: A 2023 patent application from the same university/team already disclosed the same model, 12 selected derivatives, 10 gelation positives, and included 8-AG and 8-OHG. Do not treat the overlapping paper and patent as independent discoveries or claim first disclosure/compound novelty. Hydrogel behavior does not establish clinical benefit.
- AI role: The ML model predicted gel-forming ability and selected nucleoside derivatives for external experimental validation.
Evidence sources: https://www.nature.com/articles/s41467-024-46866-9 · https://pmc.ncbi.nlm.nih.gov/articles/PMC10960799/ · https://pubmed.ncbi.nlm.nih.gov/38521777/ · https://patents.google.com/patent/CN116741306A/en
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review date: 17 Nov 2026. Review scope: Primary Nature article page and PubMed Central indexed full text for model selection and experimental outcomes; PubMed record for date and affiliations; published patent application text for same-team prior disclosure and overlapping candidates..
Editorial assessment: Impact 2/5 · Validation 3/5 · Novelty 2/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A model trained on 5,680 gauges improved average global forecast reliability from zero to five days: forecasts at days 4–5 were similar to or better than day-zero forecasts from GloFAS v4. Flood Hub covers more than 80 countries. F1 comparison across 2,092 basins for two-year return-period events; flash and urban floods remain to be covered.
Technology: Hydrological forecasting model
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: F1 scores were compared with GloFAS v4 for 2-year return-period events across 2,092 watersheds (2014-2023); reliability at 4-5 day lead times was similar to or better than GloFAS at day zero.
- Still unknown: Coverage must expand to more locations, flash and urban floods are not yet covered, and the gain is average-level rather than uniform.
- AI role: Two sequential LSTMs with a mixture density output produce probabilistic streamflow forecasts, trained on 5,680 gauges.
Evidence sources: https://research.google/blog/using-ai-to-expand-global-access-to-reliable-flood-forecasts/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: blog. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 580 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 4/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Mutational Effect Transfer Learning (METL) combines transformer pretraining on simulated protein biophysics with fine-tuning on experimental sequence–function data. Using only 64 experimental GFP examples, the authors designed 20 previously untested variants and measured their fluorescence; 16 showed measurable fluorescence, although none was brighter than wild-type GFP. The result demonstrates functional design under a small-data setup, not improved brightness over wild type. The result was present in the bioRxiv version posted 2024-03-17; the GEM@ICLR 2024 workshop paper provides the detailed 20-variant assay report, and the later journal version is the same study.
Technology: Mutational Effect Transfer Learning (METL), a transformer protein language model pretrained on simulated biophysical features and fine-tuned on experimental data
Publication: Preprint · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Preprint
- Test type: direct test
- Tested: Twenty METL-designed GFP variants were synthesized and experimentally measured; 16 showed measurable fluorescence, but none exceeded wild-type brightness.
- Still unknown: The panel contains 20 designed variants in a specific GFP task; the result does not show improvement over wild-type brightness or establish general performance across protein families.
- AI role: METL-Local was pretrained on simulated GFP sequence/biophysical data, fine-tuned using 64 experimental examples, and used with sequence optimization to select variants for wet-lab testing.
Evidence sources: https://www.biorxiv.org/content/10.1101/2024.03.15.585128v1 · https://openreview.net/pdf?id=hWISqnHBr7
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review date: 17 Nov 2026. Review scope: Primary bioRxiv v1 indexed abstract, posting date and affiliation metadata; primary GEM@ICLR workshop paper full text with methods and experimental outcomes..
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Using 632 serendipitously imaged Hubble Space Telescope asteroids found by the prior AI-assisted Hubble Asteroid Hunter survey, the authors fit trail parallax to estimate distances, absolute magnitudes, and approximate sizes. Their cumulative magnitude distribution has a shallower slope (α≈0.26) from H=15 to about H=20, extending the previously reported trend by roughly two magnitudes; the analyzed sample is dominated by potential main-belt objects smaller than 1 km.
Technology: Google Cloud deep-learning asteroid-trail detector trained on Hubble Asteroid Hunter citizen-science classifications (upstream survey)
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: Using 632 serendipitously imaged Hubble Space Telescope asteroids found by the prior AI-assisted Hubble Asteroid Hunter survey, the authors fit trail parallax to estimate distances, absolute magnitudes, and approximate sizes.
- Still unknown: The 632-object sample comprises 454 unmatched and 178 known objects; unmatched does not mean every object is independently confirmed or has a reliable orbit. Only about 40% of raw objects yielded meaningful distance estimates after filtering. Sizes assume spherical objects and a uniform albedo of 0.15. The authors state that the combined archive sample is not a dedicated or observationally unbiased survey, and the result is an extension of the size-distribution slope to H≈20 rather than a complete census.
- AI role: The primary study states that the input asteroid trails were detected in the preceding Hubble Asteroid Hunter work using a deep-learning algorithm trained on volunteer classifications. This AI-assisted search supplied the previously unidentified trail sample; the present paper then applies a parallax-fitting method and magnitude-distribution analysis to characterize it. The quantitative population result depends on this archive-scale AI-assisted detection pipeline and the present physical analysis.
Evidence sources: https://www.aanda.org/articles/aa/pdf/2024/03/aa46771-23.pdf
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: at percentile 74 of works of the same type, year and subfield (OpenAlex work metrics, 4 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 2/5 · Novelty 2/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A machine-learning classifier applied to quantitative PPI assay data and AlphaFold-Multimer predictions to prioritize SARS-CoV-2 protein interactions; 15 of 34 high-confidence interactions were not previously listed in IMEx. The team then targeted the NSP10–NSP16 methyltransferase interface in an ultra-large virtual screen. Compound 459 bound NSP10 (MST affinity about 13 μM), reduced NSP10–NSP16 interaction in cells (BRET IC50 9.2 μM), inhibited complex methyltransferase activity by about 50% at 100 μM, and reduced SARS-CoV-2 reporter replication in cells (IC50 39.5 μM).
Technology: Multi-adaptive support-vector-machine (maSVM) classifier for quantitative protein–protein interaction data, coupled with AlphaFold-Multimer complex prediction; VirtualFlow/QuickVina2 and later docking methods performed compound screening
Family AlphaFold: profile · models page
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: A machine-learning classifier applied to quantitative PPI assay data and AlphaFold-Multimer predictions to prioritize SARS-CoV-2 protein interactions; 15 of 34 high-confidence interactions were not previously listed in IMEx.
- Still unknown: The paper reports an experimental small-molecule hit, not a drug or clinical effect. Compound 459 has micromolar affinity and cellular antiviral IC50; authors explicitly state it requires extensive optimization. Methyltransferase inhibition was about 50% at 100 μM in one initial assay; subsequent data include cellular interaction and reporter-replication experiments. The NSP10–NSP16 interface was previously known and had prior virtual screening, but this paper reports a different compound scaffold and tested outcomes; shared target does not make this a duplicate. In vitro and cell assays do not demonstrate efficacy in animals or people.
- AI role: The maSVM algorithm prioritized PPI interactions using quantitative LuTHy binary-assay data and AlphaFold-Multimer predictions, supporting selection and mapping of SARS-CoV-2 interactions, including the NSP10–NSP16 interface and interface hot spots. The drug-like compound hit itself was identified by VirtualFlow and docking against the selected interface, not generated by the classifier. Compound 459 was then experimentally tested for protein binding, interaction disruption, methyltransferase activity and viral replication.
Evidence sources: https://link.springer.com/article/10.1038/s44320-024-00019-8
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 10% of works of the same type, year and subfield (OpenAlex work metrics, 37 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
MolE-XGBoost prioritized existing compounds. Of six tested commercially available compounds, elvitegravir, opicapone and ebastine showed measured growth-inhibitory effects on S. aureus; visomitin was a known-antimicrobial positive control. Effects varied: elvitegravir inhibited growth at 8 μg/mL, while the other two limited or delayed growth. This is an in-vitro finding, not a clinical antibiotic.
Technology: MolE molecular representation learning; XGBoost
Publication: Preprint · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Preprint
- Test type: direct test
- Tested: The primary source reports the study-specific result summarized here.
- Still unknown: Independent replication and broad deployment are not established by this review.
- AI role: A learned molecular representation and XGBoost rank candidates for experimental screening; selected compounds undergo MIC and growth-curve assays.
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC11986102/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 16 Nov 2026. Review scope: Primary-source indexed article or abstract and bibliographic metadata reviewed; complete source scope is given in source_checks..
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
An autonomous materials-discovery workflow used AI-feedback density-functional-theory simulations to select six previously well-explored ternary oxide systems for synthesis. In the Ca–Ru–O campaign, experiments yielded two previously unreported phases; room-temperature diffraction and follow-up characterization most clearly support an isolated low-temperature calcium-rich phase, approximately Ca1+xRuO3+y, distinct from conventional CaRuO3. The authors state the realized phases were not the specific crystal structures predicted by the simulation.
Technology: CAMD (computer-aided materials discovery), an AI-feedback density-functional-theory phase-search workflow
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: An autonomous materials-discovery workflow used AI-feedback density-functional-theory simulations to select six previously well-explored ternary oxide systems for synthesis.
- Still unknown: One low-temperature phase was isolated and extensively characterized; the second high-temperature phase remained multiphase/metastable and structurally unresolved, so this record rests on the better-supported low-temperature result. Exact composition and location of excess Ca remain under investigation. AI/DFT guided chemical-system selection but did not correctly predict the realized phase structure; authors report only one successful system among six and do not claim AI accelerated the overall process.
- AI role: CAMD mapped candidate phases across chemical systems and informed expert down-selection. Human researchers designed and ran synthesis and characterization. The successful calcium-ruthenate phases emerged experimentally in the AI-prioritized system but did not match the exact predicted crystal structures; authors describe one success among six systems and stress current AI-guided synthesis limitations.
Evidence sources: https://pubs.rsc.org/en/content/articlehtml/2024/sc/d3sc04823c
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 10% of works of the same type, year and subfield (OpenAlex work metrics, 29 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 2/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Researchers developed MutComputeX, a structure-based residual neural network for assessing amino-acid substitutions in protein–ligand environments, and used it with a 4′-O-methylnorbelladine biosensor to engineer the plant methyltransferase Nb4OMT in Escherichia coli. The paper reports engineered variants with 60% higher product titer, twofold higher catalytic activity, threefold lower off-product regioisomer formation and 5 °C higher thermostability; ligand-aware predictions helped identify the beneficial A53M substitution. These enzyme and product results do not demonstrate complete microbial production of galantamine.
Technology: MutComputeX structure-based residual neural network for protein–ligand mutation design
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: MutComputeX-guided, manually curated Nb4OMT variants were assayed using a 4NB biosensor and biochemical methods. The article reports 60% higher product titer, twofold higher catalytic activity, threefold lower off-product formation and 5 °C higher thermostability for engineered variants.
- Still unknown: The result concerns a pathway enzyme and branchpoint intermediate in the reported microbial/biochemical systems; full pathway production of galantamine and industrial manufacturing were not demonstrated.
- AI role: MutComputeX predicted sequence substitutions in a modeled protein–ligand microenvironment and guided selection of Nb4OMT variants for biosensor screening and enzyme assays; ligand-aware predictions supported A53M.
Evidence sources: https://www.nature.com/articles/s41467-024-46356-y
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review date: 17 Nov 2026. Review scope: Primary Nature Communications full text as indexed, including methods/results, author contributions and affiliation block..
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A machine-learning-generated MinE homolog, synMinEv25, restored Min oscillations and wild-type-like cell growth and morphology when substituted for the native minE gene in an E. coli strain lacking minDE. Researchers generated variants with a multiple-sequence-alignment variational autoencoder, used AlphaFold2-based structural scoring to select candidates, and screened them in synthetic-cell mimics before in-vivo testing. The validated variant is a distant homolog with less than 50% sequence identity to E. coli MinE.
Technology: Multiple-sequence-alignment variational autoencoder (MSA-VAE), AlphaFold2 Multimer, and a structure-based divide-and-conquer scoring pipeline
Family AlphaFold: profile · models page
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Researchers generated variants with a multiple-sequence-alignment variational autoencoder, used AlphaFold2-based structural scoring to select candidates, and screened them in synthetic-cell mimics before in-vivo testing.
- Still unknown: This is a proof-of-principle result in one engineered E. coli MinDE system, not a general demonstration that generated proteins can replace natural genes. The paper notes weak correlation between several in-silico scores and individual measured protein properties; it reports one best-performing variant and does not establish applications outside this assay.
- AI role: The MSA-VAE generated MinE homolog sequences; structure-based scoring estimated membrane binding, MinD complex formation and dimerization to prioritize 48 candidates from 4,000 generated sequences. Cell-free synthetic-cell assays identified 14 variants with Min oscillations, and researchers tested these in E. coli; synMinEv25 restored oscillations, growth and morphology. AI generated and prioritized candidates, while wet-lab experiments established function.
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC10914801/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 10% of works of the same type, year and subfield (OpenAlex work metrics, 23 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
CoSaNN combined neural structure prediction, learned sequence optimization and solubility ranking to design variants of an induced-fit ROK carbohydrate kinase. The authors measured expression, thermal stability and activity in vitro. The paper reports measurable expression for 189 of 348 designs in one table, while the strategy-level counts do not sum to that total; no success-rate claim is made here.
Technology: CoSaNN: AlphaFold2-based conformation modeling, ProteinMPNN sequence design and SolvIT graph neural network ranking
Family AlphaFold: profile · models page
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: The authors applied CoSaNN to redesign the Arthrobacter polyphosphate glucokinase ROK kinase and performed expression, thermal-stability and enzyme-activity assays on designed variants. The primary Results table reports 189 designs with detectable soluble expression among 348 listed designs; its strategy counts (119+95+94 plus 40 in another strategy; rows in the prose/table are internally inconsistent) do not reconcile cleanly, so the record does not use a single total as a success-rate denominator.
- Still unknown: The study tests a company-designed enzyme set in vitro; it does not establish industrial deployment. The primary paper’s count inconsistency remains explicitly disclosed.
- AI role: CoSaNN uses deep-learning structure prediction (AlphaFold2), sequence optimization with ProteinMPNN and a learned solubility classifier (SolvIT) to rank enzyme designs for experimental characterization.
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC10945820/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review date: 17 Nov 2026. Review scope: Primary PMC full text and PNAS DOI record for CoSaNN-guided glucokinase design, Enzymit affiliations, experimental activity results and the discrepancy in reported design totals..
Editorial assessment: Impact 2/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Regression and k-nearest-neighbor models with a genetic algorithm guided optimization from a shrimp lipopolysaccharide-binding peptide to LBD_A–D. The article reports in-vitro tests, NMR structure for LBD_B and improved survival in a kuruma prawn co-injection experiment. This is not human therapy and does not establish treatment efficacy after infection.
Technology: Regression and k-nearest-neighbor models; genetic algorithm
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Article reports synthesized/tested peptide variants, NMR characterization and a prawn co-injection assay.
- Still unknown: Full exact assay/organism scope and peptide sequence novelty/independent replication not exhaustively compared.
- AI role: Activity models and genetic search guided sequence variants which were synthesized and biologically tested.
Evidence sources: https://doi.org/10.1016/j.jare.2024.02.016 · https://pubmed.ncbi.nlm.nih.gov/38431124/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 16 Nov 2026. Review scope: Primary source checked for identity, date, model role and bounded reported result..
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A six-round engineering campaign combined mutational scans, structure-guided libraries and Gaussian-process models. The model-aided library yielded M3, reported with 22-fold higher apparent activity than wild type in the UV assay. Three subsequent conventional mutagenesis rounds produced M6, a 10-substitution descendant with 64-fold higher apparent kcat, but also about 12-fold higher apparent KM; do not attribute M6 as a direct model output or treat kcat as production yield. The authors report a 100 g/L laboratory process test with ≥98% conversion and 99.7% diastereomeric excess after 30 hours. These are process-relevant laboratory results, not a commercial production run or drug efficacy.
Technology: Gaussian-process regression
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: M3 and later M6 were tested for activity; article reports preparative reactions at 100 g/L substrate and conversion/stereoselectivity under specified laboratory conditions.
- Still unknown: M6 follows conventional site-mutagenesis rounds after the AI-informed step; apparent kcat differs from process yield and KM changed; no commercial production/clinical efficacy claim.
- AI role: Gaussian-process predictions informed smaller follow-up variant libraries within a broader workflow that also used mutational scanning, rational design and conventional evolution.
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC10902378/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review date: 16 Nov 2026. Review scope: Primary PMC full article, methods/results, publication history, affiliations and chronology of model and conventional rounds..
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
ModelAngelo combines cryo-EM density, protein sequence and local geometry to build atomic models. Beyond method benchmarks, the paper identifies six previously unassigned proteins in a photosynthetic supercomplex and four additional radial-spoke proteins plus two central-apparatus proteins in Chlamydomonas reinhardtii. These assignments use existing experimental cryo-EM maps and sequence searches; the paper leaves the newly assigned ciliary proteins' functional roles for further investigation.
Technology: ModelAngelo graph neural network with hidden Markov model sequence search
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: Models built by ModelAngelo were compared with those of human experts, and its protein identification was compared with human experts on proteins of unknown sequence.
- Still unknown: Not reported in the abstract read; the nucleotide result is limited to backbone accuracy.
- AI role: A graph neural network combining cryo-EM density, protein sequence and local geometry to build atomic models and identify proteins.
Evidence sources: https://www.ebi.ac.uk/europepmc/webservices/rest/search?query=DOI:10.1038/s41586-024-07215-4&resultType=core&format=json
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 713 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A deep-learning design pipeline combining sequence generation and classifiers, transfer learning, pretrained protein embeddings and C. acnes-specific training data produced 42 novel peptides for testing. Five showed high-potency, selective activity against C. acnes, with MICs of 2–4 µg/mL; additional in-vitro assays tested activity against other microbes, hemolysis and keratinocyte cytotoxicity.
Technology: Deep-learning peptide generators and activity/selectivity classifiers using transfer learning and pretrained protein embeddings
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: A deep-learning design pipeline combining sequence generation and classifiers, transfer learning, pretrained protein embeddings and C. acnes-specific training data produced 42 novel peptides for testing.
- Still unknown: All evidence is in vitro, with no animal or human testing. The 42 synthesized candidates are a selected subset of model output, and the abstract's five high-potency peptides have measured C. acnes MICs; activity against other organisms and toxicity is based on separate in-vitro assays. Do not imply topical efficacy or clinical safety.
- AI role: The models generated peptide sequences, ranked antimicrobial/hemolytic properties and incorporated C. acnes-focused training data. The team synthesized and tested 42 selected designs; five had reported C. acnes MICs of 2–4 µg/mL.
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC10894229/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 10% of works of the same type, year and subfield (OpenAlex work metrics, 22 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 2/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Kernel ridge regression reconstructed present and future air–sea CO₂-flux variability in five Earth-system models. The analysis found dissolved inorganic carbon (DIC) and alkalinity to be the main drivers, while DIC’s relative importance generally declines in warming scenarios—especially in the equatorial Pacific, Southern Ocean and North Atlantic—as anthropogenic carbon uptake flattens vertical DIC gradients. This result is conditional on the models and scenarios analyzed.
Technology: Kernel ridge regression and machine-learning clustering/classification
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: Kernel ridge regression reconstructed present and future air–sea CO₂-flux variability in five Earth-system models.
- Still unknown: This is an analysis of outputs from five Earth-system models and four historical/future scenarios, not a measurement of future ocean conditions. The interpretation that anthropogenic-carbon-driven gradient flattening contributes to reduced DIC importance is the authors’ proposed mechanism; relative predictor importance does not alone establish causal attribution.
- AI role: Kernel ridge regression reconstructed CO₂-flux variability from Earth-system-model variables; machine-learning analysis compared predictor importance among spatial clusters and climate scenarios to identify the changing roles of DIC and alkalinity.
Evidence sources: https://www.nature.com/articles/s43247-024-01257-2
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: at percentile 88 of works of the same type, year and subfield (OpenAlex work metrics, 13 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
In a multicenter study of 702 participants, targeted plasma metabolomics and machine learning produced a 10-metabolite gastric-cancer diagnostic model and separate 28-metabolite survival-risk model. The diagnostic model was evaluated in an independent 95-person cohort (AUROC 0.920; sensitivity 0.905, specificity 0.75) and detected stage IA disease with AUROC 0.909. Among 181 patients followed for a median 40 months, the prognostic model achieved test-set concordance index 0.83. Succinate, uridine and lactate were among diagnostic features.
Technology: LASSO, random forest and random survival forest
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: The diagnostic model was evaluated in an independent 95-person cohort (AUROC 0.920; sensitivity 0.905, specificity 0.75) and detected stage IA disease with AUROC 0.909.
- Still unknown: Case-control cohorts do not establish population screening performance. External diagnostic test n=95; prognosis cohort median follow-up 40 months with right-censored outcomes. The paper says direct clinical use is premature and requires quality-control samples. Associations are not causal and no model-guided benefit was tested.
- AI role: LASSO selected ten plasma metabolites for a diagnostic model evaluated in internal and independent external cohorts. Random survival forest selected a 28-metabolite prognostic panel, evaluated in a held-out patient set.
Evidence sources: https://www.nature.com/articles/s41467-024-46043-y.pdf
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 289 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A model trained only on past experimental data predicts instabilities up to 300 ms ahead, and an AI controller adjusted plasma shape and beam power to avoid them in a real DIII-D experiment. The authors call it a proof of concept on one machine and one type of instability; paper published in Nature on February 21, 2024.
Technology: Prediction and reinforcement-learning control
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: The reinforcement-learning controller was tested in experiments on the DIII-D tokamak, where it kept the tearing likelihood below a threshold, including under unfavorable conditions.
- Still unknown: The abstract does not state the prediction horizon and lists no other limitations; tearing instability is hard to forecast.
- AI role: A deep reinforcement learning controller, trained in a multimodal dynamic model of tearing instability, adjusted plasma operation to keep it stable.
Evidence sources: https://www.ebi.ac.uk/europepmc/webservices/rest/search?query=DOI:10.1038/s41586-024-07024-9&resultType=core&format=json
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: press release. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 195 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 4/5 · Validation 3/5 · Novelty 4/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Random-forest models trained on UCSF records (749 expert-diagnosed Alzheimer’s disease cases and 250,545 controls) identified clinical features predictive up to seven years before the recorded onset; the seven-year model’s AUROC was 0.72 using clinical features alone. Matched-cohort follow-up in external University of California health-system records found faster Alzheimer’s diagnosis after osteoporosis in women. Genetic colocalization analyses supported a female-specific shared signal between Alzheimer’s disease, low heel bone mineral density and MS4A6A expression near the MS4A gene cluster. The paper also reports shared APOE-locus support for hyperlipidemia and Alzheimer’s disease.
Technology: Random-forest models on longitudinal electronic health records, interpreted with the SPOKE biomedical knowledge network
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: Matched-cohort follow-up in external University of California health-system records found faster Alzheimer’s diagnosis after osteoporosis in women.
- Still unknown: This is hypothesis-generating observational work. EHR features may reflect healthcare access, clinician behavior, coding and biological processes; the authors explicitly state that further validation across omics data types is needed. Genetic colocalization and matched cohort associations do not establish that osteoporosis causes Alzheimer’s disease or that treating osteoporosis prevents it. The prediction model is not a prospective clinical-impact evaluation, and early risk discrimination is modest with the low-prevalence outcome.
- AI role: Random-forest models ranked EHR-derived clinical features associated with future Alzheimer’s diagnosis. The team used matched models to prioritize hypotheses, then used SPOKE to connect clinical predictors to biological networks and independent health-system data plus GWAS/eQTL colocalization to examine the osteoporosis and hyperlipidemia associations. The model-generated prioritization led to follow-up of the female-specific osteoporosis signal near MS4A6A.
Evidence sources: https://www.nature.com/articles/s43587-024-00573-8.pdf
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 101 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A machine-learning workflow guided virtual screening for parasite N-myristoyltransferase, followed by testing of selected compounds in parasite assays. The primary preprint reports a bounded set of measured antimalarial candidates, including LabMol-394, while target assignment remains computational. This is not clinical efficacy or proof of target engagement.
Technology: Selective-bias virtual screening; machine-learning target/activity models
Publication: Preprint · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Preprint
- Test type: direct test
- Tested: Primary preprint full text reports tested selected candidates and parasite activity.
- Still unknown: Exact earliest journal version and compound prior art remain open; target-specific binding is not experimentally established.
- AI role: Model-based selective-bias workflow scores target/activity candidates for experimental testing.
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC10925453/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 16 Nov 2026. Review scope: PMC-indexed primary preprint full text; PubMed primary record/abstract and affiliations. Material checked for the specific claims stated in the record..
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
An interpretable machine-learning analysis of multi-omic systemic-sclerosis data identified an unexplored keratinocyte-centered signature, validated by protein staining, and a previously unreported interaction between myeloid-cell and fibroblast states and macrophage HLA signaling. The same study also analyzed CD4 T-cell clonal expansion, but this record treats the paper's disease-mechanism findings as one discovery.
Technology: SLIDE (Significant Latent Factor Interaction Discovery and Exploration)
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: An interpretable machine-learning analysis of multi-omic systemic-sclerosis data identified an unexplored keratinocyte-centered signature, validated by protein staining, and a previously unreported interaction between myeloid-cell and fibroblast states and macrophage HLA signaling.
- Still unknown: The paper reports computationally identified relationships and validation of one signature by protein staining; these findings do not by themselves establish causal mechanisms or clinical utility.
- AI role: SLIDE integrated single-cell and spatial multi-omic data, identified interacting latent factors associated with systemic-sclerosis disease severity, and surfaced candidate mechanisms; the keratinocyte signal was checked with protein staining by the study team.
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC11588359/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 10% of works of the same type, year and subfield (OpenAlex work metrics, 37 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
SLIDE, an interpretable latent-factor interaction method, was applied to paired single-cell RNA-seq and T-cell receptor data from islet-derived CD4 T cells in NOD mice. The analysis identified latent components associated with clonal expansion and grouped inhibitory-receptor and intracellular signaling/apoptosis-related genes among the associated programs. The result is an inference from existing multiomic data, not a causal perturbation; it is a distinct T1D application in the same article as the separately catalogued systemic-sclerosis analysis.
Technology: SLIDE interpretable latent-factor interaction discovery and exploration
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: The method was applied to paired single-cell RNA-seq/TCR-seq from islet-derived CD4 T cells in NOD mice; the analysis reports latent components associated with clone expansion and compares the findings against an independent published dataset and alternative methods.
- Still unknown: The inferred gene programs are associations from existing data; the T1D analysis does not prospectively perturb the proposed pathways or establish causality.
- AI role: SLIDE uses interpretable machine learning to identify interactions among latent factors in high-dimensional omics data and was applied to connect T-cell gene programs with clonal expansion.
Evidence sources: https://www.nature.com/articles/s41592-024-02175-z · https://pubmed.ncbi.nlm.nih.gov/38374265/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review date: 17 Nov 2026. Review scope: Primary Nature Methods indexed full article, including T1D application, Methods, results, author affiliations and publication/version dates; PubMed primary abstract and affiliation metadata..
Editorial assessment: Impact 3/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
Directed relationships from this entry; they do not merge catalog counts.
A correlator convolutional neural network trained on projective snapshots from numerical simulations of the Kitaev–Heisenberg model identified a characteristic correlation pattern in the intermediate gapless phase when analyzing spin snapshots perpendicular to the applied field. The authors interpret the learned pattern as Friedel oscillations of gapless spinons forming a Fermi surface, supporting earlier theoretical proposals for this phase and suggesting measurable correlations for future experiments.
Technology: Quantum-classical QuCl approach using a correlator convolutional neural network and regularization-path analysis
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: A correlator convolutional neural network trained on projective snapshots from numerical simulations of the Kitaev–Heisenberg model identified a characteristic correlation pattern in the intermediate gapless phase when analyzing spin snapshots perpendicular to the applied field.
- Still unknown: This is a numerical/theoretical result based on DMRG wavefunctions and simulated snapshots, not an experimental observation of a quantum spin liquid. The authors interpret the learned motif as evidence consistent with a spinon Fermi surface and offer it as a prediction to guide experiments; they do not claim it is already experimentally confirmed.
- AI role: Analyzed simulated projective-measurement snapshots and surfaced the phase signature: spin correlations in the direction perpendicular to the magnetic field, interpreted as the spatial pattern of Friedel oscillations associated with gapless spinons.
Evidence sources: https://www.nature.com/articles/s42005-024-01542-8
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: at percentile 60 of works of the same type, year and subfield (OpenAlex work metrics, 8 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 2/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A self-supervised, attention-based deep-learning regression model predicted continuous homologous-recombination deficiency (HRD) scores and tumor-microenvironment biomarkers from H&E pathology slides. On TCGA internal validation, HRD classification AUROC exceeded 0.70 in five of seven cancer types; external CPTAC validation reached AUROC 0.81 for lung adenocarcinoma and 0.96 for endometrial cancer. In 2,297 colorectal-cancer patients in the DACHS cohort, continuous regression scores for leukocyte fraction and lymphocyte-infiltration-related signatures were independently prognostic in multivariable survival analysis.
Technology: CAMIL regression, a self-supervised contrastive-clustering and attention-based multiple-instance deep-learning model
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: On TCGA internal validation, HRD classification AUROC exceeded 0.70 in five of seven cancer types; external CPTAC validation reached AUROC 0.81 for lung adenocarcinoma and 0.96 for endometrial cancer.
- Still unknown: Retrospective computational pathology analysis; the model predicts molecular and immune-related biomarkers from slides and does not replace confirmatory molecular testing or establish treatment selection benefit. HRD performance varied by cancer type, with weak internal performance in some cohorts. The reported survival associations in DACHS are prognostic associations, not evidence that using model predictions improves patient outcomes. Scores and endpoints were tested in the named datasets.
- AI role: The model analyzed whole-slide H&E images to estimate continuous HRD and tumor-microenvironment biomarker scores. The team compared regression with classification methods across cohorts, externally validated HRD predictions in CPTAC, and tested prognostic associations of predicted scores in a separate retrospective colorectal-cancer cohort.
Evidence sources: https://www.nature.com/articles/s41467-024-45589-1.pdf
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 126 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 2/5 · Novelty 2/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A programmable robotic synthesis platform used inline HPLC, Raman or NMR feedback and closed-loop optimization to explore selected reaction spaces. In a tosMIC campaign, sequential model-based optimization with a Gaussian-process surrogate identified a reaction between toluenesulphonylmethyl isocyanide and benzylidenemalononitrile; subsequent optimization increased yield from the 27% initial condition to 49% (22 percentage points) over 50 reactions. The article also reports a second previously unreported reaction and an exploratory trifluoromethylation campaign that discovered new molecules.
Technology: ChemputationOptimizer, a sensor-integrated reaction engine with sequential model-based optimization (Gaussian-process surrogate); the platform also exposes random search, design-of-experiments, Bayesian-optimization and genetic-algorithm routines
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: The article also reports a second previously unreported reaction and an exploratory trifluoromethylation campaign that discovered new molecules.
- Still unknown: The platform's reported yield gains are specific to these selected reaction spaces and objectives, not general chemistry performance. The tosMIC and phloroglucinol campaigns are reported as previously unreported reactions; product characterization and complete procedure details are in the article/supplement. The system includes several optimization algorithms and the authors note random search can be a strong baseline; the reported tosMIC optimization used SMBO with a Gaussian-process surrogate. No claim is made that the reactions are industrially useful or broadly generalizable. Primary source checked; no independent replication assessed.
- AI role: For the tosMIC and phloroglucinol explorations, machine-guided combinatorial search proposed reaction parameters and inline spectroscopic feedback supplied product-yield measurements; a Gaussian-process sequential-model-based optimizer guided the reaction-space exploration and follow-up optimization. Human chemists defined reaction space and experimental constraints, and performed/characterized the experiments.
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC10858227/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 10% of works of the same type, year and subfield (OpenAlex work metrics, 61 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Researchers combined extensive X-ray diffraction with X-ray Temperature Clustering (X-TEC) to identify Bragg-glass signatures in the charge-density-wave material PdₓErTe₃. Their analysis extracted diffraction features across about 20,000 Brillouin zones and found a diverging correlation length and a phase diagram consistent with a Bragg-glass transition. The result provides bulk diffraction evidence for a phase that had been difficult to establish experimentally.
Technology: X-ray Temperature Clustering (X-TEC), a machine-learning analysis of X-ray diffraction peaks
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Researchers combined extensive X-ray diffraction with X-ray Temperature Clustering (X-TEC) to identify Bragg-glass signatures in the charge-density-wave material PdₓErTe₃.
- Still unknown: The authors frame the result as evidence for the Bragg-glass phase based on characteristic diffraction behavior, rather than direct imaging of all underlying defects. Machine learning enabled the scale of peak extraction; physical interpretation and measurements were done by the team.
- AI role: Automatically extracted and clustered charge-density-wave diffraction peaks across temperature and momentum, enabling high-throughput estimation of peak spread/correlation length from a large set of Brillouin zones; researchers interpreted these features as evidence for Bragg-glass order.
Evidence sources: https://www.nature.com/articles/s41567-023-02380-1
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: at percentile 89 of works of the same type, year and subfield (OpenAlex work metrics, 18 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
The $700,000 prize went to a submission with 11 additional columns and more than 2,000 characters (about 5% of the first scroll), using models to detect ink and automated segmentation. Papyrologists evaluated 15 anonymized columns and recovered more text from the winning submission. Segmentation remains largely manual and costly.
Technology: Ink detection with vision models and automated segmentation
Publication: Announcement only · Test context: prospective or independent check. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Announcement only
- Test type: prospective or independent check
- Tested: Papyrologists independently evaluated 15 anonymized columns, and the winning submission yielded the most text for every reviewer, meeting the stated criteria.
- Still unknown: The comprehensive study is not yet published, only rough draft transcriptions exist, segmentation is largely manual, and about 5% of the first scroll has been read.
- AI role: Models detected ink in CT scans (a TimeSformer-based model was strongest), and an automated tool produced the segmentation of the papyrus surfaces.
Evidence sources: https://scrollprize.org/grandprize
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: press release. Paper: not linked in this catalog; this does not mean no paper exists.
Review scope and date are not recorded; source checked does not establish full-text review.
Editorial assessment: Impact 5/5 · Validation 4/5 · Novelty 5/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
LogiMML, trained on high-content images of human cardiac fibroblasts, predicted that Src inhibition suppresses actin–myosin stress fibers through PI3K signaling. Follow-up PI3K inhibition reduced stress-fiber formation and procollagen I production in the fibroblasts, supporting the proposed pathway mechanism.
Technology: LogiMML, a logic-based mechanistic network coupled to ridge regression
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: LogiMML, trained on high-content images of human cardiac fibroblasts, predicted that Src inhibition suppresses actin–myosin stress fibers through PI3K signaling.
- Still unknown: The validation is in vitro in primary human ventricular cardiac fibroblasts; it supports a pathway contribution to measured cell phenotypes, not therapeutic efficacy in an animal model or patients. The model's PI3K effect is described as partial.
- AI role: LogiMML fitted the relationship between high-content microscopy phenotypes and perturbations to an experimentally informed fibroblast signaling network, then predicted pathway mechanisms for pirfenidone and the Src inhibitor WH-4-023. Researchers tested the PI3K component of the Src-inhibitor prediction with PI3K inhibition in human cardiac fibroblasts.
Evidence sources: https://www.pnas.org/doi/pdf/10.1073/pnas.2303513121
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 10% of works of the same type, year and subfield (OpenAlex work metrics, 23 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 2/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Elastic-net and multilayer-perceptron models integrated genome-wide gene-deletion data and simulated fluxes to predict beneficial and detrimental reactions for E. coli growth on 30 carbon sources. Beyond training-set examples, the authors experimentally validated model predictions about the glyoxylate shunt, pyruvate dehydrogenase, and redundant purine-biosynthesis reactions. The findings showed that effects of energy-generating pathways depend on carbon source, while biosynthetic pathways were more generally growth-promoting across the tested conditions.
Technology: Elastic-net regression and multilayer-perceptron models
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Beyond training-set examples, the authors experimentally validated model predictions about the glyoxylate shunt, pyruvate dehydrogenase, and redundant purine-biosynthesis reactions.
- Still unknown: Reported reaction effects are for E. coli under the tested carbon-source and genetic-perturbation conditions; they do not establish universal metabolic-control rules. The results should not be interpreted as direct clinical or industrial outcomes. Primary PubMed abstract and PMC article record checked; no independent replication assessed.
- AI role: The models combined gene-deletion phenotypes with simulated metabolic fluxes to prioritize reactions whose perturbation would benefit or harm growth across carbon conditions; selected out-of-training-set predictions were experimentally tested.
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC10912204/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: at percentile 85 of works of the same type, year and subfield (OpenAlex work metrics, 16 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Researchers trained machine-learning models on 288 modularly synthesized ruthenium arene Schiff-base complexes and used them to prioritize 54 new compounds for antibacterial-activity testing. The model-selected set had a 53.7% hit rate against methicillin-resistant Staphylococcus aureus, compared with 9.4% in the original library, an approximately 5.7-fold enrichment.
Technology: Machine-learning activity-prediction models trained on ruthenium arene Schiff-base complex structures and antibacterial assay data
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Researchers trained machine-learning models on 288 modularly synthesized ruthenium arene Schiff-base complexes and used them to prioritize 54 new compounds for antibacterial-activity testing.
- Still unknown: This is a study-specific in-vitro antibacterial hit-rate enrichment, not animal efficacy, clinical validation, or an approved antibiotic. The comparison is 54 model-selected compounds vs the 288-complex source library, not a randomized prospective comparison with all possible metal complexes.
- AI role: Models were trained using assay data from 288 modularly synthesized ruthenium complexes and used to predict activity for a new set of 54 compounds. Those prioritized compounds were then assessed for antibacterial activity; the paper reports higher MRSA hit-rate in the model-selected set than in the original library.
Evidence sources: https://pubmed.ncbi.nlm.nih.gov/38088924/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 10% of works of the same type, year and subfield (OpenAlex work metrics, 50 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 2/5 · Validation 3/5 · Novelty 2/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Across two UC San Diego Health emergency departments, 6,217 adult septic patients were studied before and after nurse-facing alerts driven by the COMPOSER model. Deployment was associated with a 1.9-point absolute (17% relative) drop in in-hospital sepsis mortality (95% CI 0.3-3.5%) and a 5-point rise in bundle compliance. The design is quasi-experimental, not randomized, and several authors co-founded a company that commercializes the tools.
Technology: COMPOSER (deep learning sepsis prediction model)
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: A quasi-experimental before-and-after comparison across two UC San Diego emergency departments (6,217 septic patients) measured in-hospital sepsis mortality and bundle compliance before and after COMPOSER-driven nurse alerts.
- Still unknown: The design is quasi-experimental, not randomized; other limitations were not reported in the fetched abstract.
- AI role: The COMPOSER deep-learning sepsis model triggered nurse-facing alerts.
Evidence sources: https://www.ebi.ac.uk/europepmc/webservices/rest/search?query=DOI:10.1038/s41746-023-00986-6&resultType=core&format=json
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 207 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 5/5 · Validation 2/5 · Novelty 4/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A deep-learning-based catalog pipeline processed 18 months of seismic-array data and located 1,891 shallow earthquakes, more than twice the 849 events in the prior manual catalog. The denser catalog shows that shallow Kabaw Fault seismicity disappears south of about 22.8°N while deeper events occur west of the fault, and it fills part of a previously under-sampled seismicity gap; the authors interpret this contrast as potentially reflecting along-strike variation in deformation associated with oblique Indian-plate subduction.
Technology: DEPALORE pipeline using Earthquake Transformer, transfer-learned PickNet, REAL association, NonLinLoc location and HypoDD relocation
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: A deep-learning-based catalog pipeline processed 18 months of seismic-array data and located 1,891 shallow earthquakes, more than twice the 849 events in the prior manual catalog.
- Still unknown: The source reports catalog-derived patterns and offers a tentative tectonic explanation; the catalog contains 1,891 localized events and location uncertainties of several kilometers. The AI enabled detections and resolution, while geological interpretation remains the authors’ analysis.
- AI role: Deep-learning detectors and phase-picking models extracted seismic arrivals from continuous array waveforms; association and location stages produced the higher-completeness catalog from which researchers interpreted fault patterns.
Evidence sources: https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2023GL105159
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 10% of works of the same type, year and subfield (OpenAlex work metrics, 16 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 2/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A language model proposing constructions and a symbolic engine deriving proofs solved 25 of 30 problems in the IMO-AG-30 benchmark (the previous method solved 10; the average gold medalist score was 25.9), trained on 100 million synthetic examples. It covers geometry only and a 30-problem benchmark.
Technology: AlphaGeometry (language model and symbolic engine)
Family AlphaGeometry: profile · models page
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: retrospective or benchmark
- Tested: On the IMO-AG-30 benchmark of 30 Olympiad geometry problems, the system solved 25 versus 10 for the previous method, with each solution checked by computer.
- Still unknown: It covers geometry only, roughly a third of Olympiad problems, and approaches but does not match gold-medalist performance.
- AI role: A language model proposes auxiliary constructions and a symbolic engine derives proofs, looping until a proof is found.
Evidence sources: https://deepmind.google/blog/alphageometry-an-olympiad-level-ai-system-for-geometry/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: blog. Paper: linked separately.
Review date: 16 Nov 2026. Review scope: Original institutional announcement checked for the specific results, AI contribution and limitations cited in the antibiotic dossier or mathematical follow-up pilot. Full-paper reanalysis, external replication and an exhaustive follow-up search were not performed..
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 494 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 2/5 · Novelty 4/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
- same-team extension · 25 Jul 2024: DeepMind describes AlphaGeometry 2 as an improved version used for the IMO 2024 geometry problem. The combined result also used AlphaProof; it does not belong to AlphaGeometry alone. Source. Recorded 16 Nov 2026.
Follow-up checked 16 Nov 2026 within scope: Bounded check of the original announcements cited by the events recorded here, including explicit links to later program work. This is not an exhaustive literature search or ongoing monitoring commitment..
No source-supported relationships recorded.
Researchers developed a fragmentation strategy for AlphaFold-Multimer and applied it to 62 human protein interactions associated with neurological disorders. Eighteen predicted structures were judged correct or likely correct, and experiments corroborated six previously unreported interfaces: FBXO23–STX1B, STX1B–VAMP2, ESRRG–PSMC5, PEX3–PEX19, PEX3–PEX16 and SNRPB–GIGYF1.
Technology: AlphaFold-Multimer with a protein-fragmentation strategy
Family AlphaFold: profile · models page
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Eighteen predicted structures were judged correct or likely correct, and experiments corroborated six previously unreported interfaces: FBXO23–STX1B, STX1B–VAMP2, ESRRG–PSMC5, PEX3–PEX19, PEX3–PEX16 and SNRPB–GIGYF1.
- Still unknown: The article reports 18 correct or likely correct models among 62 interactions and experimental support for six interfaces; these are not a comprehensive interactome map. AlphaFold-Multimer showed limited specificity and was sensitive to input-fragment length, as the source explicitly notes.
- AI role: AlphaFold-Multimer predicted domain–motif interface structures; researchers tested predictions and used the model-informed fragmentation strategy to prioritize interaction interfaces. Wet-lab assays corroborated six novel contacts, while the tool had limited specificity and performance declined with longer/full-length input sequences.
Evidence sources: https://link.springer.com/article/10.1038/s44320-023-00005-6
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 119 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
The authors combined automatic feature engineering with active learning and high-throughput experiments for oxidative coupling of methane. Of 36 selected catalysts, 30 reached combined ethane and ethylene yields of at least 15% in the reported tests. This is a small-data, reaction-specific catalyst screening outcome; the paper also tests feature engineering on other catalysis datasets. A ChemRxiv version predates the journal publication, and this record does not claim journal-version priority.
Technology: Automatic feature engineering with active learning
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: The article reports 36 selected oxidative coupling of methane catalysts experimentally assessed; 30 achieved combined ethane and ethylene yields of at least 15% under the reported test conditions.
- Still unknown: The paper also appeared as an earlier ChemRxiv version; this check did not establish whether every quantitative detail appeared in that version. The result is limited to the reported catalyst set and reaction conditions.
- AI role: Automatic feature engineering extracted catalyst descriptors, and active learning selected catalysts for high-throughput experimental testing.
Evidence sources: https://d-nb.info/1324333405/34 · https://www.nature.com/articles/s42004-023-01086-y
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review date: 17 Nov 2026. Review scope: Full published article text in German National Library repository and publisher-indexed primary metadata/abstract; earlier ChemRxiv version noted but not compared..
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
With the PNNL laboratory, Microsoft screened 32.6 million materials using AI models and high-performance computing, narrowing them to 500,000, then 800, 150, and finally an electrolyte replacing some lithium with sodium (about 70% less lithium). PNNL synthesized it and built a working solid-state battery. Validation and optimization are ongoing; the post is by a Microsoft executive.
Technology: Azure Quantum Elements (AI and DFT simulations)
Publication: Preprint · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Preprint
- Test type: direct test
- Tested: The top candidates in the Na_xLi_(3-x)YCl6 series were synthesized and experimentally characterized for structure and conductivity; the abstract does not mention battery cell testing.
- Still unknown: The approximately half-million materials are predictions of potential stability, other candidates are still under investigation, and no explicit limitations are listed.
- AI role: AI and physics-based models on cloud HPC screened over 32 million candidates and identified 18 promising electrolyte candidates.
Evidence sources: https://arxiv.org/abs/2401.04070
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: blog. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
A model applied to about 2 million GB of Sentinel-1 radar imagery (2017–2021) maps vessels and infrastructure: about 75% of industrial fishing vessels are not publicly tracked, over 25% of shipping and energy activity is missing from AIS data, wind turbines more than doubled, and global fishing fell 12% during COVID. Published in Nature; ‘fishing’ classification is inferred; source is the organization’s own release.
Technology: Deep learning on radar imagery
Publication: Peer-reviewed paper · Test context: Test context unspecified in annotation. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: not reported
- Tested: The page describes no ground-truth check, accuracy metric or comparison for the vessel and infrastructure mapping, so the main claim is not evaluated in the source.
- Still unknown: Not all boats broadcast their position, the 'fishing' classification is inferred, and the page gives no accuracy figures for that step.
- AI role: Machine learning detected vessels and offshore infrastructure in satellite imagery and inferred which non-broadcasting vessels were likely fishing; the model type is not named.
Evidence sources: https://globalfishingwatch.org/press-release/new-research-harnesses-ai-and-satellite-imagery-to-reveal-the-expanding-footprint-of-human-activity-at-sea/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: press release. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 314 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 4/5 · Validation 1/5 · Novelty 4/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
In a proteomic study, researchers measured paired CSF and serum samples in a discovery cohort and used random-forest machine learning to select a 19-protein CSF panel and an eight-protein serum panel to distinguish mild cognitive impairment due to Alzheimer disease from cognitively normal participants. Targeted proteomics verified candidate proteins in an independent multicenter cohort. In small held-out test groups, the CSF panel reached AUC 0.984 and the serum panel AUC 0.881; the CSF panel also distinguished Alzheimer-related MCI from frontotemporal dementia, ALS and Huntington disease in the reported analyses.
Technology: Random-forest machine-learning classifiers on CSF and serum proteomic profiles
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: In a proteomic study, researchers measured paired CSF and serum samples in a discovery cohort and used random-forest machine learning to select a 19-protein CSF panel and an eight-protein serum panel to distinguish mild cognitive impairment due to Alzheimer disease from cognitively normal participants.
- Still unknown: The AUC test sets were small (CSF: 9 Alzheimer-related MCI and 21 controls; serum: 8 and 21). The study used separate independent multicenter targeted-proteomics validation for candidate proteins; that does not make the small classifier test cohorts large or establish clinical utility. The serum panel was less effective at distinguishing Alzheimer-related MCI from other neurodegenerative diseases than the CSF panel. These are research panels, not routine tests.
- AI role: Random forests prioritized 19 CSF proteins and eight serum proteins to form disease-specific MCI classifiers after proteomic discovery and targeted validation; the selected panels were evaluated in held-out samples and against other neurodegenerative diseases.
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC10794110/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review scope and date are not recorded; source checked does not establish full-text review.
Citations: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 66 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
Literature-trained models and iterative experimental feedback guided a predetermined ammonia-catalyst search. Selected formulations exceeded many reported comparator activities in the authors' laboratory measurements. Initial literature-only predictions were insufficient, making feedback essential. This is a bounded formulation result, not demonstrated commercial green-ammonia performance or invention of promoted ruthenium catalysis.
Technology: Active-learning regression trained on literature and experimental feedback
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: Researchers synthesized and tested model-selected catalyst formulations and updated the model.
- Still unknown: Independent formulation priority, replication and industrial performance were not assessed.
- AI role: Active-learning regression trained on literature and experimental feedback supported the stated study; experimental and interpretive work remained with researchers.
Evidence sources: https://pubs.rsc.org/en/content/articlelanding/2024/ta/d3ta05939a
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review date: 16 Nov 2026. Review scope: Indexed primary publisher abstract, author affiliations and publication metadata; prior frozen method review retained, author manuscript freshly blocked..
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.
The authors combined an 80-amine high-throughput synthesis dataset, chemistry-informed molecular descriptors, subgroup discovery and a support-vector classifier to rank organic spacers for 2D AgBi iodide perovskites. Of 13 model-selected candidates taken forward, eight were successfully synthesized and characterized by single-crystal X-ray diffraction and powder X-ray diffraction. The study reports a 61.5% success rate in this selected validation set; it does not establish universal synthesis performance across other perovskite families.
Technology: Subgroup discovery and support-vector classification using physicochemical and molecular-topology descriptors
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Peer-reviewed paper
- Test type: direct test
- Tested: The primary article reports successful synthesis of 8 of 13 machine-learning-selected 2D AgBi iodide perovskites (61.5%), with crystal structures determined by single-crystal X-ray diffraction and bulk-phase purity confirmed by PXRD.
- Still unknown: The validation concerns a selected set under common laboratory synthesis conditions and one AgBi iodide family; generalization to other material families is not demonstrated.
- AI role: Subgroup discovery identifies an applicable data region and a support-vector classifier estimates synthesis feasibility from molecular descriptors; the model ranks candidate organic spacers for experimental synthesis.
Evidence sources: https://www.nature.com/articles/s41467-023-44236-5.pdf · https://kjc.seu.edu.cn/2024/0111/c29280a478760/page.htm
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review date: 17 Nov 2026. Review scope: Primary Nature Communications publisher PDF indexed full text including Results, Methods, author contributions and article metadata; Southeast University institutional announcement checked for corroboration..
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.