A robotic electrochemistry stand coupled to Bayesian optimization tested 140 aqueous electrolyte formulas over 40 hours and selected a non-intuitive sodium mixture of NaClO4 and NaNO3. On platinum, the blend had a wider electrochemical stability window than the NaClO4 baseline despite lower total salt concentration; follow-up tests reported a 24 mV increase in the OER potential at the stated leakage-current criterion and 58% lower current density at 2 V. The result is a measured electrolyte formulation under the paper’s electrochemical protocols, not evidence of a commercial battery or field-ready cell.
Technology: Dragonfly Bayesian optimization software coupled to the Otto robotic electrochemistry test stand
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 robotic electrochemistry stand coupled to Bayesian optimization tested 140 aqueous electrolyte formulas over 40 hours and selected a non-intuitive sodium mixture of NaClO4 and NaNO3.
- Still unknown: The study measured electrochemical behavior primarily on platinum electrodes and reports a wider stability window for the selected aqueous blend under its stated protocol. This does not establish long-term performance, commercial readiness, or general improvement across battery chemistries.
- AI role: Dragonfly selected electrolyte formulations sequentially from real-time measurements and guided autonomous experiment requests to Otto; the paper reports the mixed-anion sodium formulation as the optimization result.
Evidence sources: https://doi.org/10.1016/j.xcrp.2020.100264
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 10% 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 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 deep multimodal network combining CT imaging with electronic health record data reached AUROC 0.947 (95% CI 0.946-0.948) on a held-out test set for pulmonary embolism. The abstract reports no sensitivity, specificity, study size or external validation sites, so the evidence is limited to retrospective internal testing.
Technology: Multimodal deep neural network fusing CT and EHR 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: Several fusion architectures were compared on a held-out test set, and the best late-fusion model reached AUROC 0.947 against imaging-only and EMR-only models.
- Still unknown: The abstract reports no sensitivity, specificity, study size or external validation and states no limitations.
- AI role: Deep neural networks classified pulmonary embolism from CT angiography volumes combined with EHR data.
Evidence sources: https://www.ebi.ac.uk/europepmc/webservices/rest/search?query=DOI:10.1038/s41598-020-78888-w&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 10% of works of the same type, year and subfield (OpenAlex work metrics, 263 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 graph neural network trained on a detailed N-body cosmological dark-matter simulation was distilled by symbolic regression into an analytic estimator for a halo’s overdensity from its mass and the masses, positions and velocities of nearby halos. On the analyzed simulation snapshot, the new expression reduced mean absolute error to 0.0882, compared with 0.121 for the authors’ existing hand-designed estimator. The formula’s mass-weighted neighborhood term decreases with separation and includes dependence on the central halo’s mass.
Technology: Sparse-representation Graph Neural Network followed by Eureqa genetic-programming symbolic 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: A graph neural network trained on a detailed N-body cosmological dark-matter simulation was distilled by symbolic regression into an analytic estimator for a halo’s overdensity from its mass and the masses, positions and velocities of nearby halos.
- Still unknown: The formula was derived and evaluated on a single final-time-step dataset from an N-body simulation (215,854 halos), not validated against observations or independent cosmological simulations. The authors note that some terms’ physical interpretation was unclear and reserved detailed interpretation for future astrophysics work. It is a new useful estimator for the studied simulated halo population, not a confirmed universal law of dark-matter structure.
- AI role: The team trained a graph neural network to predict dark-matter halo overdensity using a cosmological N-body simulation. L1 regularization encouraged a compact learned message representation; Eureqa symbolic regression then fitted explicit equations to the learned graph-network components and replaced them with a closed-form model. This process yielded a previously unreported analytic halo-environment estimator and enabled comparison against the existing hand-designed formula.
Evidence sources: https://proceedings.neurips.cc/paper/2020/hash/c9f2f917078bd2db12f23c3b413d9cba-Abstract.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.
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.
AlphaFold achieved a median GDT score of 92.4 across the target set (about 1.6 Å mean error) and 87.0 on the hardest targets, which organizers judged a solution to the problem of single-protein structures. DeepMind notes that protein complexes and interactions with other molecules remain open challenges.
Technology: AlphaFold 2
Family AlphaFold: profile · models page
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: Blind assessment at CASP14 on recently determined protein structures: median GDT of 92.4 across targets and 87.0 on the hardest free-modelling targets, judged by the organizers as a solution to single-protein structure prediction.
- Still unknown: Protein complexes, interactions with DNA, RNA or small molecules, and full side-chain placement remain open; not every predicted structure will be perfect.
- AI role: Attention-based neural network using evolutionary sequence alignments and residue-pair representations to predict 3D protein structure from sequence.
Evidence sources: https://deepmind.google/blog/alphafold-a-solution-to-a-50-year-old-grand-challenge-in-biology/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: blog. 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.
CAMEO, using Bayesian active learning and closed-loop X-ray measurements, identified the Ge4Sb6Te7 alloy in 19 experimental cycles (about 10 hours), selected from 177 compositions. GST467 showed twice the optical contrast of GST225 and was demonstrated in a phase-change memory device; the result concerns this library and application, not a commercial memory.
Technology: CAMEO, closed-loop Bayesian 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: CAMEO ran a closed loop over 177 candidate compositions in 19 experimental cycles; GST467 was measured by X-ray diffraction, showed twice the optical contrast of the reference alloy, and was demonstrated in a phase-change memory device.
- Still unknown: Long-term performance and larger-scale behavior are not discussed; the tenfold time reduction is an estimate, not a measured result.
- AI role: Closed-loop Bayesian active learning chose which compositions to measure next and ran the X-ray measurements.
Evidence sources: https://www.nist.gov/news-events/news/2020/11/nist-ai-system-discovers-new-material · https://www.ebi.ac.uk/europepmc/webservices/rest/search?query=DOI:10.1038/s41467-020-19597-w&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, 637 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 model trained on volunteer recordings recognized 98.5% of coughs associated with confirmed infections in the test reported by MIT. This is an early result on data collected online, with health status also self-reported; it is not a diagnosis or prospective clinical validation.
Technology: Neural networks for audio
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: Model trained on 4,256 subjects and tested on 1,064 held-out subjects from web-collected cough recordings, with results validated against official COVID-19 test outcomes; sensitivity 98.5% and specificity 94.2% overall, but 83.2% specificity in asymptomatic subjects.
- Still unknown: Not a diagnosis for symptomatic people and does not separate Covid from flu or asthma; status was self-reported; false-positive results for healthy participants not reported; app would need approval and a larger, more diverse dataset.
- AI role: Neural networks trained on cough biomarkers (vocal cords, sentiment in speech, lung performance) combined with a muscular-degradation estimate to flag Covid-positive coughs.
Evidence sources: https://news.mit.edu/2020/covid-19-cough-cellphone-detection-1029 · https://www.ebi.ac.uk/europepmc/webservices/rest/search?query=DOI:10.1109/OJEMB.2020.3026928&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, 609 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 QUBRICS survey applied a machine-learning selection pipeline to southern optical and infrared catalogs and followed candidates spectroscopically. The paper reports 224 newly identified bright quasars at z≥2.5, including 15 at z>4, and provides a large southern sample for cosmological observations.
Technology: QUBRICS machine-learning photometric selection with Canonical Correlation Analysis redshift 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 QUBRICS survey applied a machine-learning selection pipeline to southern optical and infrared catalogs and followed candidates spectroscopically.
- Still unknown: The 224 are cumulative QUBRICS discoveries across its pilot and follow-up publications. The paper estimates a 68% success rate for the current selection; contaminants include lower-redshift quasars/AGN. Five objects in the sample had already appeared in literature and were independently selected, so the total is a survey result, not 224 exclusive first publications in this paper.
- AI role: Machine-learning selection ranked photometric sources for follow-up; the team obtained and analyzed optical/infrared spectra to identify and measure redshifts. The paper reports the cumulative survey sample, not 224 objects newly found by a single autonomous run.
Evidence sources: https://doi.org/10.3847/1538-4365/abafc1
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, 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.
FermiNet approximates molecular ground-state electron wavefunctions from first principles. Against the coupled-cluster reference, it recovered 99.8% of the correlation energy beyond a single Slater determinant for the smallest molecules and at least 97% for bicyclobutane (30 electrons). It also modeled stretched nitrogen and hydrogen chains. These are computational results from the original work announced on 19 October 2020. The source page was updated on 22 August 2024 with a separate excited-state study; that later study's carbon-dimer error is not a 2020 result.
Technology: FermiNet
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: Energies compared with coupled-cluster and variational Monte Carlo reference calculations: 99.8% of the correlation energy beyond a single determinant recovered for the smallest molecules and at least 97% for bicyclobutane.
- Still unknown: The coupled-cluster baseline scales poorly, so results apply to medium-sized molecules; the Monte Carlo method is approximate; excited-state results are within about 0.1 eV of the most demanding calculations rather than matching them.
- AI role: Neural network wavefunction, antisymmetric in its inputs, trained by minimizing energy with Monte Carlo sampling using only nuclear positions.
Evidence sources: https://deepmind.google/blog/ferminet-quantum-physics-and-chemistry-from-first-principles/
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, 664 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.
Analysis of high-resolution images located and measured the crowns of over 1.8 billion trees across 1.3 million km² of West Africa. Validation used manual annotations and field data; the study counts trees but does not yet measure their stored carbon.
Technology: 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: Tree detections checked against manually annotated data and field data from the region; the abstract itself does not describe the validation method.
- Still unknown: The study counts trees but does not yet measure stored carbon; the method can mistake stones for trees and gets harder with more detail.
- AI role: Deep learning (fully convolutional network trained on about 90,000 hand-marked trees) on submetre satellite imagery mapped the crowns of over 1.8 billion trees.
Evidence sources: https://www.nasa.gov/centers-and-facilities/goddard/nasa-supercomputing-study-breaks-ground-for-tree-mapping-carbon-research/ · https://www.ebi.ac.uk/europepmc/webservices/rest/search?query=DOI:10.1038/s41586-020-2824-5&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, 496 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.
The system plans multistep syntheses of complex natural products, with routes in a Turing-test study largely indistinguishable from human routes; three computer-designed syntheses were carried out in the lab: dauricine, tacamonidine, and lamellodysidine A. Based on over 100,000 rules plus AI, it is not modern generative AI.
Technology: Chematica / Allchemy
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 Turing-like test, synthesis experts largely could not distinguish the program's routes from human ones; three computer-designed natural-product syntheses were carried out in the lab.
- Still unknown: How often its routes fail across other targets is not reported in the abstract.
- AI role: The planning program combines chemistry knowledge, data-based AI routines and causal relationships to plan several steps ahead.
Evidence sources: https://doi.org/10.1038/s41586-020-2855-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, 295 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.
Akita, a convolutional network, predicts chromatin contact maps for about 1 Mb of DNA sequence. On held-out data it reaches a Pearson R of 0.61 and Spearman R of 0.56, and it learns an orientation-specific grammar for CTCF binding sites. Caveat: the predictions are computational; the paper applies them to in silico mutagenesis, eQTL interpretation and structural-variant prediction.
Technology: Akita (convolutional neural network, Basenji-based trunk)
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: Akita's contact-map predictions were evaluated on held-out data and applied in silico to mutagenesis, eQTL interpretation and structural-variant prediction.
- Still unknown: Not reported in the abstract read; the predictions are computational and no experimental validation is described there.
- AI role: A convolutional network predicting chromatin contact maps from about 1 Mb of DNA sequence, which learned an orientation-specific CTCF binding grammar.
Evidence sources: https://www.ebi.ac.uk/europepmc/webservices/rest/search?query=DOI:10.1038/s41592-020-0958-x&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 10% of works of the same type, year and subfield (OpenAlex work metrics, 309 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.
Learned models selected bacterial stress-combination experiments and incorporated their measurements. The study reported cross-stress protection and vulnerability patterns, with follow-up supporting roles for stress-response machinery. This is a laboratory microbiology result, not a clinical resistance prediction or evidence of treatment efficacy.
Technology: OPEX machine-learning experimental 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: Selected experiments and subsequent biological validation tested stress responses.
- Still unknown: Transfer beyond the studied organism and conditions, clinical outcomes and independent replication were not assessed.
- AI role: OPEX machine-learning experimental selection supported the stated study; experimental and interpretive work remained with researchers.
Evidence sources: https://pubmed.ncbi.nlm.nih.gov/33024104/ · https://escholarship.org/content/qt5cc185t3/qt5cc185t3.pdf?t=qlp51p · https://tagkopouloslab.ucdavis.edu/news/2020/10/new-article-in-nature-communication/
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 PubMed abstract/date/affiliations, indexed deposited publisher-paper abstract and author laboratory announcement..
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 active-learning workflow with a 12-layer convolutional network analyzed about 4.1 million LAMOST DR2 spectra and used human labeling of uncertain cases. After visual inspection and rejection of 58 poor spectra, the paper reports 4,321 spectra from about 3,788 objects, including 1,013 spectra of 948 objects not matched to SIMBAD or the specified Hou et al. 2016 catalog. Visual inspection supported emission signatures, but catalog non-match does not prove universal prior absence and proposed subtypes remain hypotheses. Count the catalog study once.
Technology: 12-layer one-dimensional convolutional neural network with uncertainty-sampling active 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: The paper reports human-oracle labeling and visual inspection of candidates, confirming emission signatures in retained spectra.
- Still unknown: “New” means not matched to the named SIMBAD/Hou catalogs at the study date; proposed object classes require further classification.
- AI role: An active-learning CNN prioritized emission-line spectra in a large archive and asked human annotators to label uncertain examples.
Evidence sources: https://arxiv.org/abs/2009.03219 · https://doi.org/10.1051/0004-6361/201936090
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 metadata/abstract and journal full indexed results, publication dates, affiliations, and stated catalog limits..
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 Gaussian-process classifier, supported by random-forest, extra-trees and multilayer-perceptron models, was applied to Kepler transit candidates. From 83 candidates meeting the study’s stringent ML and data-quality criteria, the authors newly validated 50 as planets after updated vespa false-positive checks also gave each less than a 1% false-positive probability. The result is statistical validation using Kepler pipeline and stellar data, not new telescope observations of the planets.
Technology: Gaussian process classifier, random forest, extra trees and multilayer perceptron classifiers
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 Gaussian-process classifier, supported by random-forest, extra-trees and multilayer-perceptron models, was applied to Kepler transit candidates.
- Still unknown: The 50 are statistically validated candidates from archival Kepler observations, not newly observed planets. The authors caution that their method is trained and demonstrated on Kepler, that atypical/outlier candidates may be unreliable, and that ML and vespa disagree for 33 other high-ML candidates; those were not counted as validated. This entry counts one paper-level validation result, not 50 separate records.
- AI role: Four machine-learning classifiers jointly ranked candidate transit signals against planet and false-positive scenarios. The paper required all four to cross 0.99 planet probability plus positional, stellar, binary and outlier checks; an updated vespa calculation then independently cross-checked the retained candidates.
Evidence sources: https://arxiv.org/abs/2008.10516
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, 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.
Relevance-vector-machine sparse regression applied to high-resolution idealized ocean simulations produced compact momentum and temperature-flux closures for coarse grids. In the authors’ simulated setup, the inferred momentum terms include energy transfers absent from then-current closures; the parameterizations improved selected statistics when implemented in an idealized 30-km ocean model. This is a model-derived closure result, not an observation of a previously unknown ocean process.
Technology: Relevance vector machine (RVM) sparse Bayesian regression; physically constrained convolutional neural network used as a comparison
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: Relevance-vector-machine sparse regression applied to high-resolution idealized ocean simulations produced compact momentum and temperature-flux closures for coarse grids.
- Still unknown: Derived from idealized, eddy-resolving double-gyre simulations and tested in an idealized coarse-resolution shallow-water model, not validated against observational ocean data or a global coupled climate model. The inferred tensor structure relates to earlier Reynolds-stress and deformation-based theories; the paper’s stated novelty is the additional data-derived terms/energy pathways and resulting parameterizations, not discovery of a new observed ocean phenomenon.
- AI role: RVM selected compact closed-form terms from candidate functions of simulated velocity, vorticity and deformation fields to infer subgrid momentum and temperature forcing. The authors report the expressions captured energy pathways omitted from existing parameterizations. A physically constrained CNN was also trained and evaluated, but the interpretable closure result comes from the RVM.
Evidence sources: https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2020GL088376
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, 288 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 learned model ranked candidate protein families for experimental follow-up. Testing supported two previously unknown anti-CRISPR proteins in the studied system. The much larger predicted family catalog remains computational and is not counted as experimentally validated discoveries. Assay activity does not establish native-host function or improved genome-editing performance.
Technology: Extremely randomized forest of genomic-context features
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 predicted candidates received functional laboratory tests.
- Still unknown: Broader family validation, native-host activity and practical utility remain unestablished.
- AI role: Extremely randomized forest of genomic-context features supported the stated study; experimental and interpretive work remained with researchers.
Evidence sources: https://www.nature.com/articles/s41467-020-17652-0.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 published PDF abstract, affiliations and model-training Methods; experimental scope checked against prior primary 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 structured machine-learning analysis of nine years of Whakaari seismic records identified four-hour energy bursts that occurred hours to days before most of the five major eruptive episodes studied. The authors interpret the pattern as possible charging of the vent hydrothermal system by hot magmatic fluids. In cross-validation, the prototype alert model gave advance warning for four of five unseen eruption instances, including at least four hours before the 2019 eruption.
Technology: Structured machine-learning pipeline using engineered seismic time-series features, a random forest, and an alert decision 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 structured machine-learning analysis of nine years of Whakaari seismic records identified four-hour energy bursts that occurred hours to days before most of the five major eruptive episodes studied.
- Still unknown: This is a study of five major eruptive episodes at one volcano. The 4/5 warning result is cross-validation on this historical record, not independent operational validation; the model was described as a prototype. The ensemble alert score is not a calibrated eruption probability, and alert thresholds trade missed events against false alarms. The proposed hydrothermal charging mechanism is an interpretation of the seismic pattern.
- AI role: The machine-learning pipeline analyzed continuous seismic tremor data, selected precursor patterns, and modeled short-term eruption alerts. Its feature analysis identified the four-hour energy-burst pattern; the proposed hydrothermal charging mechanism is the authors’ geophysical interpretation, not a direct measurement of subsurface fluid movement.
Evidence sources: https://www.nature.com/articles/s41467-020-17375-2
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, 144 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.
An attention-based Transformer trained on signal-peptide data from Swiss-Prot generated candidate secretion sequences. Experimental testing in an industrial Bacillus subtilis strain found that model-generated signal peptides directed secreted enzyme activity competitive with industrially used signal peptides. The result establishes functional candidates in the reported host and assay; it does not establish broad performance across proteins or production hosts. The primary institutional record gives the article publication date as 2020-07-10.
Technology: Attention-based neural network (Transformer) trained on Swiss-Prot signal peptide sequences
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 abstract reports experimental testing in an industrial Bacillus subtilis strain and secreted enzyme activity competitive with industrially used signal peptides.
- Still unknown: The bounded abstract-level claim does not quantify all peptide/host combinations or demonstrate generalization across unrelated proteins and production conditions.
- AI role: An attention-based Transformer trained on Swiss-Prot generated signal-peptide sequences, which were then tested experimentally for protein secretion.
Evidence sources: https://authors.library.caltech.edu/records/5dqa2-44d08
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 Caltech institutional record with the article abstract, author contribution statement, publication date, and affiliation; the linked PDF download was unavailable in this session..
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 mobile robot ran 688 experiments in eight days in a ten-variable space, guided by a batch Bayesian algorithm, to search for photocatalysts that produce hydrogen from water; the best mixtures were six times more active than the initial formulations. This optimizes a formulation, not a new structure; only the abstract was reviewed.
Technology: Batch Bayesian optimization and mobile robot
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: 688 autonomous experiments over eight days in a ten-variable space; the best photocatalyst mixtures were about six times more active than the initial formulations.
- Still unknown: The abstract lists no limitations of this study; it notes that atomistic simulations cannot yet capture multi-scale complexity and that robots are hard to adopt widely across diverse samples and instruments.
- AI role: A batch Bayesian algorithm chose which experiments the mobile robot ran next, automating the researcher rather than the instruments.
Evidence sources: https://www.ebi.ac.uk/europepmc/webservices/rest/search?query=DOI:10.1038/s41586-020-2442-2&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, 1704 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 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.
Models trained on DNA-encoded-library selection data prioritized roughly 2,000 compounds for biochemical testing across three targets. The study reports active and potent hits outside the training libraries. Human filtering and chemical review contributed to selection. These are biochemical hits, not established safe or effective drugs; the authors disclose commercial interests and related patent applications.
Technology: Graph-convolutional neural networks and random forests
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 in biochemical assays.
- Still unknown: Cellular efficacy, selectivity, safety and clinical benefit remain unestablished.
- AI role: Graph-convolutional neural networks and random forests supported the stated study; experimental and interpretive work remained with researchers.
Evidence sources: https://accio.github.io/AMIDD/assets/2020/13-14/McCloskey-2020-ML-DELT.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: Author-hosted published-article PDF abstract, selected experimental Results, affiliations and disclosures; prior PubMed online date retained..
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 CNN trained on contrast-enhanced CT from a Taiwanese centre reached 98.3% sensitivity on combined local test sets versus 92.9% for radiologists (difference 5.4 points, p=0.014). On a US external test set, sensitivity fell to 79.0% with specificity 97.6% (AUC 0.920), and sensitivity for tumours under 2 cm was 63.1% in that set. The study is retrospective and the authors describe it as supplementary to radiologists.
Technology: Convolutional neural network applied to CT image patches
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 CNN trained on Taiwanese contrast CT was tested on local test sets against radiologists and on an external US test set, where it reached 79.0% sensitivity and 97.6% specificity.
- Still unknown: The study is retrospective; sensitivity for tumours under 2 cm was 63.1% on the US set, and radiologist comparisons covered only the local sets.
- AI role: A CNN classified image patches as cancerous and labelled each patient by the proportion of cancerous patches; the authors describe it as supplementary to radiologists.
Evidence sources: https://www.ebi.ac.uk/europepmc/webservices/rest/search?query=EXT_ID:33328124%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 10% of works of the same type, year and subfield (OpenAlex work metrics, 233 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.
AcRanker, an XGBoost ranking model using amino-acid sequence features, ranked proteins from prophage regions in self-targeting Streptococcus genomes. Experimental testing identified two previously unknown type II-A anti-CRISPRs: AcrIIA20 (ML1), which strongly inhibited Streptococcus iniae Cas9 and weakly inhibited S. pyogenes Cas9, and AcrIIA21 (ML8), which inhibited S. pyogenes, S. aureus and S. iniae Cas9 with lower potency.
Technology: AcRanker, an XGBoost pairwise ranking model trained on amino-acid composition and grouped di- and tri-mer features
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 testing identified two previously unknown type II-A anti-CRISPRs: AcrIIA20 (ML1), which strongly inhibited Streptococcus iniae Cas9 and weakly inhibited S. pyogenes Cas9, and AcrIIA21 (ML8), which inhibited S. pyogenes, S. aureus and S. iniae Cas9 with lower potency.
- Still unknown: The authors selected 10 candidates for follow-up, successfully purified nine and reported three validated candidates across their experiments, including ML3/AcrIIA12, which had also been identified separately as an LmoCas9 inhibitor during the study. AcrIIA20 and AcrIIA21 were roughly tenfold less potent against SpyCas9 than the established AcrIIA4 control; tested potency and host range are assay-specific. The 2025 Evo paper reports distinct generated proteins EvoAcr1–5 and separate assays; shared anti-CRISPR function does not make these identical discoveries.
- AI role: The model ranked candidate proteins from predicted prophage regions for anti-CRISPR likelihood; the researchers selected candidates from those rankings and biochemically tested them. It directly prioritized the proteins later named AcrIIA20 and AcrIIA21, whose Cas9 inhibition was measured in vitro.
Evidence sources: https://academic.oup.com/nar/article/48/9/4698/5819938
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, 125 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.
In a test on 279 patients, a system combining CT scans and clinical information reached 84% sensitivity, compared with 75% for radiologists in the reported comparison. Retrospective study on data collected in China, without evidence of effectiveness in prospective clinical use.
Technology: Multimodal deep 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: Trained on 626 of 905 patients and tested on the remaining 279 against radiologists: sensitivity 84% versus 75%, AUC 0.92, and 68% of RT-PCR-positive cases with normal CT correctly identified.
- Still unknown: Intended for use when CT scans and clinical history are both available; CT alone may have limited negative predictive value; prospective clinical use not evaluated; specificity not reported.
- AI role: Combined chest CT findings with clinical symptoms, exposure history and lab results to identify COVID-19-positive patients.
Evidence sources: https://www.mountsinai.org/about/newsroom/2020/mount-sinai-first-in-us-to-use-artificial-intelligence-to-analyze-coronavirus-covid19-patients-pr · https://www.ebi.ac.uk/europepmc/webservices/rest/search?query=DOI:10.1038/s41591-020-0931-3&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, 1114 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 system analyzes OCT scans to predict whether the less affected eye of patients with macular degeneration will develop the wet form, outperforming specialists in the comparison described. The announcement gives no numerical metrics; preventive clinical studies were still to be planned.
Technology: Deep learning on OCT images
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 prediction of whether the second eye converts to wet AMD within 6 months: 80% sensitivity at 55% specificity, and better than five of six experts in the comparison described.
- Still unknown: Preventive use needs prospective trials that were only being considered; 56% of individual eyes were false positives at the high-sensitivity setting; dataset size and generalizability not stated.
- AI role: Models built on 3D OCT scans and automatically generated tissue maps to predict conversion to wet AMD.
Evidence sources: https://www.moorfields.nhs.uk/about-us/news-and-blogs/news/ai-to-develop-an-early-warning-system-for-amd · https://www.ebi.ac.uk/europepmc/webservices/rest/search?query=DOI:10.1038/s41591-020-0867-7&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, 307 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.
DFT and active machine learning identified Cu–Al catalysts that reduce CO2 to ethylene with over 80% Faradaic efficiency at 400 mA/cm² (about 66% for pure copper) and 55% half-cell energy efficiency at 150 mA/cm². Synthesized and tested in the lab; the abstract does not state how many candidates were screened.
Technology: Active machine learning and DFT
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: Cu-Al catalysts synthesized and tested in the lab: over 80% Faradaic efficiency for ethylene at 400 mA/cm2 versus about 66% for copper, and 55% half-cell energy efficiency at 150 mA/cm2; in situ X-ray absorption examined the copper environment.
- Still unknown: The abstract does not state how many candidates were screened or list explicit limitations; full text not read.
- AI role: DFT calculations combined with active machine learning identified the Cu-Al catalysts.
Evidence sources: https://www.ebi.ac.uk/europepmc/webservices/rest/search?query=DOI:10.1038/s41586-020-2242-8&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, 1624 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 robotic laboratory used model-based optimization to select thin-film compositions and processing conditions, execute experiments and learn from their measured optical and electronic properties. The reported demonstration optimized hole mobility in organic hole-transport materials used in perovskite solar cells and electronics. This description is limited to the paper abstract; it does not assert a numerical mobility gain.
Technology: Closed-loop self-driving laboratory with autonomous 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: Closed-loop composition and processing experiments on organic hole-transport thin films, with measured hole mobility as an optimization target.
- Still unknown: Abstract and publication metadata reviewed; detailed optimization history, baseline comparisons and durability were not assessed.
- AI role: A model-based optimizer selected subsequent experiments using measurements from the robotic platform.
Evidence sources: https://www.ebi.ac.uk/europepmc/webservices/rest/search?query=DOI%3A%2210.1126%2Fsciadv.aaz8867%22&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, 929 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.
Machine learning guided selection of HoB2, followed by synthesis and magnetic measurements. The paper reports a magnetic entropy change of 40.1 J kg−1 K−1 for a 5 T field change near a 15 K ferromagnetic transition. The compound was known; the reported discovery is its measured magnetocaloric response, not a demonstrated refrigerator.
Technology: Machine-learning-guided composition screening; XGBoost
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: Machine learning ranks candidate compounds for measurement; HoB2 was then synthesized and its magnetocaloric properties measured experimentally.
Evidence sources: https://www.nature.com/articles/s41427-020-0214-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 date: 16 Nov 2026. Review scope: Primary-source abstract and publication metadata reviewed; full text 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.
A COVID Symptom Study app with 2,618,862 participants and 18,401 tested users found loss of smell or taste in 65.03% of positives versus 21.71% of negatives (OR 6.74). A machine-learning model applied to 805,753 users predicted 140,312 (17.42%) as likely infected. The inputs are self-reported and the test-verified subset is small relative to the user base, so the prediction is not a clinical diagnosis.
Technology: Machine-learning symptom model on self-reported app 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: App-reported symptoms from 2,618,862 participants were compared between 18,401 test-verified users, and a model applied to 805,753 users predicted 140,312 as likely infected.
- Still unknown: Inputs are self-reported, the test-verified subset is small relative to the user base, the prediction is not a clinical diagnosis, and the abstract does not describe model validation.
- AI role: A model combining self-reported symptoms predicted likely COVID-19; the abstract does not state the method.
Evidence sources: https://www.ebi.ac.uk/europepmc/webservices/rest/search?query=DOI:10.1038/s41591-020-0916-2&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, 1548 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.
A random-forest model ranked new chromium-containing multicomponent carbide compositions by predicted entropy-forming ability. After density-functional calculations and experimental synthesis, X-ray diffraction identified four previously unreported single-phase rock-salt compositions: CrMoNbVWC₅, CrMoNbTaWC₅, CrMoTaVWC₅, and CrMoTiVWC₅. The first three were high-EFA selections; CrMoTiVWC₅ was a medium-EFA composition and required extended annealing to reach chemical homogeneity.
Technology: Random-forest models using chemical attributes, with and without CALPHAD thermodynamic features
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 random-forest model ranked new chromium-containing multicomponent carbide compositions by predicted entropy-forming ability.
- Still unknown: Only seven of the 70 model-screened candidates were selected for DFT and synthesis, and several tested compositions formed multiple phases. Single-phase identification used XRD and was supported by microstructure, elemental mapping and EBSD; the medium-EFA CrMoTiVWC₅ sample required extended annealing. The paper establishes synthesis and phase formation, not measured application performance or engineering readiness.
- AI role: The random-forest models learned entropy-forming ability from 56 previously reported carbide compositions and rapidly screened 70 new chromium-containing candidates. Model rankings guided selection of seven compositions for DFT calculations and synthesis; experimental phase outcomes agreed with predicted rankings, including four single-phase materials.
Evidence sources: https://www.nature.com/articles/s41524-020-0317-6
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, 288 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 physics-informed Gaussian process using historical data tunes the accelerator’s 24 LCLS quadrupole magnets. The authors estimate tuning is about three to five times faster, with beams tending to be more intense; this is a result on the real machine, limited to this part of the tuning process.
Technology: Gaussian-process 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: Applied to the LCLS X-ray free-electron laser; the abstract reports the Bayesian optimizer as sample-efficient and outperforming existing optimizers, and the press release estimates tuning three to five times faster.
- Still unknown: Speed and beam-quality gains are estimates; the method covers only part of the tuning process; results from three other accelerators are not reported.
- AI role: Gaussian process model predicted the effect of magnet adjustments, chose which to try, and was trained on past machine data.
Evidence sources: https://www6.slac.stanford.edu/news/2020-04-29-new-machine-learning-method-streamlines-particle-accelerator-operations · https://www.ebi.ac.uk/europepmc/webservices/rest/search?query=DOI:10.1103/PhysRevLett.124.124801&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 10% of works of the same type, year and subfield (OpenAlex work metrics, 166 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 CT model trained on 1,186 patients (521 COVID-19, 665 other pneumonia) reached 95% sensitivity and 96% specificity internally, and AUC 0.90 on an independent test set. Radiologist accuracy rose from 85% to 90% with AI assistance. The study is retrospective and the independent-test performance is lower than the internal result.
Technology: Deep-learning CT classifier used as an assistant for radiologists
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 CT model was tested on an internal split and on an independent test set from separate hospitals, and six radiologists read cases unassisted and then with the model's probabilities.
- Still unknown: The abstract states no limitations section; performance on independent data was lower (87% vs 96% accuracy) and the sensitivity confidence interval is wide.
- AI role: The AI classified COVID-19 versus other pneumonia on CT slices and its probabilities were shown to radiologists as decision support.
Evidence sources: https://www.ebi.ac.uk/europepmc/webservices/rest/search?query=DOI:10.1148/radiol.2020201491&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, 385 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.
An automated laboratory used measured feedback and fitted local optimization models to search for optically active CsPbBr3 nanocrystals. Spectroscopy measured circular dichroism and imaging informed a proposed defect-related explanation. The composition was known; the contribution is an experimentally characterized optical property and its optimization. The proposed mechanism and future device applications are not independently established.
Technology: SNOBFIT adaptive optimization with locally fitted quadratic 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: Nanocrystal optical properties and structures were measured during adaptive experiments.
- Still unknown: Independent replication, isolated causal mechanisms and device-scale utility were not assessed.
- AI role: SNOBFIT adaptive optimization with locally fitted quadratic models supported the stated study; experimental and interpretive work remained with researchers.
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC7184584/
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 and optimization Methods, affiliations 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.
PENet, a 3D deep-learning model for pulmonary embolism on CT angiography, reached AUROC 0.84 on an internal hold-out test and 0.85 on data from one unnamed external institution. The abstract does not report sensitivity, specificity or study size, and the external validation is limited to a single site.
Technology: PENet, 3D deep convolutional network on volumetric CT
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: PENet, a 3D CNN, was tested on a hold-out set from the training institution (AUROC 0.84) and on an external institution (AUROC 0.85), and was compared with other state-of-the-art 3D CNN models.
- Still unknown: The abstract states no limitations; external validation covers one institution and sensitivity, specificity and study size are not reported.
- AI role: A 77-layer 3D CNN, pretrained on Kinetics-600, detected pulmonary embolism on CT angiography and is proposed as a triage tool.
Evidence sources: https://www.ebi.ac.uk/europepmc/webservices/rest/search?query=DOI:10.1038/s41746-020-0266-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: 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, 203 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 networks predict particle mobility in simulated glasses; perturbing the network reveals growing spatial correlations as the glass transition is approached. Evidence comes from simulations: the learned correlation length does not translate directly into a measurable physical quantity.
Technology: Graph neural networks
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: Networks trained on simulated glasses predicted particle mobility and were checked against simulation ground truth, outperforming physics-inspired and state-of-the-art baselines (96% correlation at short timescales, 64% at relaxation time).
- Still unknown: The network's correlation length cannot be directly converted into a measurable physical quantity; it remains an open debate; the work models simulated glasses, not physical samples.
- AI role: Graph neural networks predicted particle mobility and probed spatial correlations by perturbing particles in shells around a marked particle.
Evidence sources: https://deepmind.google/blog/towards-understanding-glasses-with-graph-neural-networks/
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, 358 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 high-throughput robotic synthesis and characterization workflow used a convolutional neural network to recognize crystal growth and a machine-learning regressor to guide synthesis conditions. The team reported the first synthesis of (3-PLA)₂PbCl₄, a new perovskite-like single crystal with a direct band gap near 4 eV and strong blue emission. The reported growth conditions were restrictive and required a carefully tuned search.
Technology: Convolutional neural network image classifier and machine-learning synthesis-condition regressor
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 high-throughput robotic synthesis and characterization workflow used a convolutional neural network to recognize crystal growth and a machine-learning regressor to guide synthesis conditions.
- Still unknown: The authors characterize this as a new perovskite-like material and report a laboratory single-crystal synthesis and blue emission. Synthesis conditions were restrictive; the paper does not establish device performance or scale-up.
- AI role: The CNN classified crystal-growth outcomes from images, and the model trained on high-throughput experimental results selected promising conditions for successive synthesis trials. That feedback-guided sequence enabled the first synthesis of (3-PLA)₂PbCl₄.
Evidence sources: https://light.utoronto.ca/wp-content/uploads/2020/09/matter_perovskite_crystallization.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, 152 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 model forecasts precipitation at 1 km resolution and 2-minute intervals up to 8 hours ahead from radar and satellite, without physical equations. According to the page, it outperforms NOAA’s operational HRRR model up to about 7–8 hours, in seconds rather than about an hour. US coverage only, author-only comparisons, a preprint, and no stated operational use.
Technology: MetNet (neural network)
Family MetNet: profile · models page
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: Evaluated on a precipitation-rate forecasting benchmark against NWP (HRRR) and an optical-flow baseline, scored by F1 at 1.0 mm/h; outperforms NWP at forecasts up to about 7 to 8 hours.
- Still unknown: Demonstrated only for the continental US; the authors call it a small stepping stone; comparisons are author-run.
- AI role: Deep neural network with axial self-attention forecasting precipitation probabilities from radar and satellite inputs, without physical equations.
Evidence sources: https://research.google/blog/a-neural-weather-model-for-eight-hour-precipitation-forecasting/ · https://arxiv.org/abs/2003.12140
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 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 model trained on echocardiogram videos segmented the left ventricle with a Dice coefficient of 0.92 and estimated ejection fraction with a mean absolute error of 4.1% (6.0% on an external dataset from another health system). Prospective evaluation with repeated human measurements found its variance comparable to or lower than that of human experts. Caveat: training data come from one institution's echocardiograms, with external testing on a single additional system.
Technology: EchoNet-Dynamic (video 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: Segmentation and ejection fraction were tested on held-out internal data (Dice 0.92, MAE 4.1%) and an external dataset from another health system (MAE 6.0%), and repeated human measurements were compared with model variance.
- Still unknown: Not reported; the page states no limitations section.
- AI role: Deep learning model trained on echocardiogram videos to segment the left ventricle and estimate ejection fraction.
Evidence sources: https://www.ebi.ac.uk/europepmc/webservices/rest/search?query=DOI:10.1038/s41586-020-2145-8&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, 1292 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 COVNet 3D deep-learning model, trained on 4,352 CT scans from 3,322 patients, reached per-scan sensitivity 90% and specificity 96% (AUC 0.96) for COVID-19, and AUC 0.95 for community-acquired pneumonia. The study is retrospective and evaluated on Chinese cohorts, so performance in other populations is not established.
Technology: COVNet, a 3D convolutional neural network on chest CT
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 COVNet 3D model, trained on 4,352 CT scans, was evaluated on an independent test set for COVID-19 (sensitivity 90%, specificity 96%, AUC 0.96) and community-acquired pneumonia.
- Still unknown: The abstract does not state limitations and does not confirm whether the test set came from outside the training hospitals; the study is retrospective.
- AI role: COVNet, a fully automatic 3D deep-learning framework, classified chest CT scans.
Evidence sources: https://www.ebi.ac.uk/europepmc/webservices/rest/search?query=DOI:10.1148/radiol.2020200905&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, 1999 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.
Models trained on functional-gene profiles predicted microbial distribution along salinity, depth and sampling-fraction gradients in Baltic Sea observations. The study links genomic repertoire to ecological niche placement within this ecosystem. It does not demonstrate that associated genes causally determine habitat or that the models transfer elsewhere. The genome clusters are not separately counted discoveries.
Technology: Genome-feature regression and tree-based 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: Predictions were compared with observed distributions in the sampled ecosystem.
- Still unknown: Causal effects, field manipulation and cross-ecosystem transfer were not established.
- AI role: Genome-feature regression and tree-based learning supported the stated study; experimental and interpretive work remained with researchers.
Evidence sources: https://www.nature.com/articles/s42003-020-0856-x · https://epic.awi.de/52845/1/Alneberg_et_al_2020.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: Indexed primary publisher abstract, date and affiliations, plus institutional publisher-PDF metadata; Methods review retained from frozen research..
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 model recognizes signs on tablets from the Persepolis Fortification Archive with about 80% accuracy on tablets excluded from training (over 6,000 annotated images and over 100,000 signs in the database). About 20% of signs are wrong; testing is internal only, and the source is a university announcement.
Technology: DeepScribe (machine learning on images)
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: Recognition of signs on tablets excluded from training reached about 80% accuracy; the source describes the testing as internal only.
- Still unknown: About 20% of signs are still wrong; the size of the held-out set and an error breakdown are not given; other uses are described as aspirations.
- AI role: Machine-learning model trained on over 100,000 annotated signs from Persepolis Fortification Archive tablet images to recognize cuneiform signs.
Evidence sources: https://news.uchicago.edu/story/how-ai-could-help-translate-written-language-ancient-civilizations
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 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.
Knowledge- and Context-driven Machine Learning (KCML) analyzed perturbation phenotypes across molecular, cellular and population scales. In colorectal-cancer cells, it associated depletion of olfactory receptors and TGFβ/WNT pathway genes with abnormal multicellular organization. The authors integrated the predictions with patient data and found olfactory-receptor expression correlated with WNT/TGFβ signatures and worse outcomes in higher-grade tumors.
Technology: KCML, weakly supervised machine-learning framework using support-vector-machine classifiers trained on gene-ontology annotations and high-throughput perturbation phenotypes
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: Knowledge- and Context-driven Machine Learning (KCML) analyzed perturbation phenotypes across molecular, cellular and population scales.
- Still unknown: Patient data validate expression associations and outcome stratification, not a causal olfactory-receptor mechanism in tumors. The authors describe KCML as hypothesis-generating; predictions can be indirect through interacting genes. This is a distinct paper/result from later AI cancer models.
- AI role: Mapped gene-perturbation morphology to context-specific gene functions and nominated olfactory-receptor links to multicellular organization; patient-expression analyses provided an independent association check.
Evidence sources: https://ora.ox.ac.uk/objects/uuid%3Abbef6c79-e4e7-43a3-a819-c00f09a588d1/files/sv979v315p
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 62 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 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.
DeepMind openly shared AlphaFold predictions, without waiting for review, for six poorly studied viral proteins (membrane, ORF3a, Nsp2, Nsp4, Nsp6, and the C-terminal domain of papain-like protease). They were not experimentally verified; an ORF3a experimental structure deposited in the PDB in June 2020 matched the second predicted model well.
Technology: AlphaFold
Family AlphaFold: profile · models page
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: Predictions compared with experimental structures: the spike prediction matched the PDB structure, and the second-ranked ORF3a model agreed well with later experimental work.
- Still unknown: Predictions were not experimentally verified overall; the system was still in development and accuracy is not certain; released before peer review; the top ORF3a model misplaced some transmembrane helices.
- AI role: AlphaFold free-modelling predicted structures for six understudied SARS-CoV-2 proteins, with per-residue confidence scores.
Evidence sources: https://deepmind.google/blog/computational-predictions-of-protein-structures-associated-with-covid-19/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: blog. 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 4/5 · 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 international crowdsourced challenge (over 1,100 participants, 126 teams, 44 countries) tested AI on screening mammograms from 144,231 US and 166,578 Swedish examinations. The best single algorithm did not beat radiologists (US AUC 0.858), but an ensemble of top algorithms combined with US radiologists' recall decisions reached AUC 0.942 and 92.0% specificity at the same sensitivity. The study is retrospective, uses challenge data and does not report recall rates.
Technology: Ensemble of crowdsourced deep-learning mammography algorithms combined with radiologist recall
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 crowdsourced challenge of 126 teams was tested on 144,231 US and 166,578 Swedish screening examinations, with an ensemble of top algorithms combined with US radiologists' recall decisions compared on AUC and specificity.
- Still unknown: The study is retrospective, no single algorithm beat radiologists, the ensemble was evaluated only against US radiologists, and recall rates are not reported.
- AI role: Algorithms produced cancer risk scores; an ensemble of top algorithms was combined with radiologist recall decisions.
Evidence sources: https://www.ebi.ac.uk/europepmc/webservices/rest/search?query=DOI:10.1001/jamanetworkopen.2020.0265&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, 438 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.
A model trained on about 2,500 molecules screened the roughly 6,000 compounds in the Broad Institute’s Drug Repurposing Hub and identified halicin, which killed many resistant bacteria in the lab and cleared A. baumannii infections in mice within 24 hours. Cell-culture and mouse results, not human studies; the mechanism remains preliminary.
Technology: Graph neural network for predicting antibacterial activity
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: Bactericidal activity against a wide range of pathogens, including M. tuberculosis and carbapenem-resistant Enterobacteriaceae; halicin treated C. difficile and pan-resistant A. baumannii infections in mice. Of 23 tested predictions from over 107 million ZINC15 molecules, eight were antibacterial.
- Still unknown: Human safety and efficacy are not reported in the abstract.
- AI role: A deep neural network predicted antibacterial activity across chemical libraries; candidates were then tested experimentally.
Evidence sources: https://doi.org/10.1016/j.cell.2020.01.021
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, 2730 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 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
- same-team extension · 25 May 2023: MIT describes abaucin as a subsequent application of the antibiotic-screening program to A. baumannii. This is a different candidate and target-specific study, not replication of halicin. Source. Recorded 16 Nov 2026.
- same-team extension · 14 Aug 2025: MIT identifies halicin and abaucin as earlier screening results motivating generative design of new NG1/DN1 candidates. The later study does not establish clinical efficacy of these earlier candidates. 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.
A model estimating cycle life from the first 100 cycles, paired with Bayesian optimization, assessed 224 fast-charging protocols in 16 days; the authors estimate exhaustive search without early prediction would take over 500 days. It optimizes protocols, not new materials; only the abstract was reviewed.
Technology: Bayesian optimization with early 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: 224 fast-charging protocols tested experimentally by measuring battery cycle life; high-cycle-life protocols identified in 16 days, versus an estimated 500+ days for exhaustive search without early prediction.
- Still unknown: The abstract lists no explicit limitations; generalization to other problems is suggested but not discussed; only the abstract was read.
- AI role: An early-prediction model estimated final cycle life from the first cycles, and Bayesian optimization chose which protocols to test next.
Evidence sources: https://www.ebi.ac.uk/europepmc/webservices/rest/search?query=DOI:10.1038/s41586-020-1994-5&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, 1252 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.
Deep-learning QSAR models for parasite activity and cytotoxicity prioritized commercial candidates for assay. Primary Table 3 reports LabMol-152 EC50 values of 0.049 and 0.078 µM against 3D7 and W2; LabMol-149 values were 1.450 and 0.509 µM, inconsistent with the abstract’s under-500-nM claim. The study reports in-vitro results only; the model ranked existing compounds rather than generating them.
Technology: Deep-neural-network QSAR 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: Table 3 reports strain-specific EC50 and COS-7 cytotoxicity results for tested compounds.
- Still unknown: Abstract/results inconsistency for LabMol-149 preserved; no animal or clinical evidence; earlier public version/chemical prior art unresolved.
- AI role: Activity and cytotoxicity models ranked an existing library and filters/human inspection selected compounds for assay.
Evidence sources: https://doi.org/10.1371/journal.pcbi.1007025
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.
A machine-learning model trained on literature conductivity data screened inorganic crystal structures and narrowed the chemical space explored by high-throughput synthesis. The study reports high conductivity and thermal durability in selected Ca-(Nb,Ta)-Bi-O compositions; oxygen concentration-cell measurements gave oxide-ion transport numbers around 70–80% at 973 K. The result came from the combined informatics and experimental workflow. It does not establish a pure ionic conductor or room-temperature electrolyte.
Technology: Machine-learning materials informatics; partial least-squares 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: High-throughput film libraries and bulk samples were synthesized and characterized; oxygen concentration cells measured oxide-ion transport numbers around 70–80% at 973 K.
- Still unknown: The high-temperature transport result does not establish a pure ionic conductor or room-temperature electrolyte; supplementary details were not checked in this source review.
- AI role: A machine-learning model trained on literature oxide-ion conductivities screened crystal structures and helped narrow element combinations for high-throughput experimental synthesis and evaluation.
Evidence sources: https://www.nature.com/articles/s43246-019-0004-7
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 primary full text, methods/results and publication 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.
Sumitomo Dainippon Pharma and Exscientia announced the start in Japan of a Phase I study of DSP-1181, a 5-HT1A serotonin receptor agonist created with the Centaur Chemist platform, with obsessive-compulsive disorder as the first indication. The companies said exploratory research took under 12 months versus a typical 4.5 years. The announcement gives no efficacy results. In the January 31, 2022 investor-call transcript, Sumitomo said development was discontinued because the candidate did not meet the expected Phase I criterion.
Technology: Centaur Chemist (molecule-design platform)
Publication: Announcement only · 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: Announcement only
- Test type: not reported
- Tested: Not reported
- Still unknown: Not reported
- AI role: Candidate created in joint research with Exscientia using its Centaur Chemist AI platform; the page does not describe what the AI did beyond design.
Evidence sources: https://www.sumitomo-pharma.com/news/20200130.html
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: 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 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 double-blind randomised trial of 962 patients in China (484 analysed with CADe, 478 with a sham system), the deep-learning system raised adenoma detection to 34% (165/484) from 28% (132/478), OR 1.36 (95% CI 1.03-1.79). The primary analysis was per protocol, the trial was single-centre, and the abstract reports no ITT analysis.
Technology: Deep-learning computer-aided detection (CADe) system for polyps during colonoscopy
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: A double-blind randomised trial of 962 patients compared colonoscopy with a deep-learning CADe system against a sham system; the primary outcome was adenoma detection rate, analysed per protocol (CADe 34% vs sham 28%).
- Still unknown: The trial is single-centre, the primary analysis is per protocol with no ITT analysis reported, and the effect on interval colorectal cancer still needs investigation.
- AI role: A deep-learning CADe system gave real-time visual alarms for possible polyps during colonoscopy.
Evidence sources: https://www.ebi.ac.uk/europepmc/webservices/rest/search?query=EXT_ID:31981517%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, 437 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 4/5 · Validation 4/5 · 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 model integrated transcriptomic, epigenomic, metabolomic and proteomic data from Huntington-disease cell models to identify previously unknown mechanisms for protective compounds. The analysis predicted autophagy activation for one compound group and mitochondrial-respiration inhibition for another; follow-up imaging, biochemical and energetics assays experimentally confirmed these effects in the cell model.
Technology: Interpretable network-optimization model integrating multi-omics over a molecular-interaction 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 analysis predicted autophagy activation for one compound group and mitochondrial-respiration inhibition for another; follow-up imaging, biochemical and energetics assays experimentally confirmed these effects in the cell model.
- Still unknown: Findings were experimentally established in STHdhQ111 murine striatal cells, with autophagy effects also checked in human SH-SY5Y and HEK293 cell lines. The paper notes that the model prioritizes pathways but does not resolve every causal step or demonstrate therapeutic efficacy in patients.
- AI role: The model integrated multiple omics and prioritized cellular pathways as candidate mechanisms of action; researchers tested those predictions with cellular imaging, biochemical assays and energetics measurements. It revealed unexpected mechanisms for compounds with unrelated known targets.
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC6976599/
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, 76 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 deep-learning network deepDTnet predicted retinoic-acid-receptor-related orphan receptor gamma-t (RORγt) as a previously unrecognized target of topotecan. Biochemical assays showed direct inhibition (IC50 0.43 μM), and topotecan improved outcomes in a mouse experimental autoimmune encephalomyelitis model of multiple sclerosis.
Technology: deepDTnet, heterogeneous drug–gene–disease network embedding with deep 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: Biochemical assays showed direct inhibition (IC50 0.43 μM), and topotecan improved outcomes in a mouse experimental autoimmune encephalomyelitis model of multiple sclerosis.
- Still unknown: Evidence is preclinical: direct target and mouse-model assays do not establish safety or effectiveness in people with multiple sclerosis. Topotecan is an established anticancer drug, and its new target/indication remains a repurposing hypothesis.
- AI role: The model integrated 15 types of chemical, genomic, phenotypic and cellular network profiles to rank new drug targets. It nominated RORγt for topotecan, which the team tested using direct binding/transcriptional assays and a mouse EAE disease model.
Evidence sources: https://pubmed.ncbi.nlm.nih.gov/34123272/
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, 366 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.
CNNs trained on over 2.5 million stimulated Raman histology images produced intraoperative brain tumour diagnoses in under 150 seconds, with 94.6% accuracy versus 93.9% for pathologists. The accuracy comparison comes from a prospective multicentre trial with 278 patients, but the abstract does not report sensitivity or specificity separately.
Technology: Deep neural networks on stimulated Raman histology images
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: Multicenter prospective clinical trial (n = 278): CNN diagnosis of stimulated Raman images was non-inferior to pathologists reading conventional histology (94.6% versus 93.9%), in under 150 seconds.
- Still unknown: Long-term patient outcomes are not reported in the abstract.
- AI role: Convolutional neural networks trained on over 2.5 million stimulated Raman histology images.
Evidence sources: https://doi.org/10.1038/s41591-019-0715-9
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: in the top 1% of works of the same type, year and subfield (OpenAlex work metrics, 839 citations; as of 15 Nov 2026). Attention, not a measure of correctness; not part of the score.
Editorial assessment: Impact 4/5 · 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
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.
Among 25,856 women in the UK and 3,097 in the US, the system reduced false positives by 5.7% (US) and 1.2% (UK), and false negatives by 9.4% and 2.7% in absolute terms compared with radiologists. Trained only on UK data, it also outperformed specialists on US data. Retrospective evaluation (Nature): prospective studies are needed, and clinical deployment is unclear.
Technology: Neural network for mammograms
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 evaluation on UK (25,856 women) and US (3,097 women) screening mammograms, compared with individual radiologists' decisions; a reader study and a simulated double-reading workflow were also run.
- Still unknown: The authors state prospective studies are required and that how to deploy the system clinically is not yet determined; the US data were enriched; the second-reader scenario was only simulated.
- AI role: Predicted from screening mammograms alone whether a biopsy would later show cancer.
Evidence sources: https://deepmind.google/blog/international-evaluation-of-an-ai-system-for-breast-cancer-screening/ · https://www.ebi.ac.uk/europepmc/webservices/rest/search?query=DOI:10.1038/s41586-019-1799-6&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: 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, 3668 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.