Source check: Source checked. Not independent scientific confirmation.
Finding / tested result: Retrospective clinical cohort of 359 AECOPD patients in which machine learning (K-prototype, logistic regression, random forest, XGBoost) identifies sex-associated clinical features, with a nomogram AUC of 0.945; a clinically specific distinct finding with an enabling ML role.
Technology: XGBoost; random forest; K-prototype
Publication: Publication status unspecified in annotation · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- Retrospective clinical cohort of 359 AECOPD patients in which machine learning (K-prototype, logistic regression, random forest, XGBoost) identifies sex-associated clinical features, with a nomogram AUC of 0.945; a clinically specific distinct finding with an enabling ML role.
- Limitation / still unknown
- Affiliation not stated in the reviewed material. Affiliation was not available in the reviewed primary bibliographic metadata. Publication status could not be established from the reviewed primary bibliographic metadata.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://www.frontiersin.org/journals/medicine/articles/10.3389/fmed.2023.1105854/full
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: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 22 Nov 2026. Review scope: Primary result-bearing material for the stated method, result and evaluation, with bibliographic version information..
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
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.
Source check: Source checked. Not independent scientific confirmation.
Finding / tested result: Deep learning screened whole-slide liquid-based urine cytology specimens for urothelial carcinoma using supervised and weakly supervised approaches.
Technology: Deep Learning
Publication: Publication status unspecified in annotation · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- Methods describe transfer learning and whole-slide screening; Results report neoplastic-cell detection AUC in urine cytology slides.
- Limitation / still unknown
- Affiliation not stated in the reviewed material. Affiliation was not available in the reviewed primary bibliographic metadata. Publication status could not be established from the reviewed primary bibliographic metadata.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC9818219/
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: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 22 Nov 2026. Review scope: Primary result-bearing material for the stated method, result and evaluation, with bibliographic version information..
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
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.
Source check: Source checked. Not independent scientific confirmation.
Finding / tested result: XGBoost corrects wearable exoskeleton compliance error for posture; leave-participant-out comparison against motion capture supports bounded posture-estimation result.
Technology: XGBoost
Publication: Publication status unspecified in annotation · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- XGBoost corrects wearable exoskeleton compliance error for posture; leave-participant-out comparison against motion capture supports bounded posture-estimation result.
- Limitation / still unknown
- Affiliation not stated in the reviewed material. Affiliation was not available in the reviewed primary bibliographic metadata. Publication status could not be established from the reviewed primary bibliographic metadata.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC10936310/
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: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 22 Nov 2026. Review scope: Primary result-bearing material for the stated method, result and evaluation, with bibliographic version information..
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
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.
Source check: Source checked. Not independent scientific confirmation.
Finding / tested result: Retrospective cohort (630 patients with advanced cancer) comparing ML algorithms for 180-day mortality from patient-reported symptoms, a clinically specific reference assessment.
Technology: Machine learning
Publication: Publication status unspecified in annotation · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- Retrospective cohort (630 patients with advanced cancer) comparing ML algorithms for 180-day mortality from patient-reported symptoms, a clinically specific reference assessment.
- Limitation / still unknown
- Affiliation not stated in the reviewed material. Affiliation was not available in the reviewed primary bibliographic metadata. Publication status could not be established from the reviewed primary bibliographic metadata.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://doi.org/10.1007/s11136-022-03284-y
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: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 22 Nov 2026. Review scope: Primary result-bearing material for the stated method, result and evaluation, with bibliographic version information..
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
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.
Source check: Source checked. Not independent scientific confirmation.
Finding / tested result: An autoencoder integrated RNA, miRNA, and methylation features, then k-means separated stomach-adenocarcinoma patients into prognosis groups associated with survival.
Technology: Deep Learning; autoencoder
Publication: Publication status unspecified in annotation · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- An autoencoder integrated RNA, miRNA, and methylation features, then k-means separated stomach-adenocarcinoma patients into prognosis groups associated with survival.
- Limitation / still unknown
- Affiliation not stated in the reviewed material. Affiliation was not available in the reviewed primary bibliographic metadata. Publication status could not be established from the reviewed primary bibliographic metadata.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC9633210/#methods · https://pmc.ncbi.nlm.nih.gov/articles/PMC9633210/#results · https://pmc.ncbi.nlm.nih.gov/articles/PMC9633210/
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: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 22 Nov 2026. Review scope: Primary result-bearing material for the stated method, result and evaluation, with bibliographic version information..
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
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.
Source check: Source checked. Not independent scientific confirmation.
Finding / tested result: Normative MEG modeling against 621 healthy participants produced deviation maps used by an SVM to distinguish 25 mTBI cases from 20 controls; theta-band deviations were discriminative.
Technology: SVM; machine learning
Publication: Preprint · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- Normative MEG modeling against 621 healthy participants produced deviation maps used by an SVM to distinguish 25 mTBI cases from 20 controls; theta-band deviations were discriminative.
- Limitation / still unknown
- Affiliation not stated in the reviewed material. Affiliation was not available in the reviewed primary bibliographic metadata.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://www.medrxiv.org/content/10.1101/2022.09.29.22280521v1.full
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 22 Nov 2026. Review scope: Primary result-bearing material for the stated method, result and evaluation, with bibliographic version information..
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
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.
Source check: Source checked. Not independent scientific confirmation.
Finding / tested result: Original study uses measured water samples from the 90-km Lirima Gravity Flow Scheme in Uganda to model residual chlorine concentration from temperature, pH, conductivity and sample time. It reports field-specific environmental water-quality prediction and compares learned model/ensemble performance. The inference is limited to the sampled scheme and retrospective 70/30 series split; no broader safety outcome is established.
Technology: Deep Learning; Machine Learning
Publication: Publication status unspecified in annotation · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- Methods §2.1.3: repeated morning/afternoon samples at several locations, with lab-measured temperature, pH, conductivity, residual chlorine and sampling time; 70/30 training/testing subseries. Results compare 6 deep and 9 ML models; Ens2 explains 74% variance on combined data. Field measurement and validation support a narrow environmental prediction result.
- Limitation / still unknown
- Affiliation not stated in the reviewed material. Affiliation was not available in the reviewed primary bibliographic metadata. Publication status could not be established from the reviewed primary bibliographic metadata.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://downloads.hindawi.com/journals/jeph/2022/7104752.pdf
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: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 22 Nov 2026. Review scope: Primary result-bearing material for the stated method, result and evaluation, with bibliographic version information..
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
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.
Source check: Source checked. Not independent scientific confirmation.
Finding / tested result: ArXiv v1 PDF (2022-09-16) and journal source compared: two CNNs trained on simulated data were tested on clinical data; median localization error 32 mm, up to 82% transmural detection, top-three localization error 20 mm. Same result lineage as the 2023 journal article; proof-of-principle only. These measurements establish a proof of principle in this study, not deployment benefit.
Technology: CNNs
Publication: Publication status unspecified in annotation · Test context: direct test. Catalog categories, not an evidence grade.
- AI contribution
- Two CNNs trained on simulated data mapped clinical semen images; the reported localization errors and detection rate are study-specific proof-of-principle results.
- Limitation / still unknown
- Affiliation not stated in the reviewed material. Affiliation was not available in the reviewed primary bibliographic metadata. Publication status could not be established from the reviewed primary bibliographic metadata.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://link.springer.com/article/10.1007/s10815-025-03453-1 · https://api.crossref.org/works/10.1007%2Fs10815-025-03453-1 · https://europepmc.org/search?query=10.1007/s10815-025-03453-1 · https://arxiv.org/abs/2209.08095 · https://arxiv.org/pdf/2209.08095
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: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 22 Nov 2026. Review scope: Primary result-bearing material for the stated method, result and evaluation, with bibliographic version information..
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
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.
Source check: Source checked. Not independent scientific confirmation.
Finding / tested result: Original Methods/Results report a retrospective clinical prognosis-classification study using VGG-19 on full-scan melanoma pathology images from CPTAC-CM, comparing prognosis groups. This is a specific clinical prognostic-validation result; interpret as preliminary retrospective model performance, not clinical benefit.
Technology: Deep Learning; VGG-19
Publication: Publication status unspecified in annotation · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- Cohort: 53 melanoma patients with known survival status (12 deceased, 41 living); VGG-19 image classification; random patient/image split details are imperfectly described, 5-fold CV; mean AUC 0.769±0.037, while SVM deep-feature AUC is 0.65. The source supports a bounded retrospective prognostic result and explicitly says performance is not yet satisfactory.
- Limitation / still unknown
- Affiliation not stated in the reviewed material. Affiliation was not available in the reviewed primary bibliographic metadata. Publication status could not be established from the reviewed primary bibliographic metadata.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://onlinelibrary.wiley.com/doi/10.1155/2022/4864485
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: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 22 Nov 2026. Review scope: Primary result-bearing material for the stated method, result and evaluation, with bibliographic version information..
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
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.
Source check: Source checked. Not independent scientific confirmation.
Finding / tested result: An ensemble model prospectively predicted 2019 OUD diagnoses in Alberta; performance is retrospective-data based and population-specific.
Technology: Class-balanced ensemble of machine-learning classifiers
Publication: Publication status unspecified in annotation · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- A class-balanced ensemble used linked Canadian administrative-health features for individualized OUD diagnosis prediction; it was evaluated on a non-overlapping 2019 sample.
- Limitation / still unknown
- Affiliation not stated in the reviewed material. Affiliation was not available in the reviewed primary bibliographic metadata. Publication status could not be established from the reviewed primary bibliographic metadata.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://journals.sagepub.com/doi/10.1177/07067437221114094
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: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 22 Nov 2026. Review scope: Primary result-bearing material for the stated method, result and evaluation, with bibliographic version information..
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
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.
Source check: Source checked. Not independent scientific confirmation.
Finding / tested result: A random-forest model analyzes prospective multicenter follow-up to identify predictors of minimal disease activity in recent-onset PsA; the association claim remains observational.
Technology: random-forest; machine learning
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- A random-forest model analyzes prospective multicenter follow-up to identify predictors of minimal disease activity in recent-onset PsA; the association claim remains observational.
- Limitation / still unknown
- Affiliation not stated in the reviewed material. Affiliation was not available in the reviewed primary bibliographic metadata.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://link.springer.com/article/10.1186/s13075-022-02838-2
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 22 Nov 2026. Review scope: Primary result-bearing material for the stated method, result and evaluation, with bibliographic version information..
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
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.
Source check: Source checked. Not independent scientific confirmation.
Finding / tested result: A randomized ML-triggered ACP intervention showed differing conversation-rate responses across data-derived oncologist profiles; no downstream outcome was tested.
Technology: machine learning; ann
Publication: Publication status unspecified in annotation · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- A learned mortality-risk model selected patients for clinician prompts; latent profiles then stratified observed response.
- Limitation / still unknown
- Affiliation not stated in the reviewed material. Affiliation was not available in the reviewed primary bibliographic metadata. Publication status could not be established from the reviewed primary bibliographic metadata.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0267012
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: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 22 Nov 2026. Review scope: Primary result-bearing material for the stated method, result and evaluation, with bibliographic version information..
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
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.
Source check: Source checked. Not independent scientific confirmation.
Finding / tested result: A locked eight-gene XGBoost model predicts severe-dengue progression and is tested prospectively in 377 Colombian patients before progression; the abstract supports a specific clinical result.
Technology: XGBoost; machine learning
Publication: Publication status unspecified in annotation · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- A locked eight-gene XGBoost model predicts severe-dengue progression and is tested prospectively in 377 Colombian patients before progression; the abstract supports a specific clinical result.
- Limitation / still unknown
- Affiliation not stated in the reviewed material. Affiliation was not available in the reviewed primary bibliographic metadata. Publication status could not be established from the reviewed primary bibliographic metadata.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://link.springer.com/article/10.1186/s13073-022-01034-w
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: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 22 Nov 2026. Review scope: Primary result-bearing material for the stated method, result and evaluation, with bibliographic version information..
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
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.
Source check: Source checked. Not independent scientific confirmation.
Finding / tested result: SZTrack’s CNN–RNN localizes seizure onset from scalp EEG, with patient-held-out testing and cross-site evaluation; this is retrospective evidence.
Technology: CNN; neural network; CNN–RNN
Publication: Publication status unspecified in annotation · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- CNN encodings plus recurrent modeling track electrode-level seizure activity and aggregate predictions into patient-level localization.
- Limitation / still unknown
- Affiliation not stated in the reviewed material. Affiliation was not available in the reviewed primary bibliographic metadata. Publication status could not be established from the reviewed primary bibliographic metadata.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0264537
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: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 22 Nov 2026. Review scope: Primary result-bearing material for the stated method, result and evaluation, with bibliographic version information..
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
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.
Source check: Source checked. Not independent scientific confirmation.
Finding / tested result: Patient-level test evaluation supports a specific CT transition-zone localization result for small-bowel obstruction; single-center and radiologist-annotation limits apply.
Technology: deep learning; ann
Publication: Publication status unspecified in annotation · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- Patient-level test evaluation supports a specific CT transition-zone localization result for small-bowel obstruction; single-center and radiologist-annotation limits apply.
- Limitation / still unknown
- Affiliation not stated in the reviewed material. Affiliation was not available in the reviewed primary bibliographic metadata. Publication status could not be established from the reviewed primary bibliographic metadata.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://link.springer.com/article/10.1186/s13244-021-01150-y
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: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 22 Nov 2026. Review scope: Primary result-bearing material for the stated method, result and evaluation, with bibliographic version information..
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
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.
Source check: Source checked. Not independent scientific confirmation.
Finding / tested result: retrospective single-center cohort 2,272 COVID-19 patients; propensity matching/IPTW and unsupervised four-cluster stratification; reported heterogeneous anticoagulation associations across clusters, including possible benefit cluster 3. Observational confounding precludes causal treatment claim.
Technology: Machine Learning
Publication: Publication status unspecified in annotation · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- Unsupervised clustering defines subphenotypes that structure a treatment-outcome association; observational analysis, not causal proof.
- Limitation / still unknown
- Affiliation not stated in the reviewed material. Affiliation was not available in the reviewed primary bibliographic metadata. Publication status could not be established from the reviewed primary bibliographic metadata.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://www.frontiersin.org/journals/medicine/articles/10.3389/fmed.2021.786414/full
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: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 22 Nov 2026. Review scope: Primary result-bearing material for the stated method, result and evaluation, with bibliographic version information..
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
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.
Source check: Source checked. Not independent scientific confirmation.
Finding / tested result: LSTM-SOM clustering of TCGA mutations using flanking sequence features yielded ten mutation classes associated with pathology and patient clinical/survival groupings; these are computational associations, not causal mechanisms.
Technology: machine learning
Publication: Publication status unspecified in annotation · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- LSTM extracts sequence features and SOM clusters 2,141,527 SBS from 9,596 TCGA patients; ten mutation-blot classes and K-means patient groups are linked with tumor pathology, clinical features, and survival.
- Limitation / still unknown
- Affiliation not stated in the reviewed material. Affiliation was not available in the reviewed primary bibliographic metadata. Publication status could not be established from the reviewed primary bibliographic metadata.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://link.springer.com/article/10.1186/s12920-021-01144-1
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: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 22 Nov 2026. Review scope: Primary result-bearing material for the stated method, result and evaluation, with bibliographic version information..
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
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.
Source check: Source checked. Not independent scientific confirmation.
Finding / tested result: An XGBoost model detected phasic sleep-bruxism episodes from mandibular movement signals in OSA patients, with validation against sleep-study labels.
Technology: XGBoost
Publication: Publication status unspecified in annotation · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- Methods describe movement-derived features and validation; Results report validation agreement and balanced accuracy.
- Limitation / still unknown
- Affiliation not stated in the reviewed material. Affiliation was not available in the reviewed primary bibliographic metadata. Publication status could not be established from the reviewed primary bibliographic metadata.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC8397703/
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: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 22 Nov 2026. Review scope: Primary result-bearing material for the stated method, result and evaluation, with bibliographic version information..
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
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.
Source check: Source checked. Not independent scientific confirmation.
Finding / tested result: Machine-learning models predicted clinically confirmed anastomotic leakage after esophagectomy in a 710-patient cohort with held-out testing.
Technology: Machine Learning; Machine-learning
Publication: Publication status unspecified in annotation · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- Methods define clinical leakage outcome and train/test model comparison; Results report performance and associated factors.
- Limitation / still unknown
- Affiliation not stated in the reviewed material. Affiliation was not available in the reviewed primary bibliographic metadata. Publication status could not be established from the reviewed primary bibliographic metadata.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC8367102/
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: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 22 Nov 2026. Review scope: Primary result-bearing material for the stated method, result and evaluation, with bibliographic version information..
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
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.
Source check: Source checked. Not independent scientific confirmation.
Finding / tested result: COVID-Net CXR-S estimates COVID-19 severity from chest radiographs using a deep convolutional model.
Technology: Convolutional Neural Network; Neural Network
Publication: Preprint · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- COVID-Net CXR-S estimates COVID-19 severity from chest radiographs using a deep convolutional model.
- Limitation / still unknown
- Affiliation not stated in the reviewed material. Affiliation was not available in the reviewed primary bibliographic metadata.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://doi.org/10.21203/rs.3.rs-580218/v1
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 22 Nov 2026. Review scope: Primary result-bearing material for the stated method, result and evaluation, with bibliographic version information..
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
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.
Source check: Source checked. Not independent scientific confirmation.
Finding / tested result: SciMARL uses cooperating agents to learn LES wall models that reproduce turbulent-flow quantities at substantially lower computational cost and generalize to unseen geometries and Reynolds numbers; this is a specific computational-physics result.
Technology: SciMARL multi-agent reinforcement learning for LES wall modeling
Publication: Preprint · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- SciMARL trains cooperating agents to learn large-eddy-simulation wall models; the paper reports turbulent-flow quantities, lower computation cost, and tests on unseen geometries and Reynolds numbers.
- Limitation / still unknown
- Affiliation not stated in the reviewed material. Affiliation was not available in the reviewed primary bibliographic metadata.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://www.researchsquare.com/article/rs-573667/v1
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 22 Nov 2026. Review scope: Primary result-bearing material for the stated method, result and evaluation, with bibliographic version information..
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
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.
Source check: Source checked. Not independent scientific confirmation.
Finding / tested result: Specific clinical biomarker/severity result arises from measured COVID cohort and ML analysis with comparison groups; qualifies narrowly as retrospective clinical result. Version is non-peer-reviewed v1 and association does not establish causation.
Technology: Machine Learning
Publication: Preprint · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- Original v1 Methods: retrospective 126 PCR-confirmed hospitalized COVID-19 patients, 50 healthy and 45 pre-pandemic hospital controls; measured PON1/inflammatory markers and ML variable selection/classification assess severity. Results support specific PON1-related variables and inflammatory biomarkers differentiating severity groups. Preprint, single-center, retrospective; markers are associations, not causal or prospective predictors.
- Limitation / still unknown
- Affiliation not stated in the reviewed material. Affiliation was not available in the reviewed primary bibliographic metadata.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://www.preprints.org/manuscript/202105.0623/v1/download
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: 22 Nov 2026. Review scope: Primary result-bearing material for the stated method, result and evaluation, with bibliographic version information..
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
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.
Source check: Source checked. Not independent scientific confirmation.
Finding / tested result: knowledge distillation produces an efficient student COVID case projection model with lower error than CDC ensemble during the stated short retrospective window; not prospective or causal evidence.
Technology: Deep Learning
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- SEIR mixture teacher generates sequences; sequence mixup enriches queries; student MLP learns projections.
- Limitation / still unknown
- Affiliation not stated in the reviewed material.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://cdn.aaai.org/ojs/17812/17812-13-21306-1-2-20210518.pdf · https://ojs.aaai.org/index.php/AAAI/article/view/17812
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: 22 Nov 2026. Review scope: Primary result-bearing material for the stated method, result and evaluation, with bibliographic version information..
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
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.
Source check: Source checked. Not independent scientific confirmation.
Finding / tested result: Among 4,144 admissions, RF sensitivity/specificity were 89.9%/85.8%, outperforming the logistic model on sensitivity.
Technology: Machine Learning
Publication: Publication status unspecified in annotation · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- Random forest compared with logistic regression for in-hospital myocarditis mortality from KID variables.
- Limitation / still unknown
- Affiliation not stated in the reviewed material. Affiliation was not available in the reviewed primary bibliographic metadata. Publication status could not be established from the reviewed primary bibliographic metadata.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://doi.org/10.3389/fped.2021.644922
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: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 22 Nov 2026. Review scope: Primary result-bearing material for the stated method, result and evaluation, with bibliographic version information..
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
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.
Source check: Source checked. Not independent scientific confirmation.
Finding / tested result: An improved U-Net segmented COVID-19 lung lesions on CT, with a reader comparison reporting segmentation and diagnostic time; this is a specific imaging result with limitations of retrospective data.
Technology: Improved U-Net for CT lung and lesion segmentation
Publication: Publication status unspecified in annotation · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- An improved U-Net segmented lungs and COVID-19 lesions on CT; the paper compares segmentation metrics and physician diagnostic time.
- Limitation / still unknown
- Affiliation not stated in the reviewed material. Affiliation was not available in the reviewed primary bibliographic metadata. Publication status could not be established from the reviewed primary bibliographic metadata.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC8009533/
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: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 22 Nov 2026. Review scope: Primary result-bearing material for the stated method, result and evaluation, with bibliographic version information..
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
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.
Source check: Source checked. Not independent scientific confirmation.
Finding / tested result: Task-based fMRI connectivity features in 80 participants were used to identify shared functional connections associated with ADHD diagnosis and behavioral task performance; the paper reports group-specific connectivity/diagnostic patterns. This is a small retrospective neuroimaging association, not a causal mechanism.
Technology: Machine-learning classifier on task-based fMRI functional-connectivity features
Publication: Publication status unspecified in annotation · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- A classifier analyzed task-based fMRI functional connectivity to identify ADHD-associated patterns and links with behavioral-task performance in 80 participants; findings are observational.
- Limitation / still unknown
- Affiliation not stated in the reviewed material. Affiliation was not available in the reviewed primary bibliographic metadata. Publication status could not be established from the reviewed primary bibliographic metadata.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://www.frontiersin.org/journals/physiology/articles/10.3389/fphys.2020.583005/full
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: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 22 Nov 2026. Review scope: Primary result-bearing material for the stated method, result and evaluation, with bibliographic version information..
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
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.
Source check: Source checked. Not independent scientific confirmation.
Finding / tested result: A radiology AI system is deployed in a clinical infrastructure to process chest radiographs for COVID-19 inference and triage.
Technology: PXS Score convolutional Siamese neural network within a radiology deployment pipeline
Publication: Preprint · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- The paper deployed the PXS Score deep-learning model for chest-radiograph severity scoring through live DICOM ingestion; prospective evaluation covered 135 hospitalized COVID-19 patients, with AUROC 0.81 for a three-day composite outcome.
- Limitation / still unknown
- Affiliation not stated in the reviewed material. Affiliation was not available in the reviewed primary bibliographic metadata.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://doi.org/10.21203/rs.3.rs-61220/v1 · https://assets-eu.researchsquare.com/files/rs-61220/v1_covered.pdf
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 22 Nov 2026. Review scope: Primary result-bearing material for the stated method, result and evaluation, with bibliographic version information..
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
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.
Source check: Source checked. Not independent scientific confirmation.
Finding / tested result: Feedforward neural network predicts adsorption capacity of DES-functionalized CNTs for 2,4-DCP removal, validated against experimental adsorption data.
Technology: Neural Network
Publication: Publication status unspecified in annotation · Test context: direct test. Catalog categories, not an evidence grade.
- AI contribution
- Feedforward neural network predicts adsorption capacity of DES-functionalized CNTs for 2,4-DCP removal, validated against experimental adsorption data.
- Limitation / still unknown
- Affiliation not stated in the reviewed material. Affiliation was not available in the reviewed primary bibliographic metadata. Publication status could not be established from the reviewed primary bibliographic metadata.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://mdpi-res.com/d_attachment/molecules/molecules-25-01511/article_deploy/molecules-25-01511.pdf
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: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 22 Nov 2026. Review scope: Primary result-bearing material for the stated method, result and evaluation, with bibliographic version information..
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
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.
Source check: Source checked. Not independent scientific confirmation.
Finding / tested result: Machine learning on parental electronic medical records estimated later autism diagnosis risk among children in a population-based cohort; the paper reports held-out predictive performance.
Technology: machine learning
Publication: Publication status unspecified in annotation · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- Methods define parent-record predictors, training/validation split and 10-fold evaluation; Results report future ASD risk prediction and limitations.
- Limitation / still unknown
- Affiliation not stated in the reviewed material. Affiliation was not available in the reviewed primary bibliographic metadata. Publication status could not be established from the reviewed primary bibliographic metadata.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC7315872/
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: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 22 Nov 2026. Review scope: Primary result-bearing material for the stated method, result and evaluation, with bibliographic version information..
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
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.