{"id":"W7117729834","doi":"10.1186/s12874-025-02703-1","title":"Machine learning models explanations as interpretations of evidence: a theoretical framework of explainability and its implications on high-stakes biomedical decision-making","year":2025,"lang":"en","type":"article","venue":"BMC Medical Research Methodology","topic":"Explainable Artificial Intelligence (XAI)","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of British Columbia","funders":"European Commission","keywords":"Operationalization; Terminology; Soundness; Interpretation (philosophy); Interpretability; Raising (metalworking)","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.02507632,0.001103128,0.0009812148,0.005835123,0.00180785,0.00623311,0.003893254,0.004099049,0.006110772],"category_scores_gemma":[0.08162474,0.0007958285,0.002017623,0.003032766,0.01892737,0.01260067,0.006075005,0.00426957,0.0003696023],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.004671479,"about_ca_system_score_gemma":0.002965666,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003098723,"about_ca_topic_score_gemma":0.002434538,"domain_scores_codex":[0.9768563,0.01807195,0.001015101,0.001710388,0.001754296,0.0005919642],"domain_scores_gemma":[0.8300922,0.1499892,0.008091643,0.007532654,0.003049299,0.001245044],"domain_codex":null,"domain_gemma":"methods","domain_candidate":"methods","domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.00001522666,0.00002322977,0.0007430965,0.0001181773,0.00004441887,0.0001253776,0.0009770395,0.01021386,0.00008314785,0.9825508,0.0003780769,0.004727578],"study_design_scores_gemma":[0.0000129906,0.00001053694,0.0001470303,0.00008509283,0.00001689036,0.00005149086,0.0001788007,0.02954436,0.00008463758,0.9684306,0.001425602,0.00001199519],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.02107355,0.001332999,0.9419158,0.02008454,0.0000942062,0.0001961801,0.0002779049,0.000157816,0.01486699],"genre_scores_gemma":[0.750447,0.0008740662,0.2449747,0.001088049,0.0002681115,0.000449016,0.0003020855,0.00006374629,0.001533252],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.9749237,"threshold_uncertainty_score":0.1326178,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.4858301340420024,"score_gpt":0.5662124117095355,"score_spread":0.08038227766753309,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}