{"id":"W4406284830","doi":"10.1038/s43246-024-00731-w","title":"Probing out-of-distribution generalization in machine learning for materials","year":2025,"lang":"en","type":"article","venue":"Communications Materials","topic":"Machine Learning in Materials Science","field":"Materials Science","cited_by":43,"is_retracted":false,"has_abstract":true,"ca_institutions":"Schwartz/Reisman Emergency Medicine Institute; Vector Institute; Structural Genomics Consortium; University of Toronto","funders":"Canada First Research Excellence Fund; University of Toronto; Alliance de recherche numérique du Canada; Western Canada Research Grid","keywords":"Generalization; Distribution (mathematics); Computer science; Artificial intelligence; Mathematics; Mathematical analysis","routes":{"ca_aff":true,"ca_fund":true,"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":[],"consensus_categories":[],"category_scores_codex":[0.008603171,0.001010975,0.0008912642,0.0009719942,0.0004842221,0.001377383,0.001405742,0.001320463,0.00253552],"category_scores_gemma":[0.03051612,0.0002812033,0.0008407675,0.0007206316,0.002060069,0.002876724,0.001916776,0.00248269,0.0005558464],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001114659,"about_ca_system_score_gemma":0.0007718395,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002449082,"about_ca_topic_score_gemma":0.002470566,"domain_scores_codex":[0.9978649,0.001124667,0.00008926931,0.0003168889,0.0004745067,0.0001298628],"domain_scores_gemma":[0.9841797,0.0112315,0.0006378277,0.002626437,0.0009355057,0.0003889539],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0005735967,0.0003570354,0.01222582,0.0005560346,0.0001702699,0.0001310131,0.0001780733,0.8841471,0.005463499,0.01636324,0.005107246,0.07472712],"study_design_scores_gemma":[0.00002395248,0.0002364104,0.001877585,0.00005212852,0.00001728577,0.00003752997,0.00005429676,0.9697325,0.004375554,0.02261942,0.0009597028,0.00001353084],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8307793,0.004458004,0.145877,0.002508054,0.0002480911,0.0001266932,0.0008072013,0.002663974,0.01253167],"genre_scores_gemma":[0.9785619,0.0002832929,0.01911938,0.0002490005,0.00004990031,0.00005146449,0.0007504618,0.0001959829,0.0007385976],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.008603171,"threshold_uncertainty_score":0.04549843,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03469835939800099,"score_gpt":0.3325907825709555,"score_spread":0.2978924231729546,"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."}}