{"id":"W3020729718","doi":"10.1016/j.jbi.2020.103424","title":"Building machine learning models without sharing patient data: A simulation-based analysis of distributed learning by ensembling","year":2020,"lang":"en","type":"article","venue":"Journal of Biomedical Informatics","topic":"Machine Learning in Healthcare","field":"Computer Science","cited_by":39,"is_retracted":false,"has_abstract":false,"ca_institutions":"Alberta Children's Hospital; University of Calgary","funders":"National Institute of Biomedical Imaging and Bioengineering; GE Healthcare; Janssen Research and Development; National Institutes of Health; H. Lundbeck A/S; Johnson and Johnson; Calgary Foundation; Northern California Institute for Research and Education; U.S. Department of Defense; Alzheimer's Disease Neuroimaging Initiative; University of Southern California; Genentech; IXICO; National Institute on Aging; Fujirebio Europe; Heart and Stroke Foundation of Canada; Merck","keywords":"Computer science; Machine learning; Artificial intelligence; Data sharing; Medicine","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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001492295,0.0001986206,0.0007020608,0.0005930822,0.0001853201,0.0001672419,0.001522998,0.000124786,0.00001475495],"category_scores_gemma":[0.003064239,0.00016755,0.0001839936,0.002342569,0.00006547842,0.001189045,0.0006681501,0.001271449,0.000001274701],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0000886819,"about_ca_system_score_gemma":0.0001957143,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0000418084,"about_ca_topic_score_gemma":7.883593e-7,"domain_scores_codex":[0.9958952,0.0001796657,0.002063713,0.0001922729,0.001341811,0.0003273118],"domain_scores_gemma":[0.9954444,0.0009191026,0.002337743,0.0003989834,0.00043214,0.0004675942],"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":[0.00002627714,0.00003931548,0.008848402,0.0001632331,0.0002661281,0.000006840916,0.003437913,0.9704285,0.00008527196,0.00007991708,0.00004217299,0.01657609],"study_design_scores_gemma":[0.0004809785,0.0004126474,0.00009503009,0.0001420844,0.0001887724,0.000004607572,0.0002148008,0.9949649,0.00004767019,0.00003788045,0.003255121,0.0001555042],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1091617,0.0001503064,0.8893646,0.00104756,0.00008503943,0.00007724861,0.00004210501,0.00006128444,0.00001012215],"genre_scores_gemma":[0.8770235,0.00001506422,0.1223134,0.0004018369,0.00004892144,5.743053e-7,0.0001847388,0.00001094166,0.000001031699],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.7678618,"threshold_uncertainty_score":0.6832492,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.0656222157227938,"score_gpt":0.3315865052443263,"score_spread":0.2659642895215325,"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."}}