{"id":"W4399301274","doi":"10.1007/s00134-024-07491-8","title":"Why federated learning will do little to overcome the deeply embedded biases in clinical medicine","year":2024,"lang":"en","type":"letter","venue":"Intensive Care Medicine","topic":"Machine Learning in Healthcare","field":"Computer Science","cited_by":6,"is_retracted":false,"has_abstract":true,"ca_institutions":"Artificial Intelligence in Medicine (Canada)","funders":"Fogarty International Center; National Institute of Biomedical Imaging and Bioengineering; Ministerium für Kultur und Wissenschaft des Landes Nordrhein-Westfalen; Deutsche Forschungsgemeinschaft; National Science Foundation; National Institutes of Health; Universitätsklinikum Essen","keywords":"Data science; Process (computing); Health care; Medicine; Scale (ratio); Quality (philosophy); Data sharing; Big data; MEDLINE; Computer science; Intensive care; Artificial intelligence; Internet privacy; Data mining; Alternative medicine; Intensive care medicine","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":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","metaepi_narrow","research_integrity"],"consensus_categories":[],"category_scores_codex":[0.002195585,0.0009403795,0.001967881,0.001087113,0.0002899793,0.0002126765,0.002425834,0.001014287,0.0003189828],"category_scores_gemma":[0.05186252,0.0005694251,0.0002735562,0.002187481,0.0007100043,0.000185741,0.001173536,0.01173317,0.0002826943],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005329965,"about_ca_system_score_gemma":0.0004356284,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004844106,"about_ca_topic_score_gemma":0.0005547372,"domain_scores_codex":[0.9908154,0.002136885,0.002065735,0.002050568,0.001700972,0.001230463],"domain_scores_gemma":[0.9842491,0.007589061,0.0005405337,0.001740939,0.005485989,0.000394427],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00002633425,0.000004881418,0.00184556,0.0004343778,0.00008023931,0.004151171,0.0183344,0.0001771389,0.000007352612,0.00008085426,0.9637058,0.01115188],"study_design_scores_gemma":[0.0008178679,0.001995085,0.001290136,0.01035953,0.0001003276,0.0003025921,0.008278782,0.00508912,0.000003897339,0.0003078529,0.9708521,0.0006026414],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"commentary","genre_gemma":"commentary","genre_scores_codex":[0.001510078,0.01226919,0.002805865,0.9698064,0.01052711,0.001135513,0.00001326855,0.0005501322,0.00138242],"genre_scores_gemma":[0.02147941,0.0002736095,0.0002660393,0.9632282,0.01233933,0.0001110914,0.0002842949,0.0001498554,0.001868204],"genre_candidate":"commentary","genre_consensus":"commentary","teacher_disagreement_score":0.04966694,"threshold_uncertainty_score":0.9996757,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.05626682739397904,"score_gpt":0.3848491796676702,"score_spread":0.3285823522736911,"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."}}