{"id":"W4385365799","doi":"10.1016/j.jacr.2023.06.025","title":"“Shortcuts” Causing Bias in Radiology Artificial Intelligence: Causes, Evaluation, and Mitigation","year":2023,"lang":"en","type":"review","venue":"Journal of the American College of Radiology","topic":"Artificial Intelligence in Healthcare and Education","field":"Medicine","cited_by":99,"is_retracted":false,"has_abstract":false,"ca_institutions":"York University","funders":"National Institute on Minority Health and Health Disparities; National Institute of Biomedical Imaging and Bioengineering; National Institute of Mental Health; Radiological Society of North America; National Bureau of Economic Research; National Heart, Lung, and Blood Institute; U.S. Department of Defense; National Cancer Institute; National Institutes of Health; National Science Foundation","keywords":"Artificial intelligence; Inference; Spurious relationship; Computer science; Machine learning; Metric (unit); Race (biology)","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":[],"consensus_categories":[],"category_scores_codex":[0.003281128,0.0002480872,0.002353018,0.0008814871,0.00008406128,0.000006166763,0.000209151,0.0002529164,0.00002370642],"category_scores_gemma":[0.003767467,0.0001673958,0.0003579848,0.00151524,0.000826682,0.00005476012,0.0000527345,0.0007845893,0.000004812059],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005520951,"about_ca_system_score_gemma":0.003003299,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0002825943,"about_ca_topic_score_gemma":0.000260964,"domain_scores_codex":[0.9952284,0.001502228,0.002370933,0.0002574471,0.0003480059,0.000293062],"domain_scores_gemma":[0.9942297,0.001689687,0.003027459,0.0003226672,0.0006097356,0.0001206933],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0001604991,0.00008752657,0.001446825,0.001864388,0.0003102242,0.00009583099,0.0004968211,0.0001019515,0.00001959437,0.0004796517,0.001517598,0.9934191],"study_design_scores_gemma":[0.0007503312,0.01661147,0.01054622,0.09948395,0.02400851,0.180023,0.04090191,0.006586774,0.001143561,0.0906039,0.5261399,0.00320054],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"review","genre_gemma":"review","genre_scores_codex":[0.0732357,0.9207501,0.00010063,0.002607661,0.001994694,0.001244802,0.00003657544,0.0000084031,0.000021477],"genre_scores_gemma":[0.0208411,0.9776968,0.0003389055,0.0001374638,0.0008711977,0.00002633187,0.00001284952,0.0000360497,0.00003933558],"genre_candidate":"review","genre_consensus":"review","teacher_disagreement_score":0.9902185,"threshold_uncertainty_score":0.6826205,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.3374095107685531,"score_gpt":0.497149325622682,"score_spread":0.1597398148541289,"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."}}