{"id":"W3203233031","doi":"10.1002/jmri.27908","title":"Deep Generative Medical Image Harmonization for Improving Cross‐Site Generalization in Deep Learning Predictors","year":2021,"lang":"en","type":"article","venue":"Journal of Magnetic Resonance Imaging","topic":"Face recognition and analysis","field":"Computer Science","cited_by":84,"is_retracted":false,"has_abstract":true,"ca_institutions":"Artificial Intelligence in Medicine (Canada)","funders":"National Institute on Drug Abuse; National Institute of Mental Health; National Institute of Biomedical Imaging and Bioengineering; Ministry of Cultural Affairs; Universität Greifswald; Bundesministerium für Bildung und Forschung; National Institute on Aging; Siemens Healthineers; Schweizerischer Nationalfonds zur Förderung der Wissenschaftlichen Forschung; National Institutes of Health; National Science Foundation","keywords":"Generalizability theory; Artificial intelligence; Deep learning; Overfitting; Computer science; Neuroimaging; Generalization; Machine learning; Harmonization; Pattern recognition (psychology); Statistics; Artificial neural network; Medicine; Mathematics","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":[],"consensus_categories":[],"category_scores_codex":[0.00416522,0.001002442,0.001007374,0.001031196,0.0003841994,0.0008306275,0.001511067,0.001058494,0.002005329],"category_scores_gemma":[0.006386033,0.0005585007,0.001424474,0.0008774612,0.0008054379,0.001017256,0.001821571,0.001970533,0.000678506],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001042738,"about_ca_system_score_gemma":0.001002691,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005160852,"about_ca_topic_score_gemma":0.005584364,"domain_scores_codex":[0.999218,0.0002765153,0.00004577541,0.000249166,0.000119323,0.00009126327],"domain_scores_gemma":[0.9980053,0.001011827,0.0002122092,0.0003740416,0.0003263924,0.0000702333],"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":[0.0003644146,0.0002331522,0.008171538,0.0001083929,0.0003181067,0.0001295676,0.0002197467,0.5752295,0.01130802,0.003500294,0.004218494,0.3961988],"study_design_scores_gemma":[0.00001685087,0.00007091897,0.001041614,0.000009213573,0.0000308598,0.00003465202,0.00001348183,0.9924647,0.003593685,0.002177351,0.0005371491,0.000009588241],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1128595,0.0007034644,0.8809805,0.0004959739,0.00008624778,0.0001282513,0.0002826443,0.003139128,0.001324282],"genre_scores_gemma":[0.8146752,0.0002559257,0.1803139,0.0005591867,0.00008842168,0.000203837,0.001056157,0.0003988616,0.002448504],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.005160852,"threshold_uncertainty_score":0.02202803,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.007359170147348658,"score_gpt":0.2544498088827585,"score_spread":0.2470906387354098,"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."}}