{"id":"W4388801023","doi":"10.1101/2023.11.15.23298499","title":"Spot the Difference: Can ChatGPT4-Vision Transform Radiology Artificial Intelligence?","year":2023,"lang":"en","type":"preprint","venue":"medRxiv","topic":"COVID-19 diagnosis using AI","field":"Medicine","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"Artificial Intelligence in Medicine (Canada)","funders":"Science Foundation Ireland; Health Service Executive; Royal College of Surgeons in Ireland; Wellcome Trust; Canadian Institute for Theoretical Astrophysics","keywords":"Computer science; Artificial intelligence; Coding (social sciences); Recall; Machine learning; F1 score; Task (project management); Psychology; Statistics; Cognitive psychology; Engineering; Mathematics","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":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.001165276,0.0005060573,0.0009040219,0.000387983,0.000205393,0.00007889803,0.0007274431,0.0006412941,0.0002684574],"category_scores_gemma":[0.0008538192,0.0003413359,0.0004135285,0.000447474,0.0004405182,0.00001770132,0.0004166059,0.001805237,0.0002730789],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003153421,"about_ca_system_score_gemma":0.0006316912,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001157256,"about_ca_topic_score_gemma":0.001309559,"domain_scores_codex":[0.9967534,0.0002180656,0.0007879028,0.001000459,0.000612206,0.0006279687],"domain_scores_gemma":[0.9970003,0.001032626,0.000207864,0.001379811,0.0001466762,0.0002326717],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"observational","study_design_scores_codex":[0.001388308,0.001814012,0.06584185,0.004104522,0.001415286,0.002019485,0.02585945,0.001519322,0.01978504,0.006602402,0.02597992,0.8436704],"study_design_scores_gemma":[0.001080039,0.001793602,0.6589819,0.005332496,0.002110345,0.0003487092,0.001323295,0.01553472,0.07985831,0.1614159,0.06966653,0.002554132],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.7011957,0.0003256841,0.003366225,0.289423,0.003645245,0.00142202,0.00006733638,0.0004386689,0.0001160981],"genre_scores_gemma":[0.9893159,0.001296401,0.0001410606,0.006298355,0.001378868,0.0002964844,0.0001651962,0.0001139645,0.000993743],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.8411162,"threshold_uncertainty_score":0.9999039,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.09104601245486954,"score_gpt":0.3611499249292693,"score_spread":0.2701039124743997,"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."}}