{"id":"W4392407790","doi":"10.2196/55508","title":"Assessing the Utility of Multimodal Large Language Models (GPT-4 Vision and Large Language and Vision Assistant) in Identifying Melanoma Across Different Skin Tones","year":2024,"lang":"en","type":"article","venue":"JMIR Dermatology","topic":"Artificial Intelligence in Healthcare and Education","field":"Medicine","cited_by":38,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Toronto; Western University","funders":"","keywords":"Computer science; Artificial intelligence; Natural language processing; Computer vision","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":true,"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":[0.005627916,0.001234607,0.0004317373,0.001306699,0.0004362781,0.001860311,0.000623741,0.001646711,0.002402367],"category_scores_gemma":[0.02183833,0.0003022913,0.0009832807,0.0004146231,0.0004706902,0.001887488,0.0014795,0.00132359,0.001364635],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006932972,"about_ca_system_score_gemma":0.0009101406,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.008224977,"about_ca_topic_score_gemma":0.009161178,"domain_scores_codex":[0.9979401,0.001063255,0.0001155663,0.0004386855,0.000297729,0.000144639],"domain_scores_gemma":[0.9902714,0.007685859,0.000328403,0.0005088146,0.0008566512,0.0003488312],"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":[0.009614641,0.002940883,0.1864857,0.000924028,0.001697935,0.000787448,0.001373393,0.1258149,0.04680328,0.002223716,0.01463323,0.6067008],"study_design_scores_gemma":[0.0002788158,0.00309246,0.06944862,0.0001348584,0.0007228967,0.0009923875,0.001191998,0.8939768,0.02168356,0.004830281,0.003443482,0.0002038406],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9298211,0.001339281,0.05599892,0.0008576269,0.0001936916,0.0004482395,0.002046148,0.001876416,0.00741859],"genre_scores_gemma":[0.9551077,0.0002516997,0.03993959,0.0004036408,0.00005332812,0.000165115,0.002100115,0.0001250039,0.001853852],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.008224977,"threshold_uncertainty_score":0.02976364,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.06756033798632126,"score_gpt":0.4919907083386785,"score_spread":0.4244303703523573,"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."}}