{"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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.000440463,0.0001422882,0.0003307291,0.0001131517,0.0001450622,0.00009176305,0.00006423678,0.0001799848,0.00004276457],"category_scores_gemma":[0.0000693749,0.00009492105,0.00004898812,0.0001663624,0.00011504,0.000288261,0.000112431,0.0003011704,0.000004013261],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00003569955,"about_ca_system_score_gemma":0.00005003025,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0005542481,"about_ca_topic_score_gemma":0.0011876,"domain_scores_codex":[0.9985579,0.0001845689,0.000452014,0.0003092697,0.0001587278,0.0003374774],"domain_scores_gemma":[0.9992024,0.0003783655,0.00007538085,0.0002265065,0.00004447733,0.00007285842],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0002863366,0.0009946111,0.3453676,0.004313489,0.0000794122,0.0007488995,0.3186052,0.000002139881,0.02184421,0.001844993,0.0006605825,0.3052524],"study_design_scores_gemma":[0.0004042193,0.0001106942,0.6327788,0.001185159,0.00004051531,0.0005402897,0.1026854,0.2414758,0.01897162,0.001166857,0.0004210494,0.000219609],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9886886,0.006732744,0.002011023,0.001759973,0.0002411273,0.0004013635,0.00002063317,0.00004080649,0.0001037595],"genre_scores_gemma":[0.9993968,0.0001796739,0.0001018871,0.0001544515,0.00005902149,0.00003255686,0.00003299857,0.00001542321,0.00002718609],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.3050328,"threshold_uncertainty_score":0.3870768,"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."}}