{"id":"W4390580666","doi":"10.1177/12034754231220934","title":"Evaluation of Vision LLMs GTP-4V and LLaVA for the Recognition of Features Characteristic of Melanoma","year":2024,"lang":"en","type":"letter","venue":"Journal of Cutaneous Medicine and Surgery","topic":"AI in cancer detection","field":"Computer Science","cited_by":9,"is_retracted":false,"has_abstract":false,"ca_institutions":"Dermatrials Research; Western University; McMaster University; University of Toronto","funders":"","keywords":"Medicine; Melanoma; Dermatology; Artificial intelligence; Cancer research","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":[0.0007761202,0.0002655312,0.0003858669,0.0008850216,0.0002518595,0.00054329,0.0004559424,0.002094758,0.00168437],"category_scores_gemma":[0.002592059,0.0001423227,0.0003555647,0.000153667,0.0003815921,0.0004394341,0.0001821485,0.001343171,0.0008992639],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005993247,"about_ca_system_score_gemma":0.0002670939,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001670443,"about_ca_topic_score_gemma":0.003579802,"domain_scores_codex":[0.9997919,0.00004971391,0.00002190143,0.00002910836,0.00004842808,0.00005896476],"domain_scores_gemma":[0.9993076,0.000334075,0.00001705334,0.00005139051,0.000188623,0.00010124],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.01114549,0.001262028,0.1365928,0.0006423339,0.0002359254,0.1133868,0.0004490007,0.001176188,0.1778925,0.001797094,0.04183998,0.5135798],"study_design_scores_gemma":[0.0009317462,0.008944578,0.2637365,0.0003427225,0.000578472,0.3555717,0.001130472,0.09729651,0.2085669,0.004306066,0.0584297,0.0001646621],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9274336,0.008051567,0.008579178,0.03206474,0.001647201,0.0002455155,0.0002478925,0.0004316939,0.02129855],"genre_scores_gemma":[0.9899867,0.0007547379,0.003372351,0.003337912,0.0005411118,0.00004251748,0.0001030401,0.00002252391,0.001839],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.002094758,"threshold_uncertainty_score":0.005634785,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.05103814608982515,"score_gpt":0.3041818787602356,"score_spread":0.2531437326704105,"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."}}