{"id":"W4396647722","doi":"10.2196/56627","title":"Advancing Accuracy in Multimodal Medical Tasks Through Bootstrapped Language-Image Pretraining (BioMedBLIP): Performance Evaluation Study","year":2024,"lang":"en","type":"article","venue":"JMIR Medical Informatics","topic":"Multimodal Machine Learning Applications","field":"Computer Science","cited_by":5,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"","keywords":"Preprint; Computer science; Artificial intelligence; Natural language processing; Computer vision; World Wide Web","routes":{"ca_aff":false,"ca_fund":false,"ca_venue":true,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.004175759,0.0002958929,0.0003417718,0.0003650638,0.0001635769,0.0003448393,0.001589371,0.000274554,0.0004911217],"category_scores_gemma":[0.001795427,0.0002530158,0.00007826392,0.001464484,0.0001438887,0.002464595,0.0005460601,0.001558397,0.0003033568],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002203098,"about_ca_system_score_gemma":0.001058936,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001928658,"about_ca_topic_score_gemma":0.00006857108,"domain_scores_codex":[0.9941694,0.0002335802,0.001232813,0.000388412,0.003344476,0.000631344],"domain_scores_gemma":[0.9976522,0.0008976618,0.0001927711,0.0007601622,0.0001177513,0.0003794655],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00001069605,0.0005136951,0.003163019,0.0003997402,0.00004937028,0.00009124005,0.1593266,0.0005187844,0.0000641084,0.001395218,0.0004634654,0.8340041],"study_design_scores_gemma":[0.001221156,0.0001506876,0.01711226,0.0004576983,0.00001239968,0.00005375276,0.004689606,0.9752806,0.00004076216,0.0000804369,0.0006263702,0.0002742663],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8336233,0.000106392,0.159129,0.001694437,0.000398202,0.001417557,0.000003854818,0.0006389476,0.002988293],"genre_scores_gemma":[0.961651,0.00002382944,0.03692906,0.0005974742,0.0001762065,0.0005421108,0.00004250274,0.00002327374,0.00001458369],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.9747618,"threshold_uncertainty_score":0.9999922,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02037923269398638,"score_gpt":0.3870688965330144,"score_spread":0.366689663839028,"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."}}