{"id":"W4403603005","doi":"10.2196/60164","title":"Health Care Language Models and Their Fine-Tuning for Information Extraction: Scoping Review","year":2024,"lang":"en","type":"article","venue":"JMIR Medical Informatics","topic":"Topic Modeling","field":"Computer Science","cited_by":17,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"","keywords":"Terminology; Computer science; Health care; Data extraction; Scopus; Unified Medical Language System; Knowledge management; MEDLINE; Data science; Artificial intelligence; Linguistics","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.1280087,0.003103831,0.009082157,0.03696923,0.001983565,0.008463203,0.005672395,0.003502913,0.007735293],"category_scores_gemma":[0.414267,0.002425609,0.01466148,0.02915619,0.003327344,0.01112822,0.006854994,0.003206893,0.001361274],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.01076192,"about_ca_system_score_gemma":0.04288815,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0126732,"about_ca_topic_score_gemma":0.01752583,"domain_scores_codex":[0.9038832,0.04683156,0.03344337,0.004245263,0.01079001,0.0008065731],"domain_scores_gemma":[0.5391566,0.3928559,0.02897199,0.01169636,0.02653685,0.0007823993],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"systematic_review","study_design_gemma":"systematic_review","study_design_scores_codex":[0.0001447588,0.00004710281,0.0009752969,0.7895033,0.004526655,0.00009618213,0.001229962,0.0007993666,0.0001651185,0.003497307,0.003698736,0.1953162],"study_design_scores_gemma":[0.00006248623,0.00006752701,0.0007194835,0.9595695,0.01074974,0.0001090448,0.0005783004,0.0004241996,0.0002690257,0.002244533,0.02515793,0.00004841869],"study_design_candidate":"systematic_review","study_design_consensus":"systematic_review","genre_codex":"review","genre_gemma":"review","genre_scores_codex":[0.001102437,0.9801501,0.009237923,0.002070409,0.0004008659,0.003907403,0.001250408,0.0001314649,0.001749039],"genre_scores_gemma":[0.02023633,0.9344189,0.02911043,0.00135507,0.0002517667,0.01271417,0.001537204,0.00008496003,0.0002910807],"genre_candidate":"review","genre_consensus":"review","teacher_disagreement_score":0.1280087,"threshold_uncertainty_score":0.6769826,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03758902317528386,"score_gpt":0.3692044030291242,"score_spread":0.3316153798538404,"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."}}