{"id":"W4392193048","doi":"10.1038/s41591-024-02855-5","title":"Adapted large language models can outperform medical experts in clinical text summarization","year":2024,"lang":"en","type":"article","venue":"Nature Medicine","topic":"Topic Modeling","field":"Computer Science","cited_by":664,"is_retracted":false,"has_abstract":false,"ca_institutions":"Artificial Intelligence in Medicine (Canada)","funders":"National Institute of Biomedical Imaging and Bioengineering; Foundation for the National Institutes of Health; National Institute of Arthritis and Musculoskeletal and Skin Diseases; Agency for Healthcare Research and Quality; National Institutes of Health; National Heart, Lung, and Blood Institute","keywords":"Automatic summarization; Computer science; Natural language processing; Artificial intelligence; Information retrieval; Medicine","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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001697275,0.000147522,0.0002728208,0.0002362946,0.00003229342,0.00003330201,0.0006487123,0.0005139924,0.0001934108],"category_scores_gemma":[0.0007093954,0.0001063249,0.00005047096,0.0005773604,0.000048225,0.0002477739,0.0001936377,0.00115871,0.000009347729],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00006827145,"about_ca_system_score_gemma":0.0002167703,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001321044,"about_ca_topic_score_gemma":0.0005467939,"domain_scores_codex":[0.9976817,0.0000949334,0.0005220096,0.0005187021,0.0008674898,0.0003151027],"domain_scores_gemma":[0.9989724,0.0002614072,0.00003921461,0.0004784951,0.00005052609,0.0001979525],"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":[0.00002515987,0.0001780246,0.001939504,0.0001536113,0.00005250004,0.001802775,0.01992762,0.0003798933,0.0001234939,0.3332027,0.02083486,0.6213799],"study_design_scores_gemma":[0.000723307,0.00005914799,0.0003317877,0.0004507411,0.000006290835,0.00002806899,0.0002219752,0.9864346,0.00004463844,0.001320864,0.01025204,0.0001265042],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.05393868,0.01582626,0.8707203,0.04340168,0.004932643,0.0002876872,0.00000458886,0.0005362181,0.0103519],"genre_scores_gemma":[0.9914723,0.0001687015,0.002546114,0.004273975,0.001012152,0.000009508309,0.00002703611,0.00001471914,0.0004754621],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9860547,"threshold_uncertainty_score":0.5034082,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02209648554058223,"score_gpt":0.3445173154923743,"score_spread":0.3224208299517921,"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."}}