{"id":"W4392736665","doi":"10.1093/jamia/ocae312","title":"A dataset and benchmark for hospital course summarization with adapted large language models","year":2024,"lang":"en","type":"article","venue":"Journal of the American Medical Informatics Association","topic":"Machine Learning in Healthcare","field":"Computer Science","cited_by":28,"is_retracted":false,"has_abstract":true,"ca_institutions":"Artificial Intelligence in Medicine (Canada)","funders":"Stanford Cardiovascular Institute, School of Medicine, Stanford University; National Institute of Biomedical Imaging and Bioengineering; National Institute of Arthritis and Musculoskeletal and Skin Diseases; Stanford Institute for Human-Centered Artificial Intelligence, Stanford University; National Institute of Nursing Research; National Heart, Lung, and Blood Institute; National Institutes of Health","keywords":"Computer science; Context (archaeology); Automatic summarization; Benchmark (surveying); Artificial intelligence; Benchmarking; Health care; Machine learning; Natural language processing","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":[0.001730135,0.0000755666,0.0001804394,0.00006725007,0.00009284715,0.0001625681,0.0004201511,0.00004365681,0.000003194719],"category_scores_gemma":[0.0007973912,0.00004517655,0.0000438225,0.0003405226,0.00003312494,0.0006564363,0.0001080918,0.0003735123,8.46926e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001389823,"about_ca_system_score_gemma":0.0003323282,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00002962015,"about_ca_topic_score_gemma":0.00001441913,"domain_scores_codex":[0.9983974,0.00009102935,0.0004150798,0.00005660283,0.0008691392,0.0001707689],"domain_scores_gemma":[0.9982178,0.0004502927,0.0008914988,0.0001537849,0.0001701134,0.000116553],"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":[0.0001815695,0.0005823238,0.06510899,0.001534479,0.001575002,0.00009028109,0.1100476,0.0109227,0.00002428854,0.05708333,0.3321122,0.4207372],"study_design_scores_gemma":[0.0002962569,0.000311249,0.002446239,0.0001653887,0.00003662028,0.00002930296,0.0007181113,0.9912221,0.000003459532,0.0005263741,0.004172999,0.00007190638],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1222046,0.0001994704,0.8564122,0.02019283,0.0004675344,0.000237308,0.0001896571,0.00003184626,0.00006460679],"genre_scores_gemma":[0.9674237,0.0000908588,0.03000834,0.002151584,0.0001824522,0.000005418142,0.00009710772,0.000009584795,0.00003095697],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9802994,"threshold_uncertainty_score":0.1842246,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.005396857601458518,"score_gpt":0.2874210560819696,"score_spread":0.2820241984805111,"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."}}