{"id":"W3207783082","doi":"","title":"BI-RADS BERT & Using Section Tokenization to Understand Radiology Reports.","year":2021,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Topic Modeling","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"Health Sciences Centre; Sunnybrook Health Science Centre; University of Toronto","funders":"","keywords":"Lexical analysis; Computer science; Artificial intelligence; Lexicon; Natural language processing; Preprocessor; Breast imaging; Section (typography); Radiology; Breast cancer; Mammography; Medicine; Cancer","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":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.0002726407,0.0002296888,0.0002969549,0.0003290581,0.0001675744,0.0001607816,0.0006006855,0.0003297581,0.00001783915],"category_scores_gemma":[0.00005371059,0.0003000943,0.0001259637,0.0007341751,0.00003427208,0.0003841844,0.001432965,0.0003785415,0.000004048749],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006766359,"about_ca_system_score_gemma":0.0003452341,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0005057376,"about_ca_topic_score_gemma":0.000115481,"domain_scores_codex":[0.9978845,0.0001587558,0.0002643663,0.001305592,0.00009511363,0.0002916599],"domain_scores_gemma":[0.9981354,0.00003270149,0.0002402928,0.001237235,0.0002049376,0.0001494295],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.000005974861,0.0000230175,0.002440045,0.00003363195,0.00004885672,0.0009126876,0.0009698269,0.98405,0.0003347646,0.010874,0.00006277131,0.0002444494],"study_design_scores_gemma":[0.0001543186,0.0000276564,0.0004747164,0.00009395221,0.00005152545,0.0001929144,0.0003665612,0.9896086,0.0002386674,0.008260828,0.0001386311,0.0003915798],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.4220063,0.00003581683,0.575702,0.00006891716,0.001526624,0.0001367706,6.205483e-7,0.0001166457,0.000406294],"genre_scores_gemma":[0.9846725,0.00004199067,0.01452395,0.0001587174,0.0001868523,2.720111e-7,0.00001062091,0.00001555863,0.0003894698],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.5626662,"threshold_uncertainty_score":0.9999451,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1166955285620192,"score_gpt":0.2011895285846554,"score_spread":0.0844940000226362,"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."}}