{"id":"W2981443541","doi":"10.2196/14782","title":"Efficient Reuse of Natural Language Processing Models for Phenotype-Mention Identification in Free-text Electronic Medical Records: A Phenotype Embedding Approach","year":2019,"lang":"en","type":"article","venue":"JMIR Medical Informatics","topic":"Machine Learning in Healthcare","field":"Computer Science","cited_by":16,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"King's College London; Medical Research Council; Engineering and Physical Sciences Research Council; National Institute for Health and Care Research; Department of Health and Social Care; South London and Maudsley NHS Foundation Trust","keywords":"Computer science; Retraining; Identification (biology); Artificial intelligence; Machine learning; Reuse; Natural language processing; Language model; Adaptation (eye); Embedding","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":[],"consensus_categories":[],"category_scores_codex":[0.002302581,0.0001680906,0.0003166629,0.0002522753,0.00006616794,0.00007488082,0.001743174,0.0002698023,0.00003997689],"category_scores_gemma":[0.001046218,0.0001490283,0.00007502458,0.0006122391,0.00006092968,0.000520387,0.0004403869,0.0008068449,0.00001749331],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002025013,"about_ca_system_score_gemma":0.0006358843,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00006014528,"about_ca_topic_score_gemma":0.00002076916,"domain_scores_codex":[0.9965441,0.0001120726,0.001077615,0.0002455025,0.001492906,0.0005277564],"domain_scores_gemma":[0.9981359,0.0002064438,0.0004815626,0.0008254283,0.0001712673,0.0001794037],"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":[0.000182708,0.0009990083,0.0006229298,0.01105654,0.00005907885,0.000005833218,0.1320097,0.04390191,0.0001029184,0.06579804,0.001717203,0.7435441],"study_design_scores_gemma":[0.0009055598,0.00009869931,0.0001439134,0.000413723,0.00000466748,0.0000103955,0.001067982,0.9956401,0.00002193379,0.001399797,0.0001399002,0.0001533448],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.3662721,0.0009389057,0.629181,0.001061689,0.0004713114,0.001118922,0.00000333006,0.0001487204,0.0008040286],"genre_scores_gemma":[0.9565853,0.00001565772,0.04269899,0.0003161236,0.00007947758,0.0001204247,0.00004596748,0.00001589719,0.0001221661],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9517382,"threshold_uncertainty_score":0.6077197,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.0110460746108385,"score_gpt":0.3162595550492472,"score_spread":0.3052134804384087,"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."}}