{"id":"W3021380283","doi":"10.1093/jamia/ocaa038","title":"Using word embeddings to improve the privacy of clinical notes","year":2020,"lang":"en","type":"article","venue":"Journal of the American Medical Informatics Association","topic":"Machine Learning in Healthcare","field":"Computer Science","cited_by":21,"is_retracted":false,"has_abstract":true,"ca_institutions":"St. Michael's Hospital; Institute for Clinical Evaluative Sciences; Vector Institute; University of Toronto","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Computer science; Readability; Security token; Word embedding; Recall; Artificial intelligence; Word (group theory); Natural language processing; Precision and recall; Machine learning; Information retrieval; Embedding; Data science; Computer security","routes":{"ca_aff":true,"ca_fund":true,"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":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.004643752,0.00008535977,0.0004250493,0.00004647798,0.0001010629,0.00006934304,0.001800046,0.00006249236,0.000006183497],"category_scores_gemma":[0.03520809,0.00004627229,0.0002186375,0.0007447855,0.00007807361,0.0002731733,0.0005378475,0.0009832561,0.00000531136],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001412714,"about_ca_system_score_gemma":0.0005301774,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00005989253,"about_ca_topic_score_gemma":0.000001593121,"domain_scores_codex":[0.9960238,0.0004164527,0.001655843,0.00006131823,0.00164556,0.0001970799],"domain_scores_gemma":[0.9917423,0.001904006,0.005352446,0.0002728818,0.0004360018,0.000292387],"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":[0.00009353897,0.0001104603,0.3672973,0.0001526725,0.0003067775,0.000004667685,0.04158835,0.007079941,0.0001885157,0.001267266,0.008315947,0.5735946],"study_design_scores_gemma":[0.00043494,0.0005520732,0.06909765,0.000149718,0.00004080033,0.0000169499,0.0006818466,0.9241002,0.0001498681,0.0003470223,0.004310984,0.0001179583],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.6918274,0.000009252762,0.2121012,0.09523495,0.0006318624,0.0001248099,0.000001328962,0.00001223649,0.00005703488],"genre_scores_gemma":[0.9112493,0.00001978453,0.06896678,0.01929333,0.000459464,6.269602e-7,1.206075e-7,0.000005617542,0.000004975638],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.9170203,"threshold_uncertainty_score":0.9729187,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.05535004671577025,"score_gpt":0.4073642148405306,"score_spread":0.3520141681247604,"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."}}