{"id":"W4405600491","doi":"10.1109/jiot.2024.3520426","title":"Building Privacy-Preserving Medical Text Models With a Pretrained Transformer","year":2024,"lang":"en","type":"article","venue":"IEEE Internet of Things Journal","topic":"Digital and Cyber Forensics","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":true,"ca_institutions":"Brandon University","funders":"National Natural Science Foundation of China","keywords":"Computer science; Transformer; Information privacy; Artificial intelligence; Computer security; Electrical engineering; Engineering","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00108699,0.0007980459,0.0006318535,0.000711488,0.0003141284,0.0009803982,0.0009930615,0.000918089,0.00202789],"category_scores_gemma":[0.00398848,0.0004536488,0.001346807,0.000470357,0.000802001,0.001919976,0.001159851,0.001578892,0.001989985],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007462743,"about_ca_system_score_gemma":0.0009441237,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002351578,"about_ca_topic_score_gemma":0.00292709,"domain_scores_codex":[0.9994917,0.0001115717,0.00003649124,0.000173541,0.0001220077,0.00006467434],"domain_scores_gemma":[0.9987889,0.0006481663,0.0001311228,0.0001807124,0.0002026056,0.00004846807],"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":[0.0007120112,0.000240766,0.003909375,0.000215395,0.0001585255,0.0006233143,0.0002721222,0.515985,0.03673675,0.02383394,0.008537517,0.4087754],"study_design_scores_gemma":[0.00001297831,0.00004524704,0.0001553265,0.000006811031,0.00001829715,0.00009506169,0.00001776807,0.9858841,0.00419463,0.008715069,0.0008473536,0.000007319033],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.02831843,0.000247407,0.9668035,0.0005258478,0.00005781939,0.00009491693,0.0003975476,0.002060201,0.00149445],"genre_scores_gemma":[0.8177189,0.0005940439,0.1701993,0.0008173457,0.0001663918,0.0002454437,0.002260497,0.0003331261,0.007664908],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.002351578,"threshold_uncertainty_score":0.006784022,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01486054228097387,"score_gpt":0.2466106563182092,"score_spread":0.2317501140372354,"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."}}