{"id":"W6968829696","doi":"10.5281/zenodo.4265632","title":"MeDAL: Medical Abbreviation Disambiguation Dataset for Natural Language Understanding Pretraining","year":2020,"lang":"en","type":"article","venue":"Zenodo (CERN European Organization for Nuclear Research)","topic":"Topic Modeling","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"McGill University","funders":"","keywords":"Download; Upload; Natural language; Text messaging; Natural (archaeology)","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":true,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0008175331,0.00009899709,0.0001082201,0.0000850925,0.0009115031,0.000667182,0.001196429,0.00005372967,0.0008882207],"category_scores_gemma":[0.002184071,0.0001028194,0.00003457852,0.0003601517,0.00004421319,0.0005819698,0.0008148404,0.0002099853,0.0003502338],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001287185,"about_ca_system_score_gemma":0.000007689485,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000004982033,"about_ca_topic_score_gemma":3.850522e-7,"domain_scores_codex":[0.9983887,0.0001699337,0.0002288857,0.0004242143,0.0005294441,0.0002587886],"domain_scores_gemma":[0.9992368,0.00007039386,0.00009108466,0.0002885196,0.0001222764,0.0001909555],"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":[0.00007932499,0.00007721168,0.000005764977,0.0002692943,0.00007282072,0.00002556056,0.02021132,0.0003604854,0.003349323,0.1087934,0.1386316,0.7281239],"study_design_scores_gemma":[0.0008685924,0.0001542115,0.0000464129,0.00005863536,0.00001132343,0.00003450311,0.001048769,0.8034505,0.0002289644,0.0005166627,0.1933643,0.0002171006],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.001295486,0.00008257735,0.9893792,0.006208519,0.0001233511,0.0003425013,0.0002316297,0.0005619967,0.001774698],"genre_scores_gemma":[0.9875156,0.00001223525,0.005352284,0.0009300704,0.0002514142,7.857115e-8,0.005627281,0.0002861429,0.00002489587],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.9862201,"threshold_uncertainty_score":0.9725394,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1124420204035989,"score_gpt":0.2983701788779148,"score_spread":0.1859281584743159,"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."}}