{"id":"W4400530140","doi":"10.1145/3626772.3657903","title":"Unbiased Validation of Technology-Assisted Review for eDiscovery","year":2024,"lang":"en","type":"article","venue":"","topic":"Machine Learning and Data Classification","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo","funders":"","keywords":"Computer science","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":[],"consensus_categories":[],"category_scores_codex":[0.000320224,0.00004805719,0.00008979298,0.0001245269,0.00002451556,0.00004733707,0.0002841471,0.00003323601,0.00001590716],"category_scores_gemma":[0.0001858731,0.00003688597,0.00004261499,0.0005657189,0.0000151842,0.0002463462,0.00004608243,0.00005152266,0.00002438109],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00001160267,"about_ca_system_score_gemma":0.00005106087,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000004742985,"about_ca_topic_score_gemma":6.460273e-7,"domain_scores_codex":[0.9994601,0.00001832927,0.0001675168,0.0001980176,0.00008837049,0.00006765873],"domain_scores_gemma":[0.999443,0.00008535486,0.00004640563,0.0003582113,0.00005363142,0.00001338509],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.000001410504,0.00002865471,0.00009673076,0.001526061,0.00001408044,5.843845e-7,0.00001536018,0.000006132803,0.005092919,0.442244,0.03107157,0.5199025],"study_design_scores_gemma":[0.0003496321,0.0002441907,0.00168507,0.00369093,0.000080512,0.00002447342,0.00001999681,0.1818081,0.04728806,0.01658702,0.7478758,0.0003462095],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0003434679,0.004167201,0.9771046,0.01546109,0.0001988832,0.0002034905,0.000007221865,0.0003674022,0.002146662],"genre_scores_gemma":[0.7396568,0.002137555,0.252843,0.0008845243,0.0001026621,0.0001801988,0.0003106038,0.00001838275,0.003866281],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.7393133,"threshold_uncertainty_score":0.1504166,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02645759788254705,"score_gpt":0.3177730236927362,"score_spread":0.2913154258101892,"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."}}