{"id":"W2969609970","doi":"10.48550/arxiv.1809.04729","title":"A Less Biased Evaluation of Out-of-distribution Sample Detectors","year":2018,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Anomaly Detection Techniques and Applications","field":"Computer Science","cited_by":21,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of British Columbia","funders":"","keywords":"Outlier; Computer science; Sample (material); Artificial intelligence; Set (abstract data type); Machine learning; Population; Scheme (mathematics); Data mining; Mathematics","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":[0.02267618,0.002488664,0.00192262,0.003567242,0.001364295,0.003416915,0.00310977,0.004145805,0.002974933],"category_scores_gemma":[0.07276268,0.0004642115,0.00130871,0.002235575,0.002168669,0.004498816,0.003604969,0.002821609,0.002033433],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001633842,"about_ca_system_score_gemma":0.001371919,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003589973,"about_ca_topic_score_gemma":0.004886957,"domain_scores_codex":[0.9717889,0.01352902,0.001864424,0.003955075,0.00784359,0.00101903],"domain_scores_gemma":[0.9465215,0.02607887,0.002962469,0.01386058,0.009101044,0.00147548],"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":[0.008084238,0.003360167,0.08624493,0.003690003,0.002627283,0.0009032607,0.0007432975,0.1707556,0.03606186,0.01917077,0.08271823,0.5856404],"study_design_scores_gemma":[0.0009010326,0.005459404,0.03772108,0.0007383371,0.0007068983,0.003465535,0.001150042,0.7914667,0.08726365,0.02352966,0.04730358,0.0002940535],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.5759568,0.01633813,0.353699,0.003796313,0.003064165,0.001431714,0.009144226,0.01125334,0.02531639],"genre_scores_gemma":[0.8011522,0.001188313,0.1703445,0.001552639,0.0004461118,0.0003753562,0.01887253,0.001081742,0.00498659],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.02267618,"threshold_uncertainty_score":0.1199245,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1871613230210542,"score_gpt":0.2461036610132556,"score_spread":0.05894233799220136,"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."}}