{"id":"W4388444930","doi":"10.48550/arxiv.2311.01479","title":"Detecting Out-of-Distribution Through the Lens of Neural Collapse","year":2023,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Anomaly Detection Techniques and Applications","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"National Cancer Institute; Canadian Institute for Advanced Research","keywords":"Generalizability theory; Leverage (statistics); Computer science; Detector; Artificial intelligence; Artificial neural network; Feature vector; Range (aeronautics); Feature (linguistics); Exploit; Machine learning; Engineering; Computer security; Mathematics; Telecommunications","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003215069,0.001083423,0.001288532,0.002229588,0.0009564671,0.00218997,0.002753261,0.001740688,0.001699151],"category_scores_gemma":[0.02196265,0.000732206,0.0009903867,0.00147635,0.002688827,0.003342621,0.004699576,0.00407493,0.0004930372],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002004045,"about_ca_system_score_gemma":0.001547759,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.008771584,"about_ca_topic_score_gemma":0.00966729,"domain_scores_codex":[0.9983165,0.0004003469,0.00007360997,0.0005098872,0.000498164,0.0002013783],"domain_scores_gemma":[0.9940918,0.002747432,0.0008583121,0.001076233,0.0009085099,0.0003176888],"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":[0.0004932933,0.0001990139,0.0404161,0.0002168222,0.0002823214,0.0009373704,0.0007083644,0.4781586,0.01360842,0.1067295,0.01281733,0.3454329],"study_design_scores_gemma":[0.000007152493,0.00001888848,0.001427538,0.00001603213,0.00001011276,0.0001194981,0.00004373731,0.9571165,0.00229464,0.03777064,0.001157517,0.0000177491],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.05112741,0.000335244,0.9446323,0.0007008738,0.00005783917,0.00004251055,0.0002006306,0.00142607,0.001477081],"genre_scores_gemma":[0.8298466,0.0003170227,0.16473,0.000651202,0.0001447478,0.0001144244,0.0007825119,0.0004503335,0.002963338],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.008771584,"threshold_uncertainty_score":0.01744109,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.133265300991219,"score_gpt":0.2266639093023236,"score_spread":0.09339860831110464,"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."}}