{"id":"W7037509623","doi":"","title":"Efficient local linearity regularization to overcome catastrophic overfitting","year":2024,"lang":"en","type":"article","venue":"Infoscience (Ecole Polytechnique Fédérale de Lausanne)","topic":"Plant Diversity and Evolution","field":"Agricultural and Biological Sciences","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Army Research Office; Agencia Estatal de Investigación; Institute for Catastrophic Loss Reduction; Hasler Stiftung; Comunidad de Madrid; Schweizerischer Nationalfonds zur Förderung der Wissenschaftlichen Forschung; National Science Foundation","keywords":"Overfitting; Regularization (linguistics); Adversarial system; Linearity; Term (time); Curvature; Function (biology); Training set","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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.000680026,0.0001679391,0.000150966,0.00007816709,0.0003999811,0.0002367543,0.0004462956,0.0001750612,0.0001237076],"category_scores_gemma":[0.00009392347,0.00009337051,0.00008822588,0.001280881,0.0001360494,0.0002410925,0.0002288589,0.0002576948,0.0001817218],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000221795,"about_ca_system_score_gemma":0.00005811042,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0008906766,"about_ca_topic_score_gemma":0.0002764651,"domain_scores_codex":[0.9983145,0.0000634376,0.0002316115,0.0004660558,0.0003666253,0.0005577762],"domain_scores_gemma":[0.9994348,0.0001179655,0.00004954503,0.00009925626,0.00006188876,0.0002365332],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00008919394,0.0002356223,0.008036044,0.00008216996,0.00001536116,0.0001315854,0.0006870322,0.0273743,0.8425121,0.01033535,0.00866838,0.1018329],"study_design_scores_gemma":[0.0002568743,0.001191078,0.2092081,0.001049537,0.00005933359,0.0002538999,0.0006063735,0.5782201,0.1083411,0.00293531,0.09625088,0.001627448],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9600193,0.0001004615,0.03567689,0.001777445,0.0003818012,0.0003894328,0.0002344979,0.0006257477,0.0007944411],"genre_scores_gemma":[0.9962813,0.00001252003,0.00222892,0.000610569,0.0001990351,0.00002448791,0.00006399253,0.000001642064,0.0005774806],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.734171,"threshold_uncertainty_score":0.3807539,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01084930365096565,"score_gpt":0.2171355408073563,"score_spread":0.2062862371563907,"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."}}