{"id":"W4385571912","doi":"10.18653/v1/2023.findings-acl.356","title":"LABO: Towards Learning Optimal Label Regularization via Bi-level Optimization","year":2023,"lang":"en","type":"article","venue":"","topic":"Machine Learning and Data Classification","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo; Université de Montréal; Huawei Technologies (Canada)","funders":"","keywords":"Regularization (linguistics); Computer science; Smoothing; Artificial intelligence; Normalization (sociology); Machine learning; Deep neural networks; Artificial neural network; Mathematical optimization; 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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0005858219,0.0001280221,0.0001117987,0.0002467477,0.0002688977,0.0002681361,0.0005109056,0.00008907728,0.00007638356],"category_scores_gemma":[0.0002606759,0.0001212865,0.00002836234,0.001558919,0.00002156616,0.0007274422,0.0002505517,0.000188948,0.000429064],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00003367159,"about_ca_system_score_gemma":0.00006158686,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00007209532,"about_ca_topic_score_gemma":0.00000271118,"domain_scores_codex":[0.998614,0.0001407925,0.0002249853,0.0004336517,0.0003327054,0.0002539261],"domain_scores_gemma":[0.9991798,0.00004123539,0.0001132566,0.0004485206,0.0001406564,0.00007653011],"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":[0.00000471465,0.00004945591,0.0005175022,0.00001666848,0.0000123637,0.000004481096,0.000389098,0.7165078,0.001895503,0.0424291,0.001850124,0.2363232],"study_design_scores_gemma":[0.0002402855,0.00004739234,0.007747804,0.000007664948,0.000003761671,0.000004176415,0.00002562827,0.9862752,0.0004013802,0.0002160551,0.004872572,0.0001580967],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.001526182,0.00001095991,0.988753,0.00342166,0.0002104756,0.00009993346,0.000002538078,0.001485702,0.004489558],"genre_scores_gemma":[0.1507036,0.00008662661,0.8129803,0.0003219337,0.0001818378,0.00003309236,0.0009206658,0.00003611099,0.03473582],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.2697674,"threshold_uncertainty_score":0.5514893,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03958658749458459,"score_gpt":0.2820888890739697,"score_spread":0.2425023015793851,"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."}}