{"id":"W4414789198","doi":"10.48550/arxiv.2509.20786","title":"LiLAW: Lightweight Learnable Adaptive Weighting to Learn Sample Difficulty & Improve Noisy Training","year":2025,"lang":"en","type":"preprint","venue":"ArXiv.org","topic":"Speech Recognition and Synthesis","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Natural Sciences and Engineering Research Council of Canada; Canadian Institutes of Health Research; University of Toronto","keywords":"Robustness (evolution); Weighting; Hyperparameter; Artificial neural network; Training set; Stochastic gradient descent; Gradient descent; Generalization; Noisy data","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":["metaepi_narrow","insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.0006957324,0.0006839639,0.0008726706,0.0004633974,0.0005043987,0.0004841501,0.002177657,0.00051858,0.0002951636],"category_scores_gemma":[0.001075268,0.0006475705,0.0004660319,0.0007026898,0.0000689856,0.0003599364,0.002709144,0.001349479,0.001248296],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000275554,"about_ca_system_score_gemma":0.0004851776,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000621777,"about_ca_topic_score_gemma":0.0002944902,"domain_scores_codex":[0.9955286,0.0002647506,0.0007713953,0.001873005,0.0005910933,0.0009711519],"domain_scores_gemma":[0.9963182,0.0009081304,0.0003952982,0.001556966,0.0004041796,0.0004172251],"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":[0.0001888306,0.0009685503,0.01940897,0.0006185839,0.001283624,0.00020683,0.03007631,0.00107089,0.01704476,0.01630169,0.01592471,0.8969063],"study_design_scores_gemma":[0.002548645,0.0006299721,0.04053039,0.003958978,0.0004567817,0.00006084032,0.002389534,0.03640204,0.09237946,0.01739107,0.7958468,0.00740555],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.3817925,0.0002588513,0.566507,0.004700642,0.004120525,0.001331358,0.000211653,0.001428379,0.03964908],"genre_scores_gemma":[0.514066,0.0001397195,0.4410612,0.003481394,0.001448094,0.0005322859,0.0001324555,0.0001107438,0.03902808],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.8895007,"threshold_uncertainty_score":0.9995975,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.09525507824844402,"score_gpt":0.2889259135805544,"score_spread":0.1936708353321104,"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."}}