{"id":"W4404238744","doi":"10.1109/ijcb62174.2024.10744521","title":"On the influence of regularization techniques on label noise robustness: Self-supervised speaker verification as a use case","year":2024,"lang":"en","type":"article","venue":"","topic":"Speech Recognition and Synthesis","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"Computer Research Institute of Montréal","funders":"Natural Sciences and Engineering Research Council of Canada; Government of Canada","keywords":"Robustness (evolution); Speaker verification; Computer science; Speech recognition; Regularization (linguistics); Speaker recognition; Artificial intelligence; Pattern recognition (psychology); Noise measurement; Noise reduction","routes":{"ca_aff":true,"ca_fund":true,"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":[0.01471009,0.001340657,0.001083715,0.0009962184,0.000985602,0.001167149,0.001168692,0.00225892,0.0007072446],"category_scores_gemma":[0.04217341,0.000441694,0.0008631204,0.0006501566,0.002291145,0.002362318,0.001782143,0.002106916,0.0003291141],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008682109,"about_ca_system_score_gemma":0.0006221733,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002946537,"about_ca_topic_score_gemma":0.00298955,"domain_scores_codex":[0.9925042,0.004717797,0.0002523717,0.0009884767,0.001276613,0.0002604983],"domain_scores_gemma":[0.9519629,0.03648758,0.002328138,0.005534579,0.003256498,0.0004303216],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.002332928,0.000613537,0.009739274,0.0005257686,0.0006883938,0.0003191525,0.0007356257,0.6232191,0.08097927,0.009244311,0.002224679,0.269378],"study_design_scores_gemma":[0.00003420538,0.0005124393,0.002659584,0.00004331199,0.00008989744,0.0002563286,0.0001099864,0.9409608,0.05128669,0.003305231,0.00069116,0.00005043597],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.3802366,0.002547021,0.6111616,0.001263481,0.0001013452,0.0001778652,0.0001024652,0.001197472,0.003212195],"genre_scores_gemma":[0.8474125,0.0005041641,0.1499697,0.0002499769,0.00007026501,0.00007581413,0.0001602457,0.0002563443,0.001300972],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01471009,"threshold_uncertainty_score":0.07779527,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02842436391590777,"score_gpt":0.2556727641753869,"score_spread":0.2272484002594791,"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."}}