{"id":"W3209976096","doi":"10.1109/icassp43922.2022.9746832","title":"Pseudo-Labeling for Massively Multilingual Speech Recognition","year":2022,"lang":"en","type":"article","venue":"ICASSP 2022 - 2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)","topic":"Speech Recognition and Synthesis","field":"Computer Science","cited_by":15,"is_retracted":false,"has_abstract":true,"ca_institutions":"McGill University","funders":"","keywords":"Computer science; Natural language processing; Speech recognition; Artificial intelligence; Language model","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002543429,0.001600304,0.0009094463,0.0008320355,0.001049961,0.001358069,0.00269747,0.001327209,0.005476344],"category_scores_gemma":[0.007083763,0.0008219437,0.001025791,0.0008283131,0.001584103,0.003338422,0.003628733,0.002798351,0.004789139],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009118405,"about_ca_system_score_gemma":0.001578786,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003252031,"about_ca_topic_score_gemma":0.007479724,"domain_scores_codex":[0.9976495,0.001070474,0.0001312486,0.0006624317,0.0003626235,0.0001236918],"domain_scores_gemma":[0.9953238,0.001909535,0.0002277775,0.001568552,0.0007973193,0.0001729974],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.0009545847,0.0003796,0.002529243,0.0004398086,0.0001654989,0.0004336806,0.0007038491,0.2118917,0.04302362,0.03096363,0.02600736,0.6825075],"study_design_scores_gemma":[0.00004248732,0.00009649493,0.0003219914,0.00002563735,0.00001721873,0.0001250896,0.00006921318,0.9341763,0.01600966,0.03901193,0.01005378,0.00005021143],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.008749971,0.0001792372,0.9800514,0.0001841143,0.0001108279,0.00009856301,0.0003983834,0.008730046,0.00149743],"genre_scores_gemma":[0.2145405,0.0001528566,0.7733205,0.000563492,0.0001311869,0.0006326283,0.004188768,0.00162,0.004850006],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.005476344,"threshold_uncertainty_score":0.01832026,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.08598030702419207,"score_gpt":0.3185981659794001,"score_spread":0.232617858955208,"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."}}