{"id":"W4385573061","doi":"10.18653/v1/2022.emnlp-industry.29","title":"SpeechNet: Weakly Supervised, End-to-End Speech Recognition at Industrial Scale","year":2022,"lang":"en","type":"article","venue":"","topic":"Speech Recognition and Synthesis","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo","funders":"","keywords":"End-to-end principle; Computer science; Speech recognition; Scale (ratio); Natural language processing; Artificial intelligence; Cartography; Geography","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":[0.001860851,0.002276466,0.00150214,0.001601849,0.0006613661,0.001418625,0.002045909,0.001458468,0.01221382],"category_scores_gemma":[0.003742822,0.0008706747,0.0005774312,0.001099898,0.0005570132,0.002721231,0.002511653,0.001574374,0.01959542],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005205235,"about_ca_system_score_gemma":0.001262064,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006027457,"about_ca_topic_score_gemma":0.01157033,"domain_scores_codex":[0.9985119,0.0003916858,0.00008601062,0.0004693455,0.0004195608,0.0001214854],"domain_scores_gemma":[0.9984171,0.0006035736,0.00006291162,0.0003578832,0.0004399787,0.0001184729],"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":[0.002510628,0.000902488,0.002976096,0.0005585771,0.0003826117,0.0004452816,0.0002382568,0.02305412,0.04875241,0.003035151,0.320278,0.5968664],"study_design_scores_gemma":[0.0004226601,0.0006601155,0.006611898,0.00006893153,0.0001161332,0.0003493048,0.0002812078,0.8488619,0.07083806,0.01065721,0.06098819,0.0001443368],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.05269546,0.001798925,0.6174735,0.0008723352,0.001182103,0.0006737034,0.03841577,0.2754617,0.01142653],"genre_scores_gemma":[0.2743265,0.0009609177,0.4975027,0.0007296113,0.000419163,0.002171035,0.1778587,0.007367766,0.03866371],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01221382,"threshold_uncertainty_score":0.04085934,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.05790377424567948,"score_gpt":0.2421033126564522,"score_spread":0.1841995384107727,"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."}}