{"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":"codex-gemma-dda1882f352a","candidate_categories":["insufficient_payload"],"consensus_categories":["insufficient_payload"],"category_scores_codex":[0.0008257673,0.0002302281,0.0002602323,0.0004075308,0.0005767041,0.0002027966,0.001071214,0.0001032331,0.04768885],"category_scores_gemma":[0.00009876734,0.0002380516,0.0001657016,0.001124529,0.00003833547,0.0004379661,0.001144058,0.0003412359,0.004700941],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003093278,"about_ca_system_score_gemma":0.000113431,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001786288,"about_ca_topic_score_gemma":0.0001560084,"domain_scores_codex":[0.9971775,0.0002710661,0.000394498,0.0007525483,0.0009084296,0.000495972],"domain_scores_gemma":[0.9986572,0.0002038585,0.00008330339,0.0006546543,0.00009643149,0.0003045559],"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":[0.00004065877,0.0001428831,0.0003354847,0.000002487097,0.00001678473,0.00006064741,0.0001801433,0.000002890324,0.004893021,0.0001717684,0.04614702,0.9480062],"study_design_scores_gemma":[0.002562726,0.0006777386,0.001334825,0.00003214773,0.0000480015,0.0009836691,0.000909698,0.005888413,0.4602509,0.00491491,0.5208666,0.001530382],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.7603413,0.0000296753,0.01296467,0.01001704,0.00302128,0.001146681,0.0002155037,0.001170846,0.211093],"genre_scores_gemma":[0.5150953,0.00004056908,0.4170086,0.01651481,0.001545808,0.0006657736,0.0003348653,0.0001331578,0.04866109],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.9464758,"threshold_uncertainty_score":0.996074,"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."}}