{"id":"W2587210085","doi":"10.1109/slt.2016.7846241","title":"Batch-normalized joint training for DNN-based distant speech recognition","year":2016,"lang":"en","type":"article","venue":"","topic":"Speech and Audio Processing","field":"Computer Science","cited_by":30,"is_retracted":false,"has_abstract":true,"ca_institutions":"Université de Montréal","funders":"","keywords":"Speech recognition; Computer science; Robustness (evolution); Speech technology; Acoustic model; Voice activity detection; Training set; Speech processing; Speaker recognition; Joint (building); Speech enhancement; Training (meteorology); Artificial intelligence; Engineering","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":[],"consensus_categories":[],"category_scores_codex":[0.0004012378,0.0001218341,0.0001533468,0.00007922511,0.0001305377,0.0001483059,0.0002859583,0.00004604553,0.000117488],"category_scores_gemma":[0.0001709919,0.00007635938,0.00009251654,0.0001649766,0.00002968683,0.0005576152,0.00003771123,0.00003739859,0.00006208938],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00004961626,"about_ca_system_score_gemma":0.0001412166,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000004704385,"about_ca_topic_score_gemma":0.000009761608,"domain_scores_codex":[0.9988661,0.00002302622,0.000235199,0.0003416476,0.0001822246,0.0003517812],"domain_scores_gemma":[0.9992716,0.0001727537,0.00009263513,0.0002490559,0.0001106818,0.0001032401],"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":[0.00001858923,0.00002065252,0.00002191744,0.0000153146,0.00000472581,0.000004844685,0.00008980097,5.539005e-7,0.0835053,0.0003036445,0.0007308539,0.9152838],"study_design_scores_gemma":[0.001509945,0.0001048862,0.00008199345,0.0001279808,0.000005240606,0.00001026145,0.00003577888,0.002686809,0.9707381,0.01961999,0.004853677,0.0002253242],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.02206728,0.00001939243,0.9697324,0.003999905,0.0002545523,0.0001783272,0.000006729721,0.0002992768,0.003442131],"genre_scores_gemma":[0.3992216,0.000003054707,0.5989761,0.001036653,0.0001072286,0.00004001497,0.000004790784,0.00001176017,0.0005987707],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.9150585,"threshold_uncertainty_score":0.3113845,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.08675578359859087,"score_gpt":0.2701628841973234,"score_spread":0.1834071005987326,"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."}}