{"id":"W4416756303","doi":"10.1109/taslpro.2025.3633090","title":"Towards Effective and Efficient Non-Autoregressive Decoders for Conformer and LLM-Based ASR Using Block-Based Attention Mask","year":2025,"lang":"","type":"article","venue":"IEEE Transactions on Audio Speech and Language Processing","topic":"Speech Recognition and Synthesis","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"National Research Council Canada","funders":"Youth Innovation Promotion Association of the Chinese Academy of Sciences","keywords":"Decoding methods; Speedup; Inference; Autoregressive model; Connectionism; Language model; Speech processing; Speech coding","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":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.0006513566,0.0005409284,0.0005739778,0.0007673726,0.001231228,0.0008128631,0.000184529,0.0003018088,0.00002994979],"category_scores_gemma":[0.00006278708,0.0005191138,0.0001944677,0.0005827691,0.0003950466,0.0003328602,0.000007703806,0.0003725462,0.00000238132],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001666139,"about_ca_system_score_gemma":0.0005312701,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00009877957,"about_ca_topic_score_gemma":0.00006749757,"domain_scores_codex":[0.9973784,0.000149268,0.0005024947,0.001018601,0.0003575674,0.0005936351],"domain_scores_gemma":[0.9983894,0.0004793016,0.0002843298,0.0002793852,0.0003034347,0.0002642153],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0003501363,0.0003376339,0.00003457094,0.001179504,0.0001176128,0.00002883731,0.001456556,0.001789681,0.01045686,0.000004643441,0.000008165743,0.9842358],"study_design_scores_gemma":[0.003112641,0.0002826204,0.0003847981,0.001871961,0.0004016734,0.00004248219,0.0009986004,0.7660984,0.2262759,0.00002753607,0.00003373365,0.0004696524],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.2448769,0.001546197,0.7511192,0.0005632878,0.0004031619,0.001167385,0.00006715002,0.0000890731,0.0001676559],"genre_scores_gemma":[0.9371591,0.00004818601,0.06148018,0.0009203591,0.00004262165,0.0001377113,0.000004333778,0.00003398172,0.0001735229],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9837661,"threshold_uncertainty_score":0.9997261,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01155468110483668,"score_gpt":0.279686951037758,"score_spread":0.2681322699329213,"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."}}