{"id":"W2015633636","doi":"10.1109/icassp.2014.6854823","title":"I-vector-based speaker adaptation of deep neural networks for French broadcast audio transcription","year":2014,"lang":"en","type":"article","venue":"","topic":"Speech Recognition and Synthesis","field":"Computer Science","cited_by":119,"is_retracted":false,"has_abstract":true,"ca_institutions":"Computer Research Institute of Montréal","funders":"","keywords":"Speech recognition; Computer science; Speaker diarisation; Word error rate; Transcription (linguistics); Feature vector; Hidden Markov model; Artificial neural network; Artificial intelligence; Speaker recognition; Acoustic model; Vector quantization; Pattern recognition (psychology); Speech processing","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":[0.0006859531,0.0006102549,0.0002904318,0.0002654243,0.0001817251,0.0002882918,0.0005072827,0.0004317035,0.003215957],"category_scores_gemma":[0.001431727,0.0002271391,0.0003660003,0.0002492857,0.0001931133,0.0004497302,0.0005068204,0.0009486963,0.001410047],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004378377,"about_ca_system_score_gemma":0.000353314,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00644605,"about_ca_topic_score_gemma":0.00905784,"domain_scores_codex":[0.9996284,0.0001391691,0.00001892623,0.00009540556,0.00007044574,0.00004757124],"domain_scores_gemma":[0.9996321,0.0001582268,0.00002098258,0.0000477787,0.0001226398,0.00001829473],"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":[0.000586083,0.0001354138,0.001106579,0.0001143654,0.0001099996,0.0001340279,0.00020678,0.1451586,0.1525994,0.001075069,0.004083545,0.6946902],"study_design_scores_gemma":[0.0000214259,0.0001374066,0.002034944,0.00001377332,0.00004035027,0.00009806074,0.00004934699,0.9198104,0.07425089,0.0008726233,0.002638055,0.00003269261],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1585557,0.0009357904,0.8246348,0.0002273546,0.0002302033,0.0001033799,0.0005158968,0.009869349,0.004927555],"genre_scores_gemma":[0.7556191,0.0003755632,0.2311812,0.0001723339,0.00007878579,0.0001710985,0.001761668,0.0005784988,0.0100618],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.00644605,"threshold_uncertainty_score":0.01281708,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02772913931601576,"score_gpt":0.2293340230087177,"score_spread":0.2016048836927019,"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."}}