{"id":"W2016084804","doi":"10.1109/icassp.2013.6639015","title":"Incoherent training of deep neural networks to de-correlate bottleneck features for speech recognition","year":2013,"lang":"en","type":"article","venue":"","topic":"Speech Recognition and Synthesis","field":"Computer Science","cited_by":47,"is_retracted":false,"has_abstract":true,"ca_institutions":"York University","funders":"","keywords":"Hidden Markov model; Computer science; Speech recognition; Artificial neural network; Artificial intelligence; Bottleneck; Pattern recognition (psychology); Feature extraction; Context (archaeology); Feature (linguistics)","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":[0.0002455144,0.0001260695,0.000181156,0.0001168062,0.00006803926,0.0001177335,0.0003507972,0.00009404073,0.0003846154],"category_scores_gemma":[0.0001265547,0.0001087627,0.0001113706,0.0002319825,0.00001663132,0.0003242476,0.00006928174,0.00009522746,0.00006957931],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00002736403,"about_ca_system_score_gemma":0.00001567827,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00007526171,"about_ca_topic_score_gemma":0.00006911137,"domain_scores_codex":[0.9989665,0.00005170612,0.0002496476,0.0002675327,0.0001420739,0.0003226107],"domain_scores_gemma":[0.9991097,0.0002703092,0.00007902209,0.0002148538,0.0001700699,0.0001560199],"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":[0.000007678837,0.00002701044,0.00006592985,0.000006383499,0.00001223322,0.000001139159,0.0003224565,0.0001574194,0.0004708701,0.0001867467,0.001781198,0.9969609],"study_design_scores_gemma":[0.0005445429,0.0002632562,0.009304726,0.00004635204,0.00001867213,0.00006243333,0.0003578008,0.9602211,0.01845729,0.0100733,0.0002932189,0.0003572827],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1033913,0.00003013329,0.8890082,0.00116676,0.0003820551,0.0005792096,0.000002393824,0.0001414289,0.005298468],"genre_scores_gemma":[0.5989331,0.000004278674,0.3992915,0.001283409,0.00008644573,0.0001299395,0.000005461041,0.00001096698,0.0002548946],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.9966037,"threshold_uncertainty_score":0.4435214,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03280178765113027,"score_gpt":0.2504833423333723,"score_spread":0.217681554682242,"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."}}