{"id":"W1981932308","doi":"10.1109/icassp.2010.5495109","title":"I-smooth for improved minimum classification error training","year":2010,"lang":"en","type":"article","venue":"","topic":"Speech Recognition and Synthesis","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"National Research Council Canada","funders":"","keywords":"Hidden Markov model; TIMIT; Discriminative model; Computer science; Boundary (topology); Interpolation (computer graphics); Generalization; Speech recognition; Word error rate; Pattern recognition (psychology); Confusion; Artificial intelligence; Training (meteorology); Training set; Task (project management); Mathematics; 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":[0.0002414986,0.00006896067,0.00008117935,0.00006280468,0.00008515879,0.0001084139,0.0003467755,0.00006254394,0.000147491],"category_scores_gemma":[0.0001487467,0.00005818025,0.0000622041,0.0001057321,0.00002109884,0.0002257634,0.00002243259,0.00007148682,0.00005911663],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000006093321,"about_ca_system_score_gemma":0.00004836241,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000003319263,"about_ca_topic_score_gemma":0.00004234031,"domain_scores_codex":[0.9993843,0.00001067155,0.0001360778,0.0002362871,0.00006644262,0.0001661511],"domain_scores_gemma":[0.999366,0.0001537894,0.00004882969,0.0002827536,0.00007651989,0.00007206872],"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":[0.000002600778,0.00002213022,0.000007535948,0.000002096969,0.000003052552,1.447221e-7,0.0001948149,1.939495e-8,0.171023,0.03204182,0.00114386,0.7955589],"study_design_scores_gemma":[0.0007192378,0.00008376841,0.003276946,0.000005280347,0.000008442775,0.0000122426,0.0004098462,0.7728646,0.0578733,0.01313829,0.1512724,0.0003356177],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.01810567,0.000001763462,0.9465584,0.005754561,0.0007472102,0.0002546975,0.000003616377,0.0002773868,0.02829671],"genre_scores_gemma":[0.4417177,5.965888e-7,0.5555419,0.0008010613,0.00009625272,0.00006514302,0.000003427215,0.000006640736,0.001767315],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.7952232,"threshold_uncertainty_score":0.2372522,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.0834741312720563,"score_gpt":0.2978591173893709,"score_spread":0.2143849861173146,"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."}}