{"id":"W2125226720","doi":"10.1109/asru.2009.5373248","title":"Discriminative training of n-gram language models for speech recognition via linear programming","year":2009,"lang":"en","type":"article","venue":"","topic":"Speech Recognition and Synthesis","field":"Computer Science","cited_by":5,"is_retracted":false,"has_abstract":true,"ca_institutions":"York University","funders":"","keywords":"Discriminative model; Computer science; Speech recognition; Word error rate; Viterbi algorithm; Artificial intelligence; n-gram; Vocabulary; Language model; Metric (unit); Pattern recognition (psychology); Viterbi decoder; Hidden Markov model; Decoding methods; Algorithm","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":[0.0003087879,0.0001119842,0.0001817699,0.0001289149,0.00006062908,0.00004711184,0.0002582095,0.00005653825,0.00002837955],"category_scores_gemma":[0.00009145481,0.0000965685,0.0001169046,0.0002306555,0.00002388544,0.0005271047,0.00002172449,0.00006104093,0.00001045454],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00001647263,"about_ca_system_score_gemma":0.00002499285,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0000166981,"about_ca_topic_score_gemma":0.00001126791,"domain_scores_codex":[0.9990616,0.00003469011,0.0002387444,0.0002601654,0.0001786849,0.0002261133],"domain_scores_gemma":[0.999355,0.0001401089,0.00009657542,0.0001823888,0.0001604207,0.00006552939],"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":[0.000007729227,0.00008890814,4.996714e-7,0.000009672585,0.000008102214,0.000003359368,0.002662947,0.000003039822,0.001644257,0.002097066,0.0000296251,0.9934448],"study_design_scores_gemma":[0.0007338928,0.0005739242,0.00005113503,0.00009455771,0.00002978832,0.00003668672,0.003194805,0.5947965,0.3113891,0.08847375,0.0002548942,0.0003709788],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.01250209,0.00001690491,0.9773389,0.0004464578,0.00006243589,0.0003846209,0.000007601104,0.0001896937,0.009051357],"genre_scores_gemma":[0.3479708,0.00000257658,0.6516185,0.0001900072,0.00003761293,0.00002185458,0.00001233202,0.00000497057,0.0001413912],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.9930738,"threshold_uncertainty_score":0.3937949,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.0839183369757153,"score_gpt":0.3085977370139907,"score_spread":0.2246794000382754,"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."}}