{"id":"W2152382718","doi":"10.3115/1626431.1626478","title":"Stabilizing minimum error rate training","year":2009,"lang":"en","type":"article","venue":"","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":46,"is_retracted":false,"has_abstract":true,"ca_institutions":"National Research Council Canada","funders":"","keywords":"Computer science; Stability (learning theory); Word error rate; Variation (astronomy); Training set; Feature (linguistics); Test data; Component (thermodynamics); Machine translation; Translation (biology); Artificial intelligence; Test (biology); Machine learning; Training (meteorology); Algorithm; Pattern recognition (psychology)","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":[0.007005925,0.001239106,0.001574055,0.0009489725,0.0009660375,0.001107691,0.002384224,0.001770847,0.002728036],"category_scores_gemma":[0.04097695,0.0006672493,0.0007532617,0.0009818812,0.00123386,0.001768999,0.001980739,0.00247132,0.002219573],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008007658,"about_ca_system_score_gemma":0.001467596,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002051772,"about_ca_topic_score_gemma":0.002477227,"domain_scores_codex":[0.9952176,0.002099339,0.0002696778,0.00103598,0.001077423,0.0002999773],"domain_scores_gemma":[0.9843457,0.008664551,0.0007954118,0.002805216,0.003203086,0.0001859254],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0007561571,0.0002740864,0.004474236,0.0002480109,0.0001534292,0.0001676577,0.000367738,0.5660141,0.01873278,0.01651222,0.006182818,0.3861167],"study_design_scores_gemma":[0.00003355581,0.0001284971,0.0005898775,0.000022489,0.00001777744,0.00007968998,0.00002742883,0.9784722,0.01280849,0.006679291,0.001124375,0.00001621476],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.02953872,0.0002268035,0.9653105,0.0001650561,0.0000720649,0.0001055024,0.00009633348,0.002564058,0.001921023],"genre_scores_gemma":[0.5118444,0.000169277,0.4814749,0.000301666,0.00009256122,0.0006143418,0.0008324355,0.0009702002,0.003700207],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.007005925,"threshold_uncertainty_score":0.03705132,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03676422810836918,"score_gpt":0.3020566031008958,"score_spread":0.2652923749925267,"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."}}