{"id":"W2091014931","doi":"10.1145/1177055.1177057","title":"Confidence estimation for NLP applications","year":2006,"lang":"en","type":"article","venue":"ACM Transactions on Speech and Language Processing","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":30,"is_retracted":false,"has_abstract":true,"ca_institutions":"National Research Council Canada; Université de Montréal","funders":"","keywords":"Computer science; Machine translation; Artificial intelligence; Natural language processing; Estimation; Confidence interval; Natural language; Machine learning; Statistics; Mathematics","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":[0.0001610031,0.0001446771,0.0001243054,0.000151958,0.0003893338,0.0003232013,0.0004842912,0.00008075377,0.000006637809],"category_scores_gemma":[0.00003117941,0.0001327451,0.00004062188,0.0003512303,0.00005292577,0.0006400461,0.00001222096,0.0001480929,0.000004479958],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00003207166,"about_ca_system_score_gemma":0.00005220659,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00008095615,"about_ca_topic_score_gemma":0.00002750119,"domain_scores_codex":[0.999042,0.0000141775,0.0001922903,0.0003698835,0.0001655964,0.0002160398],"domain_scores_gemma":[0.9992441,0.0001160421,0.00008975353,0.0003964272,0.0001034865,0.00005018498],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.000006328406,0.00004845936,0.000002585587,0.00009134108,0.000002894358,0.000003011647,0.0002768657,0.00004334034,0.003744363,0.005423162,0.00003135734,0.9903263],"study_design_scores_gemma":[0.0009902022,0.0002115519,0.00005840443,0.0003307959,0.00007103309,0.0002478876,0.0004074427,0.1613738,0.4302529,0.4030198,0.002135058,0.0009010148],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.0004443829,0.004913214,0.9926556,0.0007532103,0.00002777958,0.0003959314,0.000008731331,0.000648941,0.0001521549],"genre_scores_gemma":[0.4311251,0.000006785866,0.5683165,0.0001623534,0.00003215464,0.0001524924,0.000006679174,0.00000884474,0.0001890756],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.9894253,"threshold_uncertainty_score":0.5413188,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.00987436474930131,"score_gpt":0.2842979056426198,"score_spread":0.2744235408933184,"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."}}