{"id":"W3197275000","doi":"10.18653/v1/2021.emnlp-main.181","title":"Unsupervised Conversation Disentanglement through Co-Training","year":2021,"lang":"en","type":"article","venue":"Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing","topic":"Topic Modeling","field":"Computer Science","cited_by":20,"is_retracted":false,"has_abstract":true,"ca_institutions":"Queen's University","funders":"","keywords":"Conversation; Computer science; Classifier (UML); Session (web analytics); Artificial intelligence; Machine learning; Reinforcement learning; Training set; Speech recognition; World Wide Web","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":[0.0008683087,0.0002234103,0.0003663537,0.00008999485,0.0001257892,0.0002469372,0.000992402,0.0001086068,0.00004645347],"category_scores_gemma":[0.001101523,0.0001652533,0.0001147455,0.0009596451,0.00009094131,0.0006800954,0.0004149994,0.0005677343,0.000002051733],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00012883,"about_ca_system_score_gemma":0.0002579921,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001142511,"about_ca_topic_score_gemma":0.000003088054,"domain_scores_codex":[0.997815,0.000128023,0.0005064066,0.0006318286,0.0005403863,0.0003783323],"domain_scores_gemma":[0.9988037,0.0002112725,0.0002803793,0.0002625583,0.0003817378,0.0000603353],"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":[0.00002758117,0.0001235977,0.003793812,0.0002957644,0.0000249542,0.00001454093,0.05404985,0.00003052326,0.1551296,0.02097037,0.00004038207,0.7654991],"study_design_scores_gemma":[0.00121523,0.00006481392,0.002415319,0.001437686,0.00003535653,0.00004029653,0.02620299,0.4426259,0.4933661,0.03169112,0.000265792,0.0006394196],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.43537,0.006207095,0.500186,0.02867191,0.001560003,0.0009410989,0.000007579742,0.0002779801,0.02677826],"genre_scores_gemma":[0.607336,0.00001357089,0.3918849,0.0006067611,0.0000510395,0.00001179131,0.000001250987,0.000008463164,0.00008621411],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.7648596,"threshold_uncertainty_score":0.6738834,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1138268762368847,"score_gpt":0.4306470469350119,"score_spread":0.3168201706981271,"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."}}