{"id":"W3172351783","doi":"10.18653/v1/2021.sigdial-1.18","title":"Improving Unsupervised Dialogue Topic Segmentation with Utterance-Pair Coherence Scoring","year":2021,"lang":"en","type":"preprint","venue":"","topic":"Topic Modeling","field":"Computer Science","cited_by":35,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of British Columbia","funders":"","keywords":"Computer science; Utterance; Coherence (philosophical gambling strategy); Artificial intelligence; Inference; Segmentation; Natural language processing; Exploit; Relevance (law); Speech recognition; Task (project management); Machine learning","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":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.0002042702,0.0002985332,0.0003220084,0.00009393645,0.0001081244,0.000735531,0.001074117,0.0001496626,0.00004259382],"category_scores_gemma":[0.00002010885,0.0002658984,0.00007958883,0.0001828206,0.00002403442,0.0005354434,0.001363697,0.0004728842,0.000008179884],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001371774,"about_ca_system_score_gemma":0.0003569772,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001208445,"about_ca_topic_score_gemma":0.0002624122,"domain_scores_codex":[0.9977354,0.00008744653,0.0003582926,0.001066818,0.000410597,0.0003413854],"domain_scores_gemma":[0.9983038,0.00005566258,0.0001669766,0.001220878,0.0001578092,0.00009487151],"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":[0.00001998,0.0001816297,0.04028827,0.001863076,0.0002127476,0.0003500849,0.01183534,0.08078658,0.02113311,0.006246616,0.00007434294,0.8370082],"study_design_scores_gemma":[0.001018141,0.00008804345,0.004586868,0.0009294975,0.00003740453,0.00002983468,0.0004041291,0.9560826,0.03453926,0.001127955,0.00004588902,0.001110386],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.2489982,0.0001306148,0.7484025,0.0003398698,0.0008294025,0.0003648609,0.00000106381,0.0002775774,0.0006559135],"genre_scores_gemma":[0.6068707,0.000009572018,0.3922164,0.0003333252,0.0001613191,0.00009739875,0.00001826932,0.00001333328,0.0002797758],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.875296,"threshold_uncertainty_score":0.9999793,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02639506892366116,"score_gpt":0.2366871093001384,"score_spread":0.2102920403764772,"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."}}