{"id":"W3176581245","doi":"10.18653/v1/2021.sigdial-1.49","title":"A Brief Study on the Effects of Training Generative Dialogue Models with a Semantic loss","year":2021,"lang":"en","type":"preprint","venue":"","topic":"Topic Modeling","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"McGill University","funders":"Compute Canada; Natural Sciences and Engineering Research Council of Canada; Canadian Institute for Advanced Research","keywords":"Utterance; Computer science; Artificial intelligence; Language model; Security token; Natural language processing; Generative grammar; Training set; Set (abstract data type); Semantic similarity; Generative model; Task (project management); Context (archaeology); Principle of maximum entropy; Machine learning","routes":{"ca_aff":true,"ca_fund":true,"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":[0.008114505,0.002170611,0.001555063,0.0006679775,0.0007979129,0.002051846,0.001468004,0.002162036,0.005076698],"category_scores_gemma":[0.04585599,0.000812167,0.001321472,0.0008044898,0.001376864,0.003754177,0.00272381,0.005142096,0.000825885],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001250001,"about_ca_system_score_gemma":0.0006881308,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004799732,"about_ca_topic_score_gemma":0.005025733,"domain_scores_codex":[0.9949656,0.003447362,0.000247131,0.0006148331,0.0004552619,0.0002697477],"domain_scores_gemma":[0.9519336,0.04315516,0.000694448,0.002634361,0.00121551,0.0003669344],"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":[0.002185681,0.001037509,0.006326106,0.00110694,0.00052557,0.000278353,0.0004863686,0.6741168,0.02128336,0.01685614,0.005569738,0.2702273],"study_design_scores_gemma":[0.00008478059,0.0007701694,0.001712406,0.00007385511,0.00008001781,0.0001327999,0.0001226792,0.9755443,0.00952759,0.009625912,0.002290848,0.00003461582],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.2076923,0.01381125,0.7551509,0.003780209,0.0006586903,0.0004831939,0.0007144965,0.002749652,0.01495938],"genre_scores_gemma":[0.7665795,0.002779595,0.2209341,0.0009934996,0.0004245617,0.0004252323,0.001034815,0.0008679563,0.005960796],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.008114505,"threshold_uncertainty_score":0.04291409,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.05970064700788191,"score_gpt":0.26324544836168,"score_spread":0.203544801353798,"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."}}