{"id":"W6957621777","doi":"10.60692/5r3jg-m5e18","title":"Imperfect also Deserves Reward: Multi-Level and Sequential Reward Modeling for Better Dialog Management","year":2021,"lang":"en","type":"article","venue":"Greater South Information System","topic":"Topic Modeling","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Université de Montréal","funders":"","keywords":"Imperfect; Dialog box; Selection (genetic algorithm); Language model; Association (psychology); Subject (documents)","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":[0.003944865,0.001328217,0.0019008,0.0009347497,0.001265448,0.002503953,0.002564008,0.001994962,0.008843505],"category_scores_gemma":[0.01347585,0.0008568039,0.001252286,0.0007172627,0.0009571686,0.004270288,0.0025238,0.003423153,0.001159519],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001911142,"about_ca_system_score_gemma":0.002708549,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01511104,"about_ca_topic_score_gemma":0.02211963,"domain_scores_codex":[0.9985908,0.0006441832,0.00006283429,0.0003275033,0.0001478164,0.0002267617],"domain_scores_gemma":[0.9942577,0.004044933,0.000357667,0.0003468898,0.0004937782,0.0004989821],"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":[0.001080602,0.000356221,0.006549927,0.0002131077,0.0002202439,0.0002684245,0.0008339391,0.766485,0.00170433,0.09856153,0.009925647,0.1138009],"study_design_scores_gemma":[0.00001465757,0.00002119901,0.0001944167,0.000007572637,0.00001856396,0.00001031484,0.00001636078,0.9811126,0.00009519466,0.0180448,0.0004556302,0.000008678594],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.06339035,0.001474677,0.9225486,0.003101496,0.0003694282,0.000147796,0.000787402,0.001534598,0.006645704],"genre_scores_gemma":[0.919255,0.0003970747,0.07324536,0.0003117552,0.0002015917,0.000193658,0.0004916941,0.0002070948,0.005696874],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01511104,"threshold_uncertainty_score":0.03004616,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.09650262403361097,"score_gpt":0.2500986736970738,"score_spread":0.1535960496634629,"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."}}