{"id":"W3153046263","doi":"10.18653/v1/2021.emnlp-main.168","title":"Neural Path Hunter: Reducing Hallucination in Dialogue Systems via Path Grounding","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":75,"is_retracted":false,"has_abstract":true,"ca_institutions":"Canadian Institute for Advanced Research; University of Alberta","funders":"Alberta Machine Intelligence Institute; Compute Canada; Natural Sciences and Engineering Research Council of Canada; Canadian Institute for Advanced Research","keywords":"Computer science; Path (computing); Security token; Artificial neural network; Focus (optics); Artificial intelligence; Suite; Graph; Deep neural networks; Machine learning; Theoretical computer science; History","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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001909389,0.0002865607,0.0005212372,0.0002752946,0.000130574,0.0004588098,0.001145163,0.0001759784,0.000005660553],"category_scores_gemma":[0.001880284,0.0002197513,0.0001028781,0.001572588,0.00006375878,0.0009438494,0.0006116756,0.0009892657,8.689562e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002579718,"about_ca_system_score_gemma":0.0001744536,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001094201,"about_ca_topic_score_gemma":0.00001218186,"domain_scores_codex":[0.9970465,0.0002790141,0.0007640274,0.0008446011,0.0005632493,0.0005026054],"domain_scores_gemma":[0.9984741,0.0002918032,0.0004428079,0.0003007223,0.0004164347,0.0000741248],"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":[0.00002654382,0.0001455218,0.005825401,0.0006139281,0.00001112551,0.00004245437,0.01292047,0.0005030039,0.1266517,0.00657112,0.000007041096,0.8466817],"study_design_scores_gemma":[0.0003732201,0.00003342812,0.003220503,0.001702081,0.00000958943,0.00005473842,0.001976476,0.9759703,0.01361846,0.002732117,0.000008144219,0.0003009192],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.62996,0.008041443,0.3524436,0.003554706,0.002097578,0.0006535067,0.000002554086,0.0001463469,0.003100298],"genre_scores_gemma":[0.7652932,0.00001491378,0.234324,0.0001558667,0.000101092,0.00002459328,0.000001254603,0.00001477093,0.00007029976],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.9754673,"threshold_uncertainty_score":0.8961197,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.0556578091105373,"score_gpt":0.3775607902916954,"score_spread":0.3219029811811581,"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."}}