{"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003050691,0.001702592,0.0007517153,0.00141425,0.0007260411,0.001289517,0.002568517,0.001844539,0.003045583],"category_scores_gemma":[0.01408667,0.0004242909,0.0008198223,0.0007865705,0.001359373,0.004369046,0.004166329,0.002297516,0.001258944],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007952512,"about_ca_system_score_gemma":0.001119439,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00399889,"about_ca_topic_score_gemma":0.008544529,"domain_scores_codex":[0.9981126,0.0007978775,0.00009269069,0.0005743362,0.000297312,0.0001251837],"domain_scores_gemma":[0.9921811,0.005217609,0.0003784903,0.001447833,0.0005767373,0.000198346],"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":[0.002052036,0.0007388141,0.01466902,0.001742594,0.0003411685,0.0009658912,0.00316566,0.1310563,0.02720343,0.01077154,0.04462298,0.7626706],"study_design_scores_gemma":[0.0002187347,0.0005134633,0.002225341,0.00007450075,0.0001195321,0.0003752639,0.0009493792,0.9294496,0.01907319,0.03334948,0.01358694,0.00006450953],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.2943598,0.003279842,0.6520181,0.001757376,0.0003331092,0.0006339395,0.005272846,0.03603615,0.00630895],"genre_scores_gemma":[0.6997132,0.0004140603,0.2824313,0.0005114538,0.00009473303,0.0003399759,0.01135629,0.0006843264,0.004454535],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.00399889,"threshold_uncertainty_score":0.01613379,"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."}}