{"id":"W4226288657","doi":"10.18653/v1/2022.naacl-main.57","title":"NeuroLogic A*esque Decoding: Constrained Text Generation with Lookahead Heuristics","year":2022,"lang":"en","type":"article","venue":"Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":46,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Naval Information Warfare Center Pacific; Natural Sciences and Engineering Research Council of Canada; Defense Advanced Research Projects Agency; Allen Institute for Artificial Intelligence","keywords":"Rowan; Heuristics; Computer science; Decoding methods; Linguistics; Artificial intelligence; History; Philosophy; Algorithm","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0004386619,0.0001994801,0.000341187,0.0001486824,0.0005530578,0.00006485177,0.002925536,0.00004970709,0.000003701893],"category_scores_gemma":[0.003124433,0.0001240519,0.0001702972,0.000715278,0.000593911,0.00006816188,0.00137619,0.0004104548,1.294998e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001439681,"about_ca_system_score_gemma":0.0001250312,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00005190252,"about_ca_topic_score_gemma":0.0000328029,"domain_scores_codex":[0.9981428,0.0000387296,0.0004989765,0.0003182747,0.0007791063,0.00022212],"domain_scores_gemma":[0.9952883,0.0002397473,0.002590804,0.0003537067,0.001511047,0.00001637872],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.00006029504,0.0001757926,0.07057234,0.000179801,0.0001848095,9.148938e-7,0.00203308,0.005969734,0.01024389,0.9041865,0.0004535252,0.005939314],"study_design_scores_gemma":[0.002631224,0.003177462,0.03309932,0.0003559022,0.0007147906,0.00007967972,0.00597338,0.2224946,0.1970711,0.5314043,0.001122227,0.001875939],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9752257,0.0002858792,0.01556945,0.004700649,0.0006194418,0.00165538,0.0003889902,0.0006387811,0.0009157595],"genre_scores_gemma":[0.9596493,0.000002504616,0.0399944,0.0001281167,0.0000387522,0.00005575221,0.00000772993,0.00001494726,0.0001084602],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.3727822,"threshold_uncertainty_score":0.5436422,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.0168325293327092,"score_gpt":0.245775511115093,"score_spread":0.2289429817823838,"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."}}