{"id":"W3034806614","doi":"","title":"ConQUR: Mitigating Delusional Bias in Deep Q-Learning","year":2020,"lang":"en","type":"article","venue":"arXiv (Cornell University)","topic":"Domain Adaptation and Few-Shot Learning","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"Google (Canada); University of Alberta","funders":"","keywords":"Computer science; Class (philosophy); Variety (cybernetics); Scheme (mathematics); Delusion; Q-learning; Artificial intelligence; Simple (philosophy); Value (mathematics); Machine learning; Reinforcement learning; Mathematics; Psychology","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":[0.01011615,0.001146535,0.002006129,0.0006254471,0.0008624763,0.001351813,0.003113257,0.002128114,0.004014569],"category_scores_gemma":[0.03199128,0.0008215131,0.0005828876,0.0007050843,0.002662977,0.003657701,0.004103687,0.00422293,0.0006889937],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001424335,"about_ca_system_score_gemma":0.002659469,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00348725,"about_ca_topic_score_gemma":0.004097189,"domain_scores_codex":[0.9965703,0.001940001,0.0001399528,0.0004851143,0.0006308514,0.0002337726],"domain_scores_gemma":[0.9845989,0.01103622,0.0008647517,0.001881614,0.001180816,0.000437776],"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":[0.0006288012,0.0003942104,0.003625749,0.0003007582,0.0001521409,0.0001996649,0.0004423556,0.6702828,0.003731368,0.07837159,0.007152955,0.2347176],"study_design_scores_gemma":[0.00003859388,0.00006987802,0.0001071638,0.00001700175,0.000008061583,0.00002405399,0.00001186841,0.9742474,0.0008396914,0.02405564,0.0005720142,0.000008633784],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.01895057,0.000412198,0.9780044,0.0005131779,0.00005592038,0.00007913883,0.00004058076,0.0006181618,0.001325805],"genre_scores_gemma":[0.7136065,0.0003231246,0.2803752,0.001036893,0.0001239076,0.0003076734,0.0001951115,0.0002847532,0.003746837],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01011615,"threshold_uncertainty_score":0.05349994,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1088076188527771,"score_gpt":0.1888258779277586,"score_spread":0.08001825907498143,"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."}}