{"id":"W2904453761","doi":"10.48550/arxiv.1812.02900","title":"Off-Policy Deep Reinforcement Learning without Exploration","year":2018,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Reinforcement Learning in Robotics","field":"Computer Science","cited_by":280,"is_retracted":false,"has_abstract":true,"ca_institutions":"McGill University","funders":"","keywords":"Reinforcement learning; Computer science; Extrapolation; Artificial intelligence; Reinforcement; Class (philosophy); Uncorrelated; Space (punctuation); Action (physics); Control (management); Machine learning; Mathematics; Engineering","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":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.0004376033,0.0004687819,0.0003984674,0.0005934703,0.0003966733,0.0003678288,0.00227604,0.0003665528,0.00006599739],"category_scores_gemma":[0.0001954335,0.0005605649,0.0002121429,0.0008241462,0.0001709339,0.001131763,0.003414909,0.000973916,0.0006518754],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006114824,"about_ca_system_score_gemma":0.0004254793,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001160187,"about_ca_topic_score_gemma":0.00001215629,"domain_scores_codex":[0.9973186,0.0002327417,0.000392993,0.001185167,0.0002704489,0.0006000365],"domain_scores_gemma":[0.9970809,0.00007720346,0.0006688882,0.001615139,0.0003282726,0.0002296291],"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":[0.00001702318,0.00001533724,0.001272256,0.00004695106,0.00007843606,0.00003759088,0.0007431455,0.9362316,0.00001148601,0.06067394,0.0001389214,0.0007333177],"study_design_scores_gemma":[0.000437006,0.0001884926,0.0001752539,0.0001105666,0.00004652889,0.000002832795,0.00009162052,0.9903445,0.000127813,0.004330835,0.003579814,0.0005647122],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.009142648,0.00002239434,0.9742266,0.0001548998,0.0007084008,0.0004202038,3.367011e-7,0.000602242,0.01472229],"genre_scores_gemma":[0.9831737,0.0003554383,0.003483821,0.000160451,0.000360577,0.00000278623,0.0000442914,0.00003730545,0.01238158],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9740311,"threshold_uncertainty_score":0.9996846,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.0781399192839883,"score_gpt":0.2175502889724059,"score_spread":0.1394103696884176,"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."}}