{"id":"W3089548822","doi":"","title":"Stable Policy Optimization via Off-Policy Divergence Regularization.","year":2020,"lang":"en","type":"article","venue":"Uncertainty in Artificial Intelligence","topic":"Adversarial Robustness in Machine Learning","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"McGill University; Université de Montréal","funders":"","keywords":"Reinforcement learning; Benchmark (surveying); Regularization (linguistics); Computer science; Divergence (linguistics); Stability (learning theory); Trust region; Mathematical optimization; Adversarial system; Optimization problem; Range (aeronautics); Term (time); Artificial intelligence; Machine learning; Algorithm; 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":[0.0004730046,0.0002948607,0.0003131407,0.0004295629,0.0002912872,0.0002549687,0.001676383,0.0001513272,0.000163343],"category_scores_gemma":[0.002710195,0.0003194869,0.00008551493,0.004242264,0.0001650309,0.0009387205,0.0006679203,0.0004372454,0.0001837183],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003452065,"about_ca_system_score_gemma":0.0005280908,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00155228,"about_ca_topic_score_gemma":0.00008636462,"domain_scores_codex":[0.9969914,0.000263544,0.0007536755,0.0008144684,0.0005448907,0.0006320011],"domain_scores_gemma":[0.9984587,0.0002065648,0.0002525444,0.0005923859,0.0002356986,0.0002541206],"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":[0.0000218584,0.00003764155,0.0001586724,0.00001003325,0.000005327982,0.00001163282,0.001867349,0.703388,0.0002591815,0.2293497,0.0000294335,0.06486109],"study_design_scores_gemma":[0.0000474762,0.00006964336,0.00003446764,0.00002378914,0.000003550987,0.000003211796,0.0001663659,0.9574835,0.002097913,0.03905619,0.0006961949,0.000317669],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0004019767,0.00006047312,0.9822041,0.01500586,0.0003166337,0.0003300025,0.000004398284,0.0002999569,0.001376624],"genre_scores_gemma":[0.898549,0.0001035883,0.09778571,0.002491688,0.0008326095,0.00002918226,0.0000231595,0.00003335406,0.0001517012],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.898147,"threshold_uncertainty_score":0.9999257,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03451898926511433,"score_gpt":0.2957511071451971,"score_spread":0.2612321178800828,"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."}}