{"id":"W4308364487","doi":"10.1109/tnnls.2022.3217189","title":"Monotonic Quantile Network for Worst-Case Offline Reinforcement Learning","year":2022,"lang":"en","type":"article","venue":"IEEE Transactions on Neural Networks and Learning Systems","topic":"Reinforcement Learning in Robotics","field":"Computer Science","cited_by":16,"is_retracted":false,"has_abstract":true,"ca_institutions":"Vector Institute; University of Toronto","funders":"","keywords":"Quantile; Reinforcement learning; Quantile regression; CVAR; Computer science; Quantile function; Monotonic function; Function (biology); Bellman equation; Q-learning; Mathematical optimization; Offline learning; Artificial intelligence; Expected shortfall; Machine learning; Econometrics; Online learning; Mathematics; Cumulative distribution function; Risk management; Statistics; Economics; Probability density function","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":[0.002260657,0.001113972,0.001229191,0.0004136624,0.0003526922,0.0008757563,0.00176233,0.00119864,0.003655047],"category_scores_gemma":[0.007893527,0.0004709121,0.0004114766,0.000390052,0.001084833,0.001437314,0.001226799,0.002315091,0.0004889794],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001283477,"about_ca_system_score_gemma":0.001365267,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003931135,"about_ca_topic_score_gemma":0.003312282,"domain_scores_codex":[0.999067,0.0002990481,0.00004349989,0.0002638023,0.0001909283,0.0001358377],"domain_scores_gemma":[0.9976019,0.001561341,0.000222422,0.000166372,0.0003372266,0.0001106876],"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":[0.0001159964,0.00006603905,0.0009065384,0.00006631655,0.00003262962,0.00006241293,0.00005274611,0.9389884,0.001246271,0.01291264,0.0009889768,0.04456093],"study_design_scores_gemma":[0.000004571853,0.00001553586,0.00004670014,0.000003600716,0.000003300087,0.000006474357,0.000002396813,0.9959848,0.000189527,0.003621702,0.0001187972,0.000002546717],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01553385,0.0003144636,0.9813447,0.0002579725,0.00003640833,0.00004879785,0.00005191513,0.0004864097,0.001925448],"genre_scores_gemma":[0.9159251,0.0002833413,0.07920324,0.0003496523,0.00004993927,0.0002440013,0.0001679275,0.0001278709,0.00364898],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.003931135,"threshold_uncertainty_score":0.01222736,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02027412986570837,"score_gpt":0.2429601647000125,"score_spread":0.2226860348343041,"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."}}