{"id":"W4386815360","doi":"10.14428/esann/2023.es2023-181","title":"Multi-Fidelity Reinforcement Learning with Control Variates","year":2023,"lang":"en","type":"article","venue":"","topic":"Reinforcement Learning in Robotics","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"Microsoft (Canada)","funders":"","keywords":"Reinforcement learning; Estimator; Fidelity; Control variates; Computer science; Variance (accounting); Variance reduction; High fidelity; Monte Carlo method; Mathematical optimization; Bellman equation; Artificial intelligence; Sample (material); Function (biology); Machine learning; Mathematics; Statistics; 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":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.0006536671,0.0001732441,0.0001820731,0.0001431179,0.0002365133,0.0002365648,0.000666639,0.00005169464,0.00007709967],"category_scores_gemma":[0.000146182,0.0001315638,0.00004855378,0.0006914989,0.00003902377,0.0004464422,0.000249838,0.0002464906,0.0009033906],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00004901749,"about_ca_system_score_gemma":0.00006565083,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00006956869,"about_ca_topic_score_gemma":0.000004306017,"domain_scores_codex":[0.9983176,0.00007799958,0.000298008,0.0003634421,0.000465325,0.0004776058],"domain_scores_gemma":[0.9989127,0.0001982984,0.0001283404,0.0005196753,0.000129968,0.0001110479],"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":[0.000006494822,0.000006631789,0.003854536,0.000008890083,0.00003603528,0.0000141858,0.0002732453,0.9847144,0.000283849,0.009918606,0.000324511,0.0005585857],"study_design_scores_gemma":[0.001093778,0.0002075654,0.006982801,0.0000152295,0.000007471459,0.000003933282,0.00005849668,0.9881801,0.0001797011,0.00003536839,0.003035,0.0002005678],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0008153567,0.000005212707,0.9910876,0.0005620082,0.0001485731,0.0002390371,1.006986e-7,0.001395832,0.005746242],"genre_scores_gemma":[0.9121383,0.00001034752,0.0712885,0.0004468862,0.0000298272,0.00002511773,0.000007024358,0.00001486548,0.01603915],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.9197991,"threshold_uncertainty_score":0.9998745,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02119730371790796,"score_gpt":0.2544386487722293,"score_spread":0.2332413450543214,"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."}}