{"id":"W4399284119","doi":"10.1016/j.neucom.2024.127963","title":"Multi-fidelity reinforcement learning with control variates","year":2024,"lang":"en","type":"article","venue":"Neurocomputing","topic":"Reinforcement Learning in Robotics","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":false,"ca_institutions":"Microsoft (Canada)","funders":"Argonne National Laboratory; Office of Science; Laboratory Computing Resource Center; Advanced Scientific Computing Research; U.S. Department of Energy","keywords":"Reinforcement learning; Variance reduction; Estimator; Fidelity; Control variates; Computer science; Variance (accounting); High fidelity; Mathematical optimization; Reduction (mathematics); Function (biology); Artificial intelligence; Machine learning; Monte Carlo method; Mathematics; Statistics","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":[],"consensus_categories":[],"category_scores_codex":[0.0005802858,0.0002539216,0.0002191817,0.0001563978,0.0003154669,0.0008363657,0.0006927305,0.00005663038,0.00001379997],"category_scores_gemma":[0.0001128429,0.0002122469,0.00007941309,0.0005310476,0.00003887324,0.0005006167,0.0003142136,0.0006517294,0.0001475185],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0000667252,"about_ca_system_score_gemma":0.00009816005,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00002246827,"about_ca_topic_score_gemma":6.627552e-7,"domain_scores_codex":[0.9978154,0.0001368426,0.0004130596,0.0006336079,0.0004675237,0.0005335707],"domain_scores_gemma":[0.9988007,0.0004337291,0.0001273685,0.000424693,0.00009872337,0.0001147746],"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":[0.000004346655,0.000008398336,0.000922157,0.00005559672,0.00003695905,0.00008780636,0.0004597208,0.9865257,0.0008208893,0.004756207,0.00005467629,0.006267606],"study_design_scores_gemma":[0.0005189651,0.0002296128,0.001475692,0.0001365288,0.0000156403,0.00004801731,0.00001444892,0.9912652,0.0001816537,0.00002212609,0.005849449,0.0002426458],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.002440629,0.0000870123,0.9931993,0.0003704693,0.0005148958,0.0002874252,1.090342e-7,0.001420085,0.001680097],"genre_scores_gemma":[0.9203309,0.000004956421,0.07835925,0.0004737664,0.0001346188,0.000009492223,0.000002128871,0.00002984405,0.0006550598],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.9178903,"threshold_uncertainty_score":0.8655179,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01498700950666781,"score_gpt":0.2492363708707679,"score_spread":0.2342493613641001,"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."}}