{"id":"W3132054471","doi":"10.48550/arxiv.2102.08607","title":"On the Convergence and Sample Efficiency of Variance-Reduced Policy Gradient Method","year":2021,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Reinforcement Learning in Robotics","field":"Computer Science","cited_by":18,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Alberta","funders":"","keywords":"Convexity; Truncation (statistics); Variance (accounting); Convergence (economics); Term (time); Mathematics; Variance reduction; Applied mathematics; Reinforcement learning; Mathematical optimization; Function (biology); Sample (material); Computer science; Statistics; Economics; Physics; Artificial intelligence; Monte Carlo method","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":[0.0005012422,0.0002529846,0.0003151007,0.0002189709,0.0001737559,0.000104311,0.001737751,0.000146307,0.00003743876],"category_scores_gemma":[0.0006085421,0.0002290288,0.0001444515,0.0008898096,0.0001889616,0.0001368685,0.0020676,0.0004926563,0.000007713011],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001347085,"about_ca_system_score_gemma":0.0003962013,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000561559,"about_ca_topic_score_gemma":0.000005414974,"domain_scores_codex":[0.9981287,0.000334392,0.0002342106,0.0008290109,0.0001608655,0.0003127932],"domain_scores_gemma":[0.9968296,0.001131019,0.0003643816,0.001394115,0.0001710406,0.0001098429],"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":[0.000004577769,0.00002291858,0.0001506658,0.0000320332,0.00003563373,0.00001510647,0.000462112,0.5904998,0.0000946938,0.4085757,0.00001967917,0.0000871774],"study_design_scores_gemma":[0.000204275,0.00009417254,0.0007046755,0.0001161039,0.00003612772,0.000003446136,0.0001265202,0.9772581,0.001230023,0.01989491,0.0000738236,0.0002577789],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1096047,0.0000228916,0.8883497,0.0002854029,0.0004163471,0.0002347296,0.000005119303,0.00005654306,0.001024654],"genre_scores_gemma":[0.9812689,0.0001566457,0.017935,0.000163244,0.00002840593,8.699747e-7,0.00000362151,0.00001038133,0.0004329611],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.8716642,"threshold_uncertainty_score":0.9339523,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.06947985645106997,"score_gpt":0.2222800602288609,"score_spread":0.152800203777791,"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."}}