{"id":"W4221161302","doi":"10.48550/arxiv.2201.01666","title":"Sample Efficient Deep Reinforcement Learning via Uncertainty Estimation","year":2022,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Reinforcement Learning in Robotics","field":"Computer Science","cited_by":15,"is_retracted":false,"has_abstract":true,"ca_institutions":"Université de Montréal","funders":"","keywords":"Reinforcement learning; Weighting; Computer science; Heteroscedasticity; Variance (accounting); Sample (material); Noise (video); Bayesian probability; Probabilistic logic; Process (computing); Artificial intelligence; Machine learning; Mathematical optimization; Mathematics","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":[0.0006200192,0.0004392822,0.0003834816,0.0004231344,0.0006787614,0.0002374894,0.002391712,0.0002118332,0.0004274992],"category_scores_gemma":[0.0002188092,0.0005503385,0.0002658777,0.0008378148,0.00009318838,0.0002532934,0.005162781,0.001417767,0.0001297986],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001204467,"about_ca_system_score_gemma":0.0002378595,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0004347671,"about_ca_topic_score_gemma":0.000005261106,"domain_scores_codex":[0.9970753,0.0003025169,0.000415656,0.001229217,0.0003805862,0.0005967085],"domain_scores_gemma":[0.9971452,0.0003498039,0.0006533485,0.001485583,0.0001684843,0.0001975378],"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":[0.00001729469,0.00003051847,0.0004995906,0.00006382036,0.00006928829,0.00006082804,0.0005304369,0.9268074,0.000004334695,0.07108887,0.00003380406,0.0007938391],"study_design_scores_gemma":[0.0004092853,0.0001471148,0.00015364,0.00004213769,0.00005911256,0.000002963352,0.0001156043,0.9949781,0.00001457473,0.002326388,0.001195607,0.000555424],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.008406102,0.00002177683,0.9868518,0.00006161174,0.000950768,0.0005213412,0.000001598622,0.0005663367,0.002618711],"genre_scores_gemma":[0.9868923,0.00004277968,0.01067375,0.00009195345,0.00004256145,0.000005167472,0.0001727523,0.00002988421,0.002048882],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.9784862,"threshold_uncertainty_score":0.9996948,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04515281141028903,"score_gpt":0.2008328326548355,"score_spread":0.1556800212445464,"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."}}