{"id":"W3025133396","doi":"10.65109/uazl2918","title":"META-Learning State-based Eligibility Traces for More Sample-Efficient Policy Evaluation","year":2020,"lang":"en","type":"preprint","venue":"","topic":"Reinforcement Learning in Robotics","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":true,"ca_institutions":"Mila - Quebec Artificial Intelligence Institute; McGill University","funders":"","keywords":"Reinforcement learning; Computer science; Machine learning; Robustness (evolution); Artificial intelligence; Temporal difference learning; Sample (material); Q-learning; Bellman equation; 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":[0.003145732,0.0005371131,0.000792585,0.0003289531,0.0002554684,0.0007130731,0.001806059,0.000228762,0.0001202836],"category_scores_gemma":[0.003762926,0.0004653006,0.0006721194,0.0005118816,0.00009038871,0.0001442547,0.001204591,0.0008497976,0.00003483949],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004914316,"about_ca_system_score_gemma":0.001911756,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0003928177,"about_ca_topic_score_gemma":0.000005033645,"domain_scores_codex":[0.9949905,0.0004955934,0.0008926852,0.001450139,0.001591208,0.0005799084],"domain_scores_gemma":[0.9955316,0.0013289,0.0007353912,0.001378085,0.0007997162,0.0002262757],"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":[0.00001661451,0.00003975971,0.00008924865,0.0002287962,0.0003450014,5.787188e-7,0.0009601591,0.9893518,0.0000325368,0.004577799,0.0001598158,0.0041979],"study_design_scores_gemma":[0.0004881587,0.0001567394,0.000355815,0.00002441262,0.0005706465,3.339206e-7,0.00003571911,0.9912645,0.0008799984,0.004304953,0.001466043,0.0004526471],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.001816832,0.0001170286,0.9853511,0.008455601,0.0004274688,0.002596713,0.00002942242,0.0007008961,0.0005049022],"genre_scores_gemma":[0.6565921,0.000009234598,0.3410138,0.0009727523,0.0001588611,0.000641136,0.0002732346,0.00004501928,0.0002938516],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.6547752,"threshold_uncertainty_score":0.9997799,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1683899976315094,"score_gpt":0.3973292214127686,"score_spread":0.2289392237812592,"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."}}