{"id":"W3072315125","doi":"10.1073/pnas.1907370117","title":"Fast reinforcement learning with generalized policy updates","year":2020,"lang":"en","type":"article","venue":"Proceedings of the National Academy of Sciences","topic":"Reinforcement Learning in Robotics","field":"Computer Science","cited_by":82,"is_retracted":false,"has_abstract":true,"ca_institutions":"McGill University","funders":"","keywords":"Reinforcement learning; Computer science; Leverage (statistics); Artificial intelligence; Machine learning; Generalization; Exploit; 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":[],"consensus_categories":[],"category_scores_codex":[0.0006320578,0.0001123471,0.0001542231,0.0001463388,0.0002510226,0.00009361144,0.001951805,0.00004351865,0.000009302796],"category_scores_gemma":[0.0005228918,0.00007039613,0.00005475882,0.001596803,0.0004430374,0.0008462062,0.0004701646,0.000214671,0.000003881366],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00003467925,"about_ca_system_score_gemma":0.00008320865,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000006258195,"about_ca_topic_score_gemma":6.563727e-9,"domain_scores_codex":[0.9976212,0.000008244897,0.000328174,0.0002699801,0.001566626,0.0002057745],"domain_scores_gemma":[0.9990275,0.00005921402,0.000567832,0.00001103851,0.000270005,0.00006441613],"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":[0.000008447074,0.000005222345,0.002208882,0.00003399441,0.00001518287,8.465909e-9,0.0006487586,0.6896371,0.02221732,0.2847225,0.0003032238,0.0001994134],"study_design_scores_gemma":[0.0004723302,0.0003409732,0.005686073,0.00009385088,0.000009927547,0.000008041959,0.0001802359,0.8371372,0.1494705,0.0047195,0.001677635,0.0002038268],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.4610352,0.0002539772,0.1234626,0.2471734,0.0001165092,0.001966177,0.000005298393,0.000607773,0.1653791],"genre_scores_gemma":[0.9764817,0.00001672479,0.02161152,0.001473641,0.0000822487,0.000004844334,9.786515e-8,0.000003919005,0.0003252908],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.5154465,"threshold_uncertainty_score":0.3626971,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03975041601643639,"score_gpt":0.2885275840620797,"score_spread":0.2487771680456433,"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."}}