{"id":"W2945374007","doi":"10.1007/978-3-030-18305-9_68","title":"Safe Policy Learning with Constrained Return Variance","year":2019,"lang":"en","type":"book-chapter","venue":"Lecture notes in computer science","topic":"Reinforcement Learning in Robotics","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":false,"ca_institutions":"McGill University; Mila - Quebec Artificial Intelligence Institute","funders":"","keywords":"Computer science; Variance (accounting); Artificial intelligence; Machine learning; Accounting","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":[0.0009777768,0.0007371934,0.0007241208,0.001120347,0.0003486682,0.0009246892,0.004065661,0.0003995896,0.00004505752],"category_scores_gemma":[0.0002805772,0.0006345338,0.0001218196,0.001100728,0.001072986,0.0008128109,0.001410073,0.001980544,0.000179492],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000528103,"about_ca_system_score_gemma":0.002758163,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00003075833,"about_ca_topic_score_gemma":0.00001571752,"domain_scores_codex":[0.994888,0.00006453341,0.0006121797,0.001840994,0.001537522,0.001056763],"domain_scores_gemma":[0.9962356,0.0006899589,0.0006151683,0.001876211,0.0003561,0.0002269534],"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":[0.000008055686,0.000006436781,0.0001585955,0.00004180079,0.00001791151,0.00008291264,0.0004792728,0.8400615,0.0000574368,0.06309342,0.000007902089,0.09598473],"study_design_scores_gemma":[0.0005305195,0.0005839366,0.0001039413,0.0007267168,0.0000100922,0.0002027325,2.783664e-7,0.9778859,0.0002139759,0.01182897,0.00695263,0.0009602833],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.00001116595,0.00009569497,0.9542358,0.001435343,0.0009599048,0.0004935893,0.000001396329,0.0002730445,0.04249405],"genre_scores_gemma":[0.1692175,0.00006410784,0.8123914,0.002861975,0.0009014324,0.000007678427,0.00001213153,0.0001068881,0.01443684],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.1692064,"threshold_uncertainty_score":0.9996106,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01184008478179542,"score_gpt":0.2357360834831554,"score_spread":0.2238959987013599,"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."}}