{"id":"W2884439071","doi":"10.1017/s0269888921000035","title":"Safe option-critic: learning safety in the option-critic architecture","year":2021,"lang":"en","type":"article","venue":"The Knowledge Engineering Review","topic":"Reinforcement Learning in Robotics","field":"Computer Science","cited_by":24,"is_retracted":false,"has_abstract":true,"ca_institutions":"McGill University","funders":"","keywords":"Computer science; Reinforcement learning; Consistency (knowledge bases); State space; Variance (accounting); Function (biology); Artificial intelligence; Grid; Architecture; 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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001899035,0.0008115152,0.000737307,0.0004039898,0.0002689522,0.0008881319,0.001536203,0.001056235,0.00243098],"category_scores_gemma":[0.004377479,0.0004414347,0.0004495564,0.0003526579,0.001352541,0.001177466,0.0007913068,0.002176063,0.0003744255],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001156152,"about_ca_system_score_gemma":0.001118374,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004230028,"about_ca_topic_score_gemma":0.003478758,"domain_scores_codex":[0.9994429,0.0002896112,0.00002447894,0.00007544319,0.0001221962,0.0000453819],"domain_scores_gemma":[0.9981077,0.001255164,0.0001400535,0.0001104677,0.00027106,0.0001155988],"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":[0.00005439365,0.00002717869,0.0005430199,0.00008852348,0.00005457103,0.00004121206,0.00004197015,0.9240856,0.0006314852,0.03856549,0.0009390312,0.03492761],"study_design_scores_gemma":[0.0000120426,0.00001833206,0.00005728444,0.00001341974,0.000006689853,0.000009385441,0.00000270337,0.9802126,0.0002701087,0.01894643,0.0004459972,0.000005000377],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.02087674,0.001304909,0.971394,0.0007230595,0.00006123701,0.000034587,0.0000340141,0.0005067214,0.005064765],"genre_scores_gemma":[0.847465,0.001231755,0.1458494,0.0002450774,0.00006660423,0.0001189484,0.00007270573,0.0001241207,0.004826389],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.004230028,"threshold_uncertainty_score":0.01004314,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01161079066951347,"score_gpt":0.255187380824089,"score_spread":0.2435765901545756,"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."}}