{"id":"W4315588467","doi":"10.48550/arxiv.2301.02952","title":"Learning Symbolic Representations for Reinforcement Learning of Non-Markovian Behavior","year":2023,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Reinforcement Learning in Robotics","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Natural Sciences and Engineering Research Council of Canada; Government of Canada; Canadian Institute for Advanced Research; Agencia Nacional de Investigación y Desarrollo; Microsoft Research","keywords":"Reinforcement learning; Computer science; Artificial intelligence; Abstraction; Action (physics); Representation (politics); Function (biology); Markov process; State (computer science); Temporal difference learning; Markov decision process; Machine learning; Theoretical computer science; Algorithm; Mathematics","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0006317978,0.0006602121,0.0007134179,0.0005331028,0.0002888779,0.0007877703,0.001124031,0.0007444279,0.003045564],"category_scores_gemma":[0.005093711,0.0004114164,0.00061236,0.0004993194,0.001057405,0.001692622,0.001089553,0.001893525,0.0004454922],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001253719,"about_ca_system_score_gemma":0.001055542,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003555199,"about_ca_topic_score_gemma":0.006328937,"domain_scores_codex":[0.9996321,0.0001236183,0.00002337759,0.0000979886,0.00007860153,0.00004423412],"domain_scores_gemma":[0.9979795,0.001427818,0.0001866556,0.0002182287,0.0001180492,0.00006988272],"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":[0.00008618954,0.00006654269,0.001014339,0.00009076063,0.00003393971,0.00005900305,0.0001107481,0.8939163,0.001356845,0.03762094,0.0008395768,0.06480484],"study_design_scores_gemma":[0.000005656771,0.0000103393,0.0000372825,0.000006310657,0.000002943367,0.000004237967,0.000005929156,0.9784421,0.0002682822,0.02101494,0.000199664,0.000002287767],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.02832949,0.0002435643,0.9687007,0.0002495848,0.00002432475,0.00003868206,0.0001490087,0.0008852819,0.001379434],"genre_scores_gemma":[0.8188559,0.0002976799,0.1777462,0.0001194224,0.00003051814,0.0002132031,0.0005414082,0.0001345543,0.002061145],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.003555199,"threshold_uncertainty_score":0.0101884,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.08623908006050819,"score_gpt":0.2431459818392111,"score_spread":0.1569069017787029,"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."}}