{"id":"W2890047470","doi":"10.48550/arxiv.1811.09013","title":"An Off-policy Policy Gradient Theorem Using Emphatic Weightings","year":2018,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Reinforcement Learning in Robotics","field":"Computer Science","cited_by":4,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Alberta","funders":"","keywords":"Unicode; Counterexample; Reinforcement learning; Computer science; Key (lock); Mathematical economics; Mathematics; Mathematical optimization; Artificial intelligence; Discrete mathematics; Computer security","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":[0.0005013978,0.0005464406,0.0004707989,0.001121659,0.0004293174,0.0004359483,0.004022846,0.0003693894,0.0000329348],"category_scores_gemma":[0.0001561581,0.0006149259,0.0002641007,0.001516822,0.0003760344,0.0008040623,0.003087251,0.0006216188,0.0001598966],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001035347,"about_ca_system_score_gemma":0.001117977,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0008401272,"about_ca_topic_score_gemma":0.00001199376,"domain_scores_codex":[0.9967833,0.0003255153,0.0003818591,0.0014408,0.0002432403,0.0008253332],"domain_scores_gemma":[0.9959084,0.0001033606,0.0006459751,0.002639437,0.0003017199,0.0004010713],"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":[0.000008286493,0.00004234383,0.0008377396,0.0000449344,0.00006041215,0.00007143701,0.0007048257,0.6821077,0.00003659066,0.3156889,0.00004609282,0.0003507597],"study_design_scores_gemma":[0.0003329636,0.0001486849,0.0004650459,0.000144059,0.00005974347,0.00001349129,0.00005601369,0.9572458,0.0001625125,0.04017904,0.0005746679,0.0006180115],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.2473134,0.00001688806,0.7469126,0.0001130817,0.0005151219,0.0002897156,0.000003707238,0.0003822064,0.004453248],"genre_scores_gemma":[0.9868869,0.00009647676,0.01052261,0.000331883,0.0007137642,5.946937e-7,0.00001527618,0.0000489517,0.001383575],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.7395734,"threshold_uncertainty_score":0.9996302,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.06837137357235425,"score_gpt":0.2310214684100404,"score_spread":0.1626500948376862,"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."}}