{"id":"W2966234803","doi":"10.24963/ijcai.2019/445","title":"Hill Climbing on Value Estimates for Search-control in Dyna","year":2019,"lang":"en","type":"preprint","venue":"","topic":"Reinforcement Learning in Robotics","field":"Computer Science","cited_by":8,"is_retracted":false,"has_abstract":true,"ca_institutions":"Vector Institute; University of Toronto; Huawei Technologies (Canada); University of Alberta","funders":"","keywords":"Hill climbing; Climbing; Reinforcement learning; Trajectory; Computer science; Bellman equation; Langevin dynamics; Value (mathematics); Sampling (signal processing); Artificial intelligence; Mathematical optimization; Mathematics; Machine learning; Statistics; Engineering","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":[0.001421606,0.000323229,0.0004651127,0.000330888,0.00006444968,0.0004270705,0.001859926,0.0002822005,0.00001983703],"category_scores_gemma":[0.000341273,0.0002960058,0.0001803813,0.00009558348,0.00003169982,0.0001734728,0.001212877,0.0008443655,0.0001984736],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002075218,"about_ca_system_score_gemma":0.0002138213,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001025106,"about_ca_topic_score_gemma":0.00000374755,"domain_scores_codex":[0.997497,0.0001198367,0.0004977122,0.0008212276,0.00047186,0.0005923885],"domain_scores_gemma":[0.9971222,0.001219867,0.000185327,0.001280904,0.0001031243,0.00008853056],"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":[0.00001220595,0.00002247558,0.001101149,0.0001579621,0.00002915266,0.000003317383,0.0001492355,0.9108411,0.00002496145,0.08531212,0.0002289043,0.002117409],"study_design_scores_gemma":[0.0007796391,0.0001900094,0.00060036,0.0002742534,0.000009702925,0.000001083166,0.00001165297,0.9945856,0.0002190175,0.002799442,0.0002034093,0.0003258724],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0005584119,0.00002637605,0.9790287,0.001193778,0.001205352,0.001471186,0.000003896317,0.0002482038,0.01626408],"genre_scores_gemma":[0.7895821,0.0000103139,0.2068227,0.001155941,0.0001155013,0.0001089471,0.00002737995,0.00004216022,0.002135024],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.7890236,"threshold_uncertainty_score":0.9999492,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03048606102777367,"score_gpt":0.2988601795223709,"score_spread":0.2683741184945972,"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."}}