{"id":"W2998461398","doi":"10.1609/aaai.v34i04.5784","title":"Fixed-Horizon Temporal Difference Methods for Stable Reinforcement Learning","year":2020,"lang":"en","type":"article","venue":"","topic":"Reinforcement Learning in Robotics","field":"Computer Science","cited_by":21,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Toronto; Huawei Technologies (Canada); University of Alberta","funders":"Natural Sciences and Engineering Research Council of Canada; Alberta Innovates; DeepMind","keywords":"Reinforcement learning; Bellman equation; Horizon; Temporal difference learning; Function (biology); Time horizon; Value (mathematics); Stability (learning theory); Computer science; Function approximation; Mathematics; Mathematical optimization; Artificial intelligence; Machine learning; Artificial neural network","routes":{"ca_aff":true,"ca_fund":true,"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":[],"consensus_categories":[],"category_scores_codex":[0.0005750986,0.0002233643,0.0002790182,0.00006150659,0.0002494002,0.0003062802,0.001049083,0.00007581697,0.0001044194],"category_scores_gemma":[0.0004962818,0.0001983732,0.0001118051,0.0003731598,0.00003092031,0.0004183268,0.0005247176,0.0002813631,0.00007023827],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0000570749,"about_ca_system_score_gemma":0.00009123614,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00002304372,"about_ca_topic_score_gemma":5.268042e-7,"domain_scores_codex":[0.9981241,0.0001315987,0.0004265705,0.000502308,0.0003058839,0.0005095368],"domain_scores_gemma":[0.998696,0.0003489527,0.0001898382,0.0004077763,0.0001245456,0.0002329086],"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":[0.00002030552,0.000009763421,0.0008841428,0.00007300606,0.00003555861,0.000001721718,0.0008168794,0.8768551,0.003173952,0.06458395,0.001505949,0.05203969],"study_design_scores_gemma":[0.000395527,0.0009545786,0.00006612436,0.00001120956,0.000007884082,9.765346e-7,0.00006324314,0.9186433,0.005834373,0.0002324384,0.07353801,0.0002523276],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.0001337728,0.00003383684,0.9890354,0.001716786,0.0002730745,0.0004457372,1.84282e-7,0.0005711516,0.007790058],"genre_scores_gemma":[0.2067023,0.00001198407,0.7830129,0.0009232698,0.00009826534,0.00004797069,0.00001172498,0.0000207308,0.009170874],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.2065685,"threshold_uncertainty_score":0.8089423,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.05269488141129239,"score_gpt":0.328042082555781,"score_spread":0.2753472011444886,"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."}}