{"id":"W4392780713","doi":"10.48550/arxiv.2403.05385","title":"Switching the Loss Reduces the Cost in Batch (Offline) Reinforcement Learning","year":2024,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Supply Chain and Inventory Management","field":"Business, Management and Accounting","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Natural Sciences and Engineering Research Council of Canada; Canadian Institute for Advanced Research","keywords":"Reinforcement learning; Reinforcement; Computer science; Artificial intelligence; Psychology; Social psychology","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":[0.004790623,0.001964795,0.00194376,0.0004647421,0.0006512866,0.001660396,0.003507878,0.002306007,0.005860519],"category_scores_gemma":[0.02822915,0.0009146708,0.0006976527,0.0004440666,0.002666088,0.004342893,0.002813674,0.00525224,0.001554408],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001977968,"about_ca_system_score_gemma":0.002948792,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004439359,"about_ca_topic_score_gemma":0.003871443,"domain_scores_codex":[0.9974614,0.0009321052,0.0001204564,0.0006039132,0.0005944885,0.0002875261],"domain_scores_gemma":[0.9873959,0.008562913,0.0007684322,0.001755251,0.000941748,0.0005757185],"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":[0.0006993983,0.0007674325,0.003217846,0.0001790318,0.00008707564,0.0001091028,0.0001644826,0.8405204,0.005218612,0.02574295,0.004233418,0.1190603],"study_design_scores_gemma":[0.00003307958,0.0001223201,0.000138253,0.00001024061,0.000007166778,0.00001346573,0.00000689288,0.9910375,0.0007126777,0.007620913,0.0002894886,0.000008061113],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0322152,0.0002805671,0.9604788,0.0007689235,0.00009972427,0.0001757406,0.0000807383,0.00200401,0.003896317],"genre_scores_gemma":[0.8152253,0.000170606,0.1788355,0.0008587093,0.0001247734,0.0004263668,0.0002145216,0.000438342,0.00370591],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.005860519,"threshold_uncertainty_score":0.02533549,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.06598399699070621,"score_gpt":0.1898768654910096,"score_spread":0.1238928685003034,"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."}}