{"id":"W4414969206","doi":"10.48550/arxiv.2510.03494","title":"Trajectory Data Suffices for Statistically Efficient Policy Evaluation in Fixed-Horizon Offline RL with Linear $q^π$-Realizability and Concentrability","year":2025,"lang":"en","type":"preprint","venue":"ArXiv.org","topic":"Nuclear reactor physics and engineering","field":"Engineering","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; Trajectory; Function (biology); Sample (material); Value (mathematics); Sample complexity; Work (physics); Q-learning","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":[0.007203913,0.001795931,0.002940811,0.000608829,0.0007480087,0.001935909,0.002545153,0.002526994,0.003849552],"category_scores_gemma":[0.04604023,0.001286531,0.001295688,0.0006839532,0.003134326,0.005438661,0.003505019,0.006421866,0.0005073654],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.003039632,"about_ca_system_score_gemma":0.003542961,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005753909,"about_ca_topic_score_gemma":0.003709189,"domain_scores_codex":[0.9963007,0.001606959,0.000173504,0.0008153951,0.000668434,0.0004349813],"domain_scores_gemma":[0.9444655,0.04860784,0.002304345,0.002447234,0.001340907,0.0008341541],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0003514837,0.0001391499,0.00111528,0.0001406353,0.00006621672,0.0001080238,0.00009995812,0.9375998,0.0009173641,0.04371634,0.0005892181,0.01515648],"study_design_scores_gemma":[0.00002600054,0.00004424629,0.0001010942,0.00001172205,0.000007560615,0.00001208018,0.000007064616,0.9792123,0.000588448,0.01984836,0.0001337034,0.000007429604],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.04423339,0.0002652108,0.9508273,0.001175203,0.00003061731,0.00009878725,0.0002004883,0.0006240719,0.00254492],"genre_scores_gemma":[0.8502439,0.0002965263,0.1444414,0.0005312133,0.0001112951,0.0004591027,0.000485034,0.0003089219,0.003122613],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.007203913,"threshold_uncertainty_score":0.03809834,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.05665222199401565,"score_gpt":0.3138556402433033,"score_spread":0.2572034182492877,"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."}}