{"paper":{"title":"Introducing Feedback Delays to Non-Delayed Adversarial MDPs: Optimal $\\sqrt S$ and $\\sqrt K$ Dependencies in Tabular and Linear AMDPs","arxiv_id":"2611.17214","primary_cat":"cs.LG","submitted_at":"2026-11-12T18:10:46+00:00","version_compiled":"v1"},"notice":"Every quoted sentence is the authors' text from the arXiv listing abstract. Labels and probabilities are the outputs of classifier-1.13.0 on the question bank abstract-p1-20260923, exactly as returned. No sentence in this record was written by a language model.","schema":"pith.compiled-record/v0","status":"compiled","abstract":{"source":"arXiv listing abstract v1","sentences":[{"i":0,"p":0.76,"id":"s0","own":true,"role":"background","text":"We study regret minimization in episodic Adversarial Markov Decision Processes (AMDPs) with unknown transitions and bandit feedback.","hedge":{"p":0.81,"label":"asserted"},"own_p":0.52,"main_claim_p":0.02,"quantitative":false,"quantitative_p":0.03,"role_probabilities":{"gap":0.0,"method":0.0,"result":0.0,"background":0.76,"limitation":0.0,"future_work":0.0,"contribution":0.24,"interpretation":0.0}},{"i":1,"p":0.83,"id":"s1","own":false,"role":"gap","text":"In tabular AMDPs with $S$ states, $A$ actions, and horizon $H$, the best known $K$-episode regret is $\\tilde O(HS\\sqrt{AK})$, whose $\\sqrt S$ gap to the $\\Omega(H\\sqrt{SAK})$ lower bound has long remained open.","hedge":{"p":1.0,"label":"asserted"},"own_p":0.04,"main_claim_p":0.03,"quantitative":true,"quantitative_p":0.99,"role_probabilities":{"gap":0.83,"method":0.0,"result":0.0,"background":0.17,"limitation":0.0,"future_work":0.0,"contribution":0.0,"interpretation":0.0}},{"i":2,"p":0.52,"id":"s2","own":false,"role":"background","text":"In linear AMDPs with $d$-dimensional features, after a five-year line of improvements starting from $K^{14/15}$, computationally efficient algorithms without a simulator or exploratory assumptions still only attain $\\tilde O(K^{3/4})$ regret.","hedge":{"p":1.0,"label":"asserted"},"own_p":0.07,"main_claim_p":0.02,"quantitative":true,"quantitative_p":0.98,"role_probabilities":{"gap":0.48,"method":0.0,"result":0.0,"background":0.52,"limitation":0.0,"future_work":0.0,"contribution":0.0,"interpretation":0.0}},{"i":3,"p":0.87,"id":"s3","own":false,"role":"background","text":"A common challenge in both settings is to estimate the Q-function of a changing policy: past data comes from earlier policies that differ from (and are correlated with) the current one, while freezing the policy for future episodes worsens the dependency on $K$.","hedge":{"p":1.0,"label":"asserted"},"own_p":0.12,"main_claim_p":0.02,"quantitative":false,"quantitative_p":0.16,"role_probabilities":{"gap":0.13,"method":0.0,"result":0.0,"background":0.87,"limitation":0.0,"future_work":0.0,"contribution":0.0,"interpretation":0.0}},{"i":4,"p":0.65,"id":"s4","own":true,"role":"method","text":"We instead propose a general technique called introduced feedback delays: we estimate the Q-function of the policy using future episodes while still updating the policy, as if the feedback were delayed.","hedge":{"p":0.99,"label":"asserted"},"own_p":0.9,"main_claim_p":0.08,"quantitative":false,"quantitative_p":0.06,"role_probabilities":{"gap":0.0,"method":0.65,"result":0.0,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.35,"interpretation":0.0}},{"i":5,"p":0.67,"id":"s5","own":true,"role":"interpretation","text":"While feedback delays are usually an obstacle imposed by the environment, here they are introduced voluntarily and, perhaps surprisingly, come at no extra cost.","hedge":{"p":0.87,"label":"asserted"},"own_p":0.66,"main_claim_p":0.07,"quantitative":false,"quantitative_p":0.09,"role_probabilities":{"gap":0.0,"method":0.04,"result":0.12,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.17,"interpretation":0.67}},{"i":6,"p":0.98,"id":"s6","own":true,"role":"method","text":"We develop Geometric Resampling with Rejection (GRR) to account for policy changes during the delay, and threshold the visitation probabilities to avoid long delays on rarely visited state-action pairs.","hedge":{"p":0.99,"label":"asserted"},"own_p":0.91,"main_claim_p":0.03,"quantitative":false,"quantitative_p":0.05,"role_probabilities":{"gap":0.0,"method":0.98,"result":0.0,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.02,"interpretation":0.0}},{"i":7,"p":0.97,"id":"s7","own":true,"role":"result","text":"With this unified technique, our policy optimization algorithms achieve, with high probability, $\\tilde O(\\text{poly}(H)\\sqrt{SAK})$ regret in tabular AMDPs, resolving the open problem; and $\\tilde O(\\text{poly}(d,H)\\sqrt K)$ regret in linear AMDPs, giving the first efficient algorithm optimal in $K$.","hedge":{"p":1.0,"label":"asserted"},"own_p":0.96,"main_claim_p":0.94,"quantitative":true,"quantitative_p":0.98,"role_probabilities":{"gap":0.0,"method":0.0,"result":0.97,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.03,"interpretation":0.0}},{"i":8,"p":0.95,"id":"s8","own":true,"role":"result","text":"They also naturally extend to delayed AMDPs with a total feedback delay $\\mathfrak D$, yielding the first bounds optimal in $S$ and $\\mathfrak D$ (tabular) and in $K$ and $\\mathfrak D$ (linear).","hedge":{"p":1.0,"label":"asserted"},"own_p":0.92,"main_claim_p":0.15,"quantitative":true,"quantitative_p":0.61,"role_probabilities":{"gap":0.0,"method":0.0,"result":0.95,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.04,"interpretation":0.01}}]},"compiled":{"main_claim":{"p":1.0,"text":"With this unified technique, our policy optimization algorithms achieve, with high probability, $\\tilde O(\\text{poly}(H)\\sqrt{SAK})$ regret in tabular AMDPs, resolving the open problem; and $\\tilde O(\\text{poly}(d,H)\\sqrt K)$ regret in linear AMDPs, giving the first efficient algorithm optimal in $K$.","location":"abstract s7","hedge_strength":{"p":1.0,"label":"asserted"}},"provenance":{"route":{"via":"TypeSafe API","model":"classifier-1.13.0","endpoint":"https://api.typesafe.ai/v1/systemone","provider":"TypeSafe"},"packing":"one request per paper","requests":1,"input_tokens":7479,"cost_usd_reported":0.0},"compiled_at":"2026-11-14","bank_version":"abstract-p1-20260923","model_version":"classifier-1.13.0","abstract_states":{"claims_first":0.79,"states_limitation":0.19,"human_participants":0.01,"controlled_comparison":0.08,"randomized_assignment":0.02,"states_quantitative_result":0.98,"compares_with_prior_results":0.97,"mentions_code_or_data_availability":0.02},"evidential_basis":{"p":0.99,"value":"proof or derivation","probabilities":{"argument":0.01,"literature":0.0,"proof_or_derivation":0.99,"analysis_of_existing_data":0.0,"observation_or_experiment":0.0,"simulation_or_computation":0.0}},"contribution_type":{"p":0.79,"value":"method or tool","probabilities":{"review":0.0,"method_or_tool":0.79,"empirical_study":0.0,"simulation_study":0.0,"theoretical_result":0.21,"dataset_or_benchmark":0.0,"position_or_commentary":0.0}},"contribution_statement":{"p":0.91,"text":"We instead propose a general technique called introduced feedback delays: we estimate the Q-function of the policy using future episodes while still updating the policy, as if the feedback were delayed.","location":"abstract s4"}},"coverage":{"tier":"abstract","levels_run":["P abstract"],"not_examined":["full text","figures and tables","references and citing sentences","relations to other papers"],"sources_read":["arXiv listing abstract v1"],"levels_not_run":["S sections","T body sentences","C citations","R cross-paper relations"],"sentences_examined":9,"sentences_in_abstract":9},"measured_accuracy":{"url":"https://pith.science/accuracy","measured_on":"2026-10-02","method":"two independent model families labelled a fresh random sample blind; the figure is how often the published label matches their consensus","sentence_role_agreement":0.92,"main_claim_agreement":0.908,"typed_citation_agreement":0.933,"typed_citation_agreement_95_interval":[0.917,0.945],"review_comments_accurate":0.758,"review_comments_accurate_95_interval":[0.684,0.82],"review_caught_admitted_error_found_or_related":0.448,"review_comments_judged":{"judged_by":["gpt-5.6-luna","grok-4.7"],"judged_against":"Pith's own text copy of each paper: text extracted from the PDF, the first 90,000 characters","measured_on":"2026-10-04","accurate":113,"decided":149,"sampled_reviews":200,"split_by_lane":false,"by_lane":"https://pith.science/accuracy#by-lane"}}}