ResearchTrend.AI
  • Papers
  • Communities
  • Events
  • Blog
  • Pricing
Papers
Communities
Social Events
Terms and Conditions
Pricing
Parameter LabParameter LabTwitterGitHubLinkedInBlueskyYoutube

© 2025 ResearchTrend.AI, All rights reserved.

  1. Home
  2. Papers
  3. 2311.07876
22
2

Learning Adversarial Low-rank Markov Decision Processes with Unknown Transition and Full-information Feedback

14 November 2023
Canzhe Zhao
Ruofeng Yang
Baoxiang Wang
Xuezhou Zhang
Shuai Li
ArXivPDFHTML
Abstract

In this work, we study the low-rank MDPs with adversarially changed losses in the full-information feedback setting. In particular, the unknown transition probability kernel admits a low-rank matrix decomposition \citep{REPUCB22}, and the loss functions may change adversarially but are revealed to the learner at the end of each episode. We propose a policy optimization-based algorithm POLO, and we prove that it attains the O~(K56A12dln⁡(1+M)/(1−γ)2)\widetilde{O}(K^{\frac{5}{6}}A^{\frac{1}{2}}d\ln(1+M)/(1-\gamma)^2)O(K65​A21​dln(1+M)/(1−γ)2) regret guarantee, where ddd is rank of the transition kernel (and hence the dimension of the unknown representations), AAA is the cardinality of the action space, MMM is the cardinality of the model class, and γ\gammaγ is the discounted factor. Notably, our algorithm is oracle-efficient and has a regret guarantee with no dependence on the size of potentially arbitrarily large state space. Furthermore, we also prove an Ω(γ21−γdAK)\Omega(\frac{\gamma^2}{1-\gamma} \sqrt{d A K})Ω(1−γγ2​dAK​) regret lower bound for this problem, showing that low-rank MDPs are statistically more difficult to learn than linear MDPs in the regret minimization setting. To the best of our knowledge, we present the first algorithm that interleaves representation learning, exploration, and exploitation to achieve the sublinear regret guarantee for RL with nonlinear function approximation and adversarial losses.

View on arXiv
Comments on this paper