Adversarially-Robust TD Learning with Markovian Data: Finite-Time Rates and Fundamental Limits

• URL: http://arxiv.org/abs/2502.04662v1
• Date: Fri, 07 Feb 2025 05:05:42 GMT
• Title: Adversarially-Robust TD Learning with Markovian Data: Finite-Time Rates and Fundamental Limits
• Authors: Sreejeet Maity, Aritra Mitra,
• Abstract summary: Motivated by harsh, real-world environments, we revisit the policy evaluation problem from the perspective of adversarial robustness. We develop a novel algorithm called Robust-TD and prove that its finite-time guarantees match that of vanilla TD with linear function approximation up to a small $O(epsilon)$ term. To our knowledge, these results are the first of their kind in the context of adversarial approximation schemes driven by Markov noise.
• Score: 2.07180164747172
• License:
• Abstract: One of the most basic problems in reinforcement learning (RL) is policy evaluation: estimating the long-term return, i.e., value function, corresponding to a given fixed policy. The celebrated Temporal Difference (TD) learning algorithm addresses this problem, and recent work has investigated finite-time convergence guarantees for this algorithm and variants thereof. However, these guarantees hinge on the reward observations being always generated from a well-behaved (e.g., sub-Gaussian) true reward distribution. Motivated by harsh, real-world environments where such an idealistic assumption may no longer hold, we revisit the policy evaluation problem from the perspective of adversarial robustness. In particular, we consider a Huber-contaminated reward model where an adversary can arbitrarily corrupt each reward sample with a small probability $\epsilon$. Under this observation model, we first show that the adversary can cause the vanilla TD algorithm to converge to any arbitrary value function. We then develop a novel algorithm called Robust-TD and prove that its finite-time guarantees match that of vanilla TD with linear function approximation up to a small $O(\epsilon)$ term that captures the effect of corruption. We complement this result with a minimax lower bound, revealing that such an additive corruption-induced term is unavoidable. To our knowledge, these results are the first of their kind in the context of adversarial robustness of stochastic approximation schemes driven by Markov noise. The key new technical tool that enables our results is an analysis of the Median-of-Means estimator with corrupted, time-correlated data that might be of independent interest to the literature on robust statistics.

Related papers

• Regret Minimization and Statistical Inference in Online Decision Making with High-dimensional Covariates [7.21848268647674]
  We integrate the $varepsilon$-greedy bandit algorithm for decision-making with a hard thresholding algorithm for estimating sparse bandit parameters. Under a margin condition, our method achieves either $O(T1/2)$ regret or classical $O(T1/2)$-consistent inference.
  arXiv  Detail & Related papers  (2024-11-10T01:47:11Z)
• Robust Q-Learning under Corrupted Rewards [2.07180164747172]
  We investigate the robustness of the celebrated Q-learning algorithm to a strong-contamination attack model. We develop a novel robust synchronous Q-learning algorithm that uses historical reward data to construct robust empirical Bellman operators. Our results continue to hold even when the true reward distributions have infinite support, provided they admit bounded second moments.
  arXiv  Detail & Related papers  (2024-09-05T04:37:02Z)
• Efficient and Sharp Off-Policy Evaluation in Robust Markov Decision Processes [44.974100402600165]
  We study the evaluation of a policy best-parametric and worst-case perturbations to a decision process (MDP) We use transition observations from the original MDP, whether they are generated under the same or a different policy. Our estimator is also estimated statistical inference using Wald confidence intervals.
  arXiv  Detail & Related papers  (2024-03-29T18:11:49Z)
• Corruption-Robust Algorithms with Uncertainty Weighting for Nonlinear Contextual Bandits and Markov Decision Processes [59.61248760134937]
  We propose an efficient algorithm to achieve a regret of $tildeO(sqrtT+zeta)$. The proposed algorithm relies on the recently developed uncertainty-weighted least-squares regression from linear contextual bandit. We generalize our algorithm to the episodic MDP setting and first achieve an additive dependence on the corruption level $zeta$.
  arXiv  Detail & Related papers  (2022-12-12T15:04:56Z)
• A Robust Phased Elimination Algorithm for Corruption-Tolerant Gaussian Process Bandits [118.22458816174144]
  We propose a novel robust elimination-type algorithm that runs in epochs, combines exploration with infrequent switching to select a small subset of actions, and plays each action for multiple time instants. Our algorithm, GP Robust Phased Elimination (RGP-PE), successfully balances robustness to corruptions with exploration and exploitation. We perform the first empirical study of robustness in the corrupted GP bandit setting, and show that our algorithm is robust against a variety of adversarial attacks.
  arXiv  Detail & Related papers  (2022-02-03T21:19:36Z)
• Anti-Concentrated Confidence Bonuses for Scalable Exploration [57.91943847134011]
  Intrinsic rewards play a central role in handling the exploration-exploitation trade-off. We introduce emphanti-concentrated confidence bounds for efficiently approximating the elliptical bonus. We develop a practical variant for deep reinforcement learning that is competitive with contemporary intrinsic rewards on Atari benchmarks.
  arXiv  Detail & Related papers  (2021-10-21T15:25:15Z)
• Risk Minimization from Adaptively Collected Data: Guarantees for Supervised and Policy Learning [57.88785630755165]
  Empirical risk minimization (ERM) is the workhorse of machine learning, but its model-agnostic guarantees can fail when we use adaptively collected data. We study a generic importance sampling weighted ERM algorithm for using adaptively collected data to minimize the average of a loss function over a hypothesis class. For policy learning, we provide rate-optimal regret guarantees that close an open gap in the existing literature whenever exploration decays to zero.
  arXiv  Detail & Related papers  (2021-06-03T09:50:13Z)
• Adversarial Robustness Guarantees for Gaussian Processes [22.403365399119107]
  Gaussian processes (GPs) enable principled computation of model uncertainty, making them attractive for safety-critical applications. We present a framework to analyse adversarial robustness of GPs, defined as invariance of the model's decision to bounded perturbations. We develop a branch-and-bound scheme to refine the bounds and show, for any $epsilon > 0$, that our algorithm is guaranteed to converge to values $epsilon$-close to the actual values in finitely many iterations.
  arXiv  Detail & Related papers  (2021-04-07T15:14:56Z)
• CoinDICE: Off-Policy Confidence Interval Estimation [107.86876722777535]
  We study high-confidence behavior-agnostic off-policy evaluation in reinforcement learning. We show in a variety of benchmarks that the confidence interval estimates are tighter and more accurate than existing methods.
  arXiv  Detail & Related papers  (2020-10-22T12:39:11Z)
• Minimax-Optimal Off-Policy Evaluation with Linear Function Approximation [49.502277468627035]
  This paper studies the statistical theory of batch data reinforcement learning with function approximation. Consider the off-policy evaluation problem, which is to estimate the cumulative value of a new target policy from logged history.
  arXiv  Detail & Related papers  (2020-02-21T19:20:57Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.