Pith. sign in

REVIEW 5 cited by

Statistical Inference for Policy Evaluation with Temporal Difference Learning

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2410.16106 v6 pith:I5KD3U3H submitted 2024-10-21 stat.ML cs.LG

Statistical Inference for Policy Evaluation with Temporal Difference Learning

classification stat.ML cs.LG
keywords learningresultsapproximationasymptoticconfidencedifferencefunctionlinear
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
read the original abstract

We investigate the statistical properties of Temporal Difference (TD) learning with Polyak-Ruppert averaging, arguably one of the most widely used algorithms in reinforcement learning, for the task of estimating the parameters of the optimal linear approximation to the value function. Assuming independent samples, we make three theoretical contributions that improve upon the current state-of-the-art results: (i) we establish refined high-dimensional Berry-Esseen bounds over the class of convex sets, achieving faster rates than the best known results, and (ii) we propose and analyze a novel, computationally efficient online plug-in estimator of the asymptotic covariance matrix; (iii) we derive sharper high probability convergence guarantees that depend explicitly on the asymptotic variance and hold under weaker conditions than those adopted in the literature. These results enable the construction of confidence regions and simultaneous confidence intervals for the linear parameters of the value function approximation, with guaranteed finite-sample coverage. We demonstrate the applicability of our theoretical findings through numerical experiments.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 5 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Gaussian Approximation and Multiplier Bootstrap for Federated Linear Stochastic Approximation

    stat.ML 2026-05 unverdicted novelty 7.0

    Establishes non-asymptotic Gaussian approximation bounds for federated LSA with explicit communication-heterogeneity trade-offs and introduces an online multiplier bootstrap for last-iterate inference with validity gu...

  2. Gaussian Approximation for Asynchronous Q-learning

    stat.ML 2026-04 unverdicted novelty 7.0

    Derived rates of order up to n^{-1/6} log^4(n S A) for the high-dimensional CLT of averaged asynchronous Q-learning iterates, plus a general martingale-difference CLT.

  3. Sharp asymptotic theory for Q-learning with LDTZ learning rate and its generalization

    stat.ML 2026-04 unverdicted novelty 6.0

    Q-learning with PD2Z/LD2Z step sizes admits sharp non-asymptotic bounds, a tail Polyak–Ruppert CLT, and a time-uniform Gaussian approximation, establishing a best-of-both-worlds rate-and-bias tradeoff.

  4. Gaussian Approximation and Multiplier Bootstrap for Stochastic Gradient Descent

    stat.ML 2025-02 unverdicted novelty 6.0

    Proves the first fully non-asymptotic bound on the accuracy of multiplier bootstrap for constructing confidence sets from Polyak-Ruppert SGD iterates, achieving convex-distance rates up to 1/sqrt(n) under regularity c...

  5. On Gaussian approximation for entropy-regularized Q-learning with function approximation

    stat.ML 2026-05 unverdicted novelty 5.0

    Establishes n^{-1/4} Gaussian approximation in convex distance for averaged entropy-regularized Q-learning with linear function approximation and polynomial stepsizes.