Pith. sign in

REVIEW 10 cited by

Constant Stepsize Q-learning: Distributional Convergence, Bias and Extrapolation

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 2401.13884 v1 pith:PFHTDIQ7 submitted 2024-01-25 stat.ML cs.LGmath.OC

Constant Stepsize Q-learning: Distributional Convergence, Bias and Extrapolation

classification stat.ML cs.LGmath.OC
keywords q-learningstepsizebiasconvergenceconstantextrapolationiteratesasymptotic
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
read the original abstract

Stochastic Approximation (SA) is a widely used algorithmic approach in various fields, including optimization and reinforcement learning (RL). Among RL algorithms, Q-learning is particularly popular due to its empirical success. In this paper, we study asynchronous Q-learning with constant stepsize, which is commonly used in practice for its fast convergence. By connecting the constant stepsize Q-learning to a time-homogeneous Markov chain, we show the distributional convergence of the iterates in Wasserstein distance and establish its exponential convergence rate. We also establish a Central Limit Theory for Q-learning iterates, demonstrating the asymptotic normality of the averaged iterates. Moreover, we provide an explicit expansion of the asymptotic bias of the averaged iterate in stepsize. Specifically, the bias is proportional to the stepsize up to higher-order terms and we provide an explicit expression for the linear coefficient. This precise characterization of the bias allows the application of Richardson-Romberg (RR) extrapolation technique to construct a new estimate that is provably closer to the optimal Q function. Numerical results corroborate our theoretical finding on the improvement of the RR extrapolation method.

discussion (0)

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

Forward citations

Cited by 10 Pith papers

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

  1. Wasserstein-p Central Limit Theorem Rates: From Local Dependence to Markov Chains

    math.PR 2026-01 unverdicted novelty 8.0

    The paper proves the first optimal O(n^{-1/2}) Wasserstein-1 CLT rates for locally dependent sequences and geometrically ergodic Markov chains, plus new W_p rates for p greater than or equal to 2 under mild moments, w...

  2. 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...

  3. Shuffling the Data, Stretching the Step-size: Sharper Bias in constant step-size SGD

    math.OC 2026-04 unverdicted novelty 7.0

    Combining random reshuffling and Richardson-Romberg extrapolation yields cubic bias refinement and better MSE for constant-step SGD on structured non-monotone variational inequalities.

  4. A Minimal-Assumption Analysis of Q-Learning with Time-Varying Policies

    cs.LG 2025-10 unverdicted novelty 7.0

    Establishes last-iterate convergence rates for on-policy Q-learning under minimal irreducibility assumptions, with sample complexity O(1/ξ²) matching off-policy up to exploration factors.

  5. From Set Convergence to Pointwise Convergence: Finite-Time Guarantees for Average-Reward Q-Learning with Adaptive Stepsizes

    cs.LG 2025-04 unverdicted novelty 7.0

    Establishes Õ(1/k) mean-square last-iterate convergence for asynchronous average-reward Q-learning with adaptive stepsizes and proves adaptivity is necessary.

  6. SGD at the Edge of Stability: Stochastic Stabilization with Large Learning Rates

    stat.ML 2026-06 unverdicted novelty 6.0

    SGD on multiclass cross-entropy loss alternates between curvature-driven oscillations and stable regimes but self-stabilizes to enable best-iterate convergence with large learning rates for linear and two-layer models.

  7. Elephant random walk with attributed steps and extractions of random sizes

    math.PR 2026-04 unverdicted novelty 6.0

    A market choice model with random-size sampling from past customers is represented as an elephant random walk variant, with proofs of almost sure convergence of S_n/n and regime-dependent distributional limits for scaled S_n.

  8. Revisiting the Constant Stepsize Stochastic Approximation with Decision-Dependent Markovian Noise

    math.OC 2026-04 unverdicted novelty 6.0

    Constant stepsize SA with decision-dependent Markovian noise has stationary bias O(alpha) under Poisson-Gateaux differentiability, plus finite-time moment bounds and weak convergence.

  9. 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.

  10. Central Limit Theorems for Asynchronous Averaged Q-Learning

    cs.LG 2025-09 unverdicted novelty 6.0

    Establishes non-asymptotic and functional central limit theorems for asynchronous averaged Q-learning with explicit rates depending on iterations, state-action space, discount factor, and exploration quality.