A new Gradient TD Algorithm with only One Step-size: Convergence Rate Analysis using $L$-$\lambda$ Smoothness

TL;DR

This paper introduces a new single-step-size Gradient TD algorithm called Impression GTD, which achieves faster convergence rates, including linear convergence under $L$-$\lambda$ smoothness, improving upon existing methods for off-policy learning.

Contribution

The paper proposes a truly single-time-scale GTD algorithm with only one step-size parameter and proves it converges at a rate of $O(1/t)$, with linear convergence under $L$-$\lambda$ smoothness, advancing the theoretical understanding of GTD algorithms.

Findings

Impression GTD converges faster than existing GTD algorithms.

The new algorithm achieves at least $O(1/t)$ convergence rate.

Empirical results show significant speedup in convergence on benchmark problems.

Abstract

Gradient Temporal Difference (GTD) algorithms (Sutton et al., 2008, 2009) are the first $O(d)$ ($d$ is the number features) algorithms that have convergence guarantees for off-policy learning with linear function approximation. Liu et al. (2015) and Dalal et. al. (2018) proved the convergence rates of GTD, GTD2 and TDC are $O(t^{-\alpha/2})$ for some $\alpha \in (0,1)$. This bound is tight (Dalal et al., 2020), and slower than $O(1/\sqrt{t})$. GTD algorithms also have two step-size parameters, which are difficult to tune. In literature, there is a "single-time-scale" formulation of GTD. However, this formulation still has two step-size parameters. This paper presents a truly single-time-scale GTD algorithm for minimizing the Norm of Expected td Update (NEU) objective, and it has only one step-size parameter. We prove that the new algorithm, called Impression GTD, converges at least as fast as $O(1/t)$. Furthermore, based on a generalization of the expected smoothness (Gower et al. 2019), called $L$-$\lambda$ smoothness, we are able to prove that the new GTD converges even faster, in fact, with a linear rate. Our rate actually also improves Gower et al.'s result with a tighter bound under a weaker assumption. Besides Impression GTD, we also prove the rates of three other GTD algorithms, one by Yao and Liu (2008), another called A-transpose-TD (Sutton et al., 2008), and a counterpart of A-transpose-TD. The convergence rates of all the four GTD algorithms are proved in a single generic GTD framework to which $L$-$\lambda$ smoothness applies. Empirical results on Random walks, Boyan chain, and Baird counterexample show that Impression GTD converges much faster than existing GTD algorithms for both on-policy and off-policy learning problems, with well-performing step-sizes in a big range.