Finite Sample Analysis of Two-Timescale Stochastic Approximation with Applications to Reinforcement Learning

[edit]

Gal Dalal, Gugan Thoppe, Balázs Szörényi, Shie Mannor ;
Proceedings of the 31st Conference On Learning Theory, PMLR 75:1199-1233, 2018.

Abstract

Two-timescale Stochastic Approximation (SA) algorithms are widely used in Reinforcement Learning (RL). Their iterates have two parts that are updated using distinct stepsizes. In this work, we develop a novel recipe for their finite sample analysis. Using this, we provide a concentration bound, which is the first such result for a two-timescale SA. The type of bound we obtain is known as “lock-in probability”. We also introduce a new projection scheme, in which the time between successive projections increases exponentially. This scheme allows one to elegantly transform a lock-in probability into a convergence rate result for projected two-timescale SA. From this latter result, we then extract key insights on stepsize selection. As an application, we finally obtain convergence rates for the projected two-timescale RL algorithms GTD(0), GTD2, and TDC.

Related Material