Challenges in High-dimensional Reinforcement Learning with Evolution Strategies

TL;DR

This paper investigates the effectiveness of Evolution Strategies in high-dimensional, stochastic reinforcement learning tasks, revealing their limitations and proposing promising algorithmic combinations for improved performance.

Contribution

It provides a comprehensive analysis of ES suitability for high-dimensional stochastic problems and suggests combining reduced-cost ES with uncertainty handling techniques.

Findings

ESs have limitations in high-dimensional stochastic settings

Certain algorithmic mechanisms improve ES performance in these problems

Combining reduced-cost ES with uncertainty handling is promising

Abstract

Evolution Strategies (ESs) have recently become popular for training deep neural networks, in particular on reinforcement learning tasks, a special form of controller design. Compared to classic problems in continuous direct search, deep networks pose extremely high-dimensional optimization problems, with many thousands or even millions of variables. In addition, many control problems give rise to a stochastic fitness function. Considering the relevance of the application, we study the suitability of evolution strategies for high-dimensional, stochastic problems. Our results give insights into which algorithmic mechanisms of modern ES are of value for the class of problems at hand, and they reveal principled limitations of the approach. They are in line with our theoretical understanding of ESs. We show that combining ESs that offer reduced internal algorithm cost with uncertainty handling techniques yields promising methods for this class of problems.