Particle Value Functions

Chris J. Maddison, Dieterich Lawson, George Tucker, Nicolas Heess, Arnaud Doucet, Andriy Mnih, Yee Whye Teh

Abstract: The policy gradients of the expected return objective can react slowly to rare rewards. Yet, in some cases agents may wish to emphasize the low or high returns regardless of their probability. Borrowing from the economics and control literature, we review the risk-sensitive value function that arises from an exponential utility and illustrate its effects on an example. This risk-sensitive value function is not always applicable to reinforcement learning problems, so we introduce the particle value function defined by a particle filter over the distributions of an agent's experience, which bounds the risk-sensitive one. We illustrate the benefit of the policy gradients of this objective in Cliffworld.

TL;DR: Value functions that use particle filters to allow reinforcement learning agents to seek or avoid areas of high variance in the reward.

Keywords: Reinforcement Learning, Games

Conflicts: google.com, stats.ox.ac.uk, cs.toronto.edu