Policy Search with Rare Significant Events: Choosing the Right Partner to Cooperate with
This addresses a specific challenge in reinforcement learning for agents in environments with rare rewards, but it is incremental as it builds on known methods without introducing new paradigms.
The paper tackles reinforcement learning problems where significant events are rare, such as an agent choosing a cooperative partner, by comparing gradient and direct policy search methods. It finds that gradient methods struggle due to scarce gradient information, while direct methods like evolution strategies remain effective, confirming their unique role in such scenarios.
This paper focuses on a class of reinforcement learning problems where significant events are rare and limited to a single positive reward per episode. A typical example is that of an agent who has to choose a partner to cooperate with, while a large number of partners are simply not interested in cooperating, regardless of what the agent has to offer. We address this problem in a continuous state and action space with two different kinds of search methods: a gradient policy search method and a direct policy search method using an evolution strategy. We show that when significant events are rare, gradient information is also scarce, making it difficult for policy gradient search methods to find an optimal policy, with or without a deep neural architecture. On the other hand, we show that direct policy search methods are invariant to the rarity of significant events, which is yet another confirmation of the unique role evolutionary algorithms has to play as a reinforcement learning method.