Reasoning and Generalization in RL: A Tool Use Perspective
This work addresses the need for better generalization evaluation in RL, offering a domain-specific approach that is incremental in nature.
The paper tackles the problem of measuring generalization in reinforcement learning by proposing a framework inspired by tool use tasks, specifically the trap-tube task, to create multiple test sets for evaluating specified generalization types, with results demonstrated through publicly available code.
Learning to use tools to solve a variety of tasks is an innate ability of humans and has been observed of animals in the wild. However, the underlying mechanisms that are required to learn to use tools are abstract and widely contested in the literature. In this paper, we study tool use in the context of reinforcement learning and propose a framework for analyzing generalization inspired by a classic study of tool using behavior, the trap-tube task. Recently, it has become common in reinforcement learning to measure generalization performance on a single test set of environments. We instead propose transfers that produce multiple test sets that are used to measure specified types of generalization, inspired by abilities demonstrated by animal and human tool users. The source code to reproduce our experiments is publicly available at https://github.com/fomorians/gym_tool_use.