Auxiliary task discovery through generate-and-test
This work addresses the challenge of manual task design in reinforcement learning for researchers and practitioners, though it appears incremental as it builds on existing representation learning and meta-learning ideas.
The paper tackles the problem of automatically discovering useful auxiliary tasks in reinforcement learning to improve data efficiency, and introduces a generate-and-test algorithm that significantly outperforms random tasks and learning without auxiliary tasks across multiple environments.
In this paper, we explore an approach to auxiliary task discovery in reinforcement learning based on ideas from representation learning. Auxiliary tasks tend to improve data efficiency by forcing the agent to learn auxiliary prediction and control objectives in addition to the main task of maximizing reward, and thus producing better representations. Typically these tasks are designed by people. Meta-learning offers a promising avenue for automatic task discovery; however, these methods are computationally expensive and challenging to tune in practice. In this paper, we explore a complementary approach to the auxiliary task discovery: continually generating new auxiliary tasks and preserving only those with high utility. We also introduce a new measure of auxiliary tasks' usefulness based on how useful the features induced by them are for the main task. Our discovery algorithm significantly outperforms random tasks and learning without auxiliary tasks across a suite of environments.