Reinforcement Learning with Unsupervised Auxiliary Tasks
This work addresses the challenge of slow learning and suboptimal performance in reinforcement learning for complex environments like video games, representing a strong specific advance rather than a broad paradigm shift.
The paper tackles the problem of improving deep reinforcement learning by training agents to simultaneously maximize multiple pseudo-reward functions alongside extrinsic rewards, resulting in significant performance gains such as 880% expert human performance on Atari and a 10x speedup in learning on Labyrinth tasks.
Deep reinforcement learning agents have achieved state-of-the-art results by directly maximising cumulative reward. However, environments contain a much wider variety of possible training signals. In this paper, we introduce an agent that also maximises many other pseudo-reward functions simultaneously by reinforcement learning. All of these tasks share a common representation that, like unsupervised learning, continues to develop in the absence of extrinsic rewards. We also introduce a novel mechanism for focusing this representation upon extrinsic rewards, so that learning can rapidly adapt to the most relevant aspects of the actual task. Our agent significantly outperforms the previous state-of-the-art on Atari, averaging 880\% expert human performance, and a challenging suite of first-person, three-dimensional \emph{Labyrinth} tasks leading to a mean speedup in learning of 10$\times$ and averaging 87\% expert human performance on Labyrinth.