Unlocking Pixels for Reinforcement Learning via Implicit Attention
This work is significant for researchers and practitioners in vision-based reinforcement learning, as it allows attention mechanisms to be applied to higher resolution inputs, potentially improving robustness to distractions and generalization.
This paper addresses the challenge of applying attention mechanisms to high-resolution visual inputs in reinforcement learning by adopting efficient attention algorithms. This enables attention-based controllers to scale to larger visual inputs, including individual pixels, and improves generalization.
There has recently been significant interest in training reinforcement learning (RL) agents in vision-based environments. This poses many challenges, such as high dimensionality and the potential for observational overfitting through spurious correlations. A promising approach to solve both of these problems is an attention bottleneck, which provides a simple and effective framework for learning high performing policies, even in the presence of distractions. However, due to poor scalability of attention architectures, these methods cannot be applied beyond low resolution visual inputs, using large patches (thus small attention matrices). In this paper we make use of new efficient attention algorithms, recently shown to be highly effective for Transformers, and demonstrate that these techniques can be successfully adopted for the RL setting. This allows our attention-based controllers to scale to larger visual inputs, and facilitate the use of smaller patches, even individual pixels, improving generalization. We show this on a range of tasks from the Distracting Control Suite to vision-based quadruped robots locomotion. We provide rigorous theoretical analysis of the proposed algorithm.