DRIBO: Robust Deep Reinforcement Learning via Multi-View Information Bottleneck

Deep reinforcement learning (DRL) agents are often sensitive to visual changes that were unseen in their training environments. To address this problem, we leverage the sequential nature of RL to learn robust representations that encode only task-relevant information from observations based on the unsupervised multi-view setting. Specifically, we introduce an auxiliary objective based on the multi-view in-formation bottleneck (MIB) principle which quantifies the amount of task-irrelevant information and encourages learning representations that are both predictive of the future and less sensitive to task-irrelevant distractions. This enables us to train high-performance policies that are robust to visual distractions and can generalize to unseen environments. We demonstrate that our approach can achieve SOTA performance on diverse visual control tasks on the DeepMind Control Suite, even when the background is replaced with natural videos. In addition, we show that our approach outperforms well-established baselines for generalization to unseen environments on the Procgen benchmark. Our code is open-sourced and available at https://github.com/JmfanBU/DRIBO.