Generalization in Reinforcement Learning with Selective Noise Injection and Information Bottleneck
This addresses generalization issues in RL for broader application of agents, though it is incremental as it modifies existing supervised learning techniques.
The paper tackled the problem of overfitting in reinforcement learning (RL) by adapting regularization techniques like noise injection and the Information Bottleneck to RL, resulting in significant performance improvements, including outperforming state-of-the-art results on the Coinrun benchmark.
The ability for policies to generalize to new environments is key to the broad application of RL agents. A promising approach to prevent an agent's policy from overfitting to a limited set of training environments is to apply regularization techniques originally developed for supervised learning. However, there are stark differences between supervised learning and RL. We discuss those differences and propose modifications to existing regularization techniques in order to better adapt them to RL. In particular, we focus on regularization techniques relying on the injection of noise into the learned function, a family that includes some of the most widely used approaches such as Dropout and Batch Normalization. To adapt them to RL, we propose Selective Noise Injection (SNI), which maintains the regularizing effect the injected noise has, while mitigating the adverse effects it has on the gradient quality. Furthermore, we demonstrate that the Information Bottleneck (IB) is a particularly well suited regularization technique for RL as it is effective in the low-data regime encountered early on in training RL agents. Combining the IB with SNI, we significantly outperform current state of the art results, including on the recently proposed generalization benchmark Coinrun.