ReRoGCRL: Representation-based Robustness in Goal-Conditioned Reinforcement Learning
This addresses the vulnerability of GCRL algorithms to adversarial attacks, which is a critical issue for deploying robust AI systems in real-world applications, representing an incremental advance in a specific domain.
The paper tackles the problem of adversarial robustness in Goal-Conditioned Reinforcement Learning (GCRL), proposing a new attack method and defense tactics that improve robustness against perturbations, with extensive experiments validating superior performance across multiple state-of-the-art GCRL algorithms.
While Goal-Conditioned Reinforcement Learning (GCRL) has gained attention, its algorithmic robustness against adversarial perturbations remains unexplored. The attacks and robust representation training methods that are designed for traditional RL become less effective when applied to GCRL. To address this challenge, we first propose the Semi-Contrastive Representation attack, a novel approach inspired by the adversarial contrastive attack. Unlike existing attacks in RL, it only necessitates information from the policy function and can be seamlessly implemented during deployment. Then, to mitigate the vulnerability of existing GCRL algorithms, we introduce Adversarial Representation Tactics, which combines Semi-Contrastive Adversarial Augmentation with Sensitivity-Aware Regularizer to improve the adversarial robustness of the underlying RL agent against various types of perturbations. Extensive experiments validate the superior performance of our attack and defence methods across multiple state-of-the-art GCRL algorithms. Our tool ReRoGCRL is available at https://github.com/TrustAI/ReRoGCRL.