Adaptive Adversarial Training for Meta Reinforcement Learning

Abstract

Meta Reinforcement Learning (MRL) enables an agent to learn from a limited number of past trajectories and extrapolate to a new task. In this paper, we attempt to improve the robustness of MRL. We build upon model-agnostic meta-learning (MAML) and propose a novel method to generate adversarial samples for MRL by using Generative Adversarial Network (GAN). That allows us to enhance the robustness of MRL to adversal attacks by leveraging these attacks during meta training process.

References in corpus (10)

• Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks
• A Simple Neural Attentive Meta-Learner
• RL${}^2$: Fast Reinforcement Learning via Slow Reinforcement Learning
• Learning to reinforcement learn
• Delving into adversarial attacks on deep policies
• Meta Reinforcement Learning with Latent Variable Gaussian Processes
• Policy Teaching via Environment Poisoning: Training-time Adversarial Attacks against Reinforcement Learning
• Adversarial Meta-Learning
• Meta-Reinforcement Learning Robust to Distributional Shift via Model Identification and Experience Relabeling
• Meta Dropout: Learning to Perturb Features for Generalization