Risk-Sensitive Generative Adversarial Imitation Learning

Abstract

We study risk-sensitive imitation learning where the agent's goal is to perform at least as well as the expert in terms of a risk profile. We first formulate our risk-sensitive imitation learning setting. We consider the generative adversarial approach to imitation learning (GAIL) and derive an optimization problem for our formulation, which we call it risk-sensitive GAIL (RS-GAIL). We then derive two different versions of our RS-GAIL optimization problem that aim at matching the risk profiles of the agent and the expert w.r.t. Jensen-Shannon (JS) divergence and Wasserstein distance, and develop risk-sensitive generative adversarial imitation learning algorithms based on these optimization problems. We evaluate the performance of our algorithms and compare them with GAIL and the risk-averse imitation learning (RAIL) algorithms in two MuJoCo and two OpenAI classical control tasks.

Cited by in corpus (7)

• Safe Imitation Learning via Fast Bayesian Reward Inference from Preferences
• Primal Wasserstein Imitation Learning
• Risk-Averse Offline Reinforcement Learning
• Deep Bayesian Reward Learning from Preferences
• Perceive, Attend, and Drive: Learning Spatial Attention for Safe Self-Driving
• Bayesian Robust Optimization for Imitation Learning
• Relational Mimic for Visual Adversarial Imitation Learning