Contrastive Variational Reinforcement Learning for Complex Observations

TL;DR

This paper introduces CVRL, a contrastive variational reinforcement learning method that effectively handles complex visual observations in robotics tasks by maximizing mutual information, leading to improved robustness and performance.

Contribution

CVRL is a novel model-based DRL approach that uses contrastive learning to better manage complex visual inputs without modeling the entire observation space.

Findings

Outperforms state-of-the-art methods on Natural Mujoco tasks

Achieves superior results on robot box-pushing with complex observations

Demonstrates robustness to dynamic shadows and complex visual features

Abstract

Deep reinforcement learning (DRL) has achieved significant success in various robot tasks: manipulation, navigation, etc. However, complex visual observations in natural environments remains a major challenge. This paper presents Contrastive Variational Reinforcement Learning (CVRL), a model-based method that tackles complex visual observations in DRL. CVRL learns a contrastive variational model by maximizing the mutual information between latent states and observations discriminatively, through contrastive learning. It avoids modeling the complex observation space unnecessarily, as the commonly used generative observation model often does, and is significantly more robust. CVRL achieves comparable performance with state-of-the-art model-based DRL methods on standard Mujoco tasks. It significantly outperforms them on Natural Mujoco tasks and a robot box-pushing task with complex observations, e.g., dynamic shadows. The CVRL code is available publicly at https://github.com/Yusufma03/CVRL.