DEAR: Disentangled Environment and Agent Representations for Reinforcement Learning without Reconstruction
This addresses the challenge of data-hungry visual RL for robotics, offering a novel approach to improve learning efficiency, though it is incremental in leveraging agent shape knowledge.
The paper tackled the problem of low sample efficiency in visual reinforcement learning for robotic control by proposing DEAR, which uses agent segmentation masks to learn disentangled representations without reconstruction, achieving state-of-the-art sample efficiency on benchmarks like Distracting DeepMind control suite and Franka Kitchen tasks with reduced parameters.
Reinforcement Learning (RL) algorithms can learn robotic control tasks from visual observations, but they often require a large amount of data, especially when the visual scene is complex and unstructured. In this paper, we explore how the agent's knowledge of its shape can improve the sample efficiency of visual RL methods. We propose a novel method, Disentangled Environment and Agent Representations (DEAR), that uses the segmentation mask of the agent as supervision to learn disentangled representations of the environment and the agent through feature separation constraints. Unlike previous approaches, DEAR does not require reconstruction of visual observations. These representations are then used as an auxiliary loss to the RL objective, encouraging the agent to focus on the relevant features of the environment. We evaluate DEAR on two challenging benchmarks: Distracting DeepMind control suite and Franka Kitchen manipulation tasks. Our findings demonstrate that DEAR surpasses state-of-the-art methods in sample efficiency, achieving comparable or superior performance with reduced parameters. Our results indicate that integrating agent knowledge into visual RL methods has the potential to enhance their learning efficiency and robustness.