Sparse Diffusion Policy: A Sparse, Reusable, and Flexible Policy for Robot Learning

TL;DR

The paper introduces Sparse Diffusion Policy (SDP), a novel, efficient, and flexible approach for robot learning that enables multitask and continual learning by selectively activating experts within a transformer-based diffusion framework.

Contribution

It proposes a sparse, reusable policy using Mixture of Experts in a diffusion model, allowing efficient multitask and continual learning without retraining the entire system.

Findings

SDP performs well in multitask scenarios with minimal active parameters.

SDP prevents catastrophic forgetting in continual learning.

SDP enables efficient transfer of skills across tasks.

Abstract

The increasing complexity of tasks in robotics demands efficient strategies for multitask and continual learning. Traditional models typically rely on a universal policy for all tasks, facing challenges such as high computational costs and catastrophic forgetting when learning new tasks. To address these issues, we introduce a sparse, reusable, and flexible policy, Sparse Diffusion Policy (SDP). By adopting Mixture of Experts (MoE) within a transformer-based diffusion policy, SDP selectively activates experts and skills, enabling efficient and task-specific learning without retraining the entire model. SDP not only reduces the burden of active parameters but also facilitates the seamless integration and reuse of experts across various tasks. Extensive experiments on diverse tasks in both simulations and real world show that SDP 1) excels in multitask scenarios with negligible increases in active parameters, 2) prevents forgetting in continual learning of new tasks, and 3) enables efficient task transfer, offering a promising solution for advanced robotic applications. Demos and codes can be found in https://forrest-110.github.io/sparse_diffusion_policy/.