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

Yixiao Wang, Yifei Zhang, Mingxiao Huo, Thomas Tian, Xiang Zhang, Yichen Xie, Chenfeng Xu, Pengliang Ji, Wei Zhan, Mingyu Ding, Masayoshi Tomizuka
Proceedings of The 8th Conference on Robot Learning, PMLR 270:649-665, 2025.

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 task-specific learning without retraining the entire model. It 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 simulators and the real world show that SDP 1) excels in multitask scenarios with negligible increases in active parameters, 2) prevents forgetting in continual learning new tasks, and 3) enables efficient task transfer, offering a promising solution for advanced robotic applications. More demos and codes can be found on our https://anonymous.4open.science/w/sparse_diffusion_policy-24E7/.

Cite this Paper

BibTeX
@InProceedings{pmlr-v270-wang25c, title = {Sparse Diffusion Policy: A Sparse, Reusable, and Flexible Policy for Robot Learning}, author = {Wang, Yixiao and Zhang, Yifei and Huo, Mingxiao and Tian, Thomas and Zhang, Xiang and Xie, Yichen and Xu, Chenfeng and Ji, Pengliang and Zhan, Wei and Ding, Mingyu and Tomizuka, Masayoshi}, booktitle = {Proceedings of The 8th Conference on Robot Learning}, pages = {649--665}, year = {2025}, editor = {Agrawal, Pulkit and Kroemer, Oliver and Burgard, Wolfram}, volume = {270}, series = {Proceedings of Machine Learning Research}, month = {06--09 Nov}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v270/main/assets/wang25c/wang25c.pdf}, url = {https://proceedings.mlr.press/v270/wang25c.html}, 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 task-specific learning without retraining the entire model. It 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 simulators and the real world show that SDP 1) excels in multitask scenarios with negligible increases in active parameters, 2) prevents forgetting in continual learning new tasks, and 3) enables efficient task transfer, offering a promising solution for advanced robotic applications. More demos and codes can be found on our https://anonymous.4open.science/w/sparse_diffusion_policy-24E7/.} }
Endnote
%0 Conference Paper %T Sparse Diffusion Policy: A Sparse, Reusable, and Flexible Policy for Robot Learning %A Yixiao Wang %A Yifei Zhang %A Mingxiao Huo %A Thomas Tian %A Xiang Zhang %A Yichen Xie %A Chenfeng Xu %A Pengliang Ji %A Wei Zhan %A Mingyu Ding %A Masayoshi Tomizuka %B Proceedings of The 8th Conference on Robot Learning %C Proceedings of Machine Learning Research %D 2025 %E Pulkit Agrawal %E Oliver Kroemer %E Wolfram Burgard %F pmlr-v270-wang25c %I PMLR %P 649--665 %U https://proceedings.mlr.press/v270/wang25c.html %V 270 %X 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 task-specific learning without retraining the entire model. It 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 simulators and the real world show that SDP 1) excels in multitask scenarios with negligible increases in active parameters, 2) prevents forgetting in continual learning new tasks, and 3) enables efficient task transfer, offering a promising solution for advanced robotic applications. More demos and codes can be found on our https://anonymous.4open.science/w/sparse_diffusion_policy-24E7/.
APA
Wang, Y., Zhang, Y., Huo, M., Tian, T., Zhang, X., Xie, Y., Xu, C., Ji, P., Zhan, W., Ding, M. & Tomizuka, M.. (2025). Sparse Diffusion Policy: A Sparse, Reusable, and Flexible Policy for Robot Learning. Proceedings of The 8th Conference on Robot Learning, in Proceedings of Machine Learning Research 270:649-665 Available from https://proceedings.mlr.press/v270/wang25c.html.

Related Material