Dynamic Rank Adjustment in Diffusion Policies for Efficient and Flexible Training

Xiatao Sun; Shuo Yang; Yinxing Chen; Francis Fan; Yiyan (Edgar) Liang; Daniel Rakita

RSS 2025poster1 citations

Dynamic Rank Adjustment in Diffusion Policies for Efficient and Flexible Training

Xiatao Sun, Shuo Yang, Yinxing Chen, Francis Fan, Yiyan (Edgar) Liang, Daniel Rakita

Abstract

Diffusion policies trained via offline behavioral cloning have recently gained traction in robotic motion generation. While effective, these policies typically require a large number of trainable parameters. This model size affords powerful representations but also incurs high computational cost during training. Ideally, it would be beneficial to

BibTeX

@inproceedings{rss2025_dynamicrankadjus,
  title = {Dynamic Rank Adjustment in Diffusion Policies for Efficient and Flexible Training},
  author = {Xiatao Sun and Shuo Yang and Yinxing Chen and Francis Fan and Yiyan (Edgar) Liang and Daniel Rakita},
  booktitle = {RSS 2025},
  year = {2025}
}