Diffusion Model Predictive Control
Abstract
We propose Diffusion Model Predictive Control (D-MPC), a novel MPC approach that learns a multi-step action proposal and a multi-step dynamics model, both using diffusion models, and combines them for use in online MPC. On the popular D4RL benchmark, we show performance that is significantly better than existing model-based offline planning methods using MPC (e.g. MBOP) and competitive with state-of-the-art (SOTA) model-based and model-free reinforcement learning methods. We additionally illustrate D-MPC’s ability to optimize novel reward functions at run time and adapt to novel dynamics, and highlight its advantages compared to existing diffusion-based planning baselines.
Cite
Text
Zhou et al. "Diffusion Model Predictive Control." Transactions on Machine Learning Research, 2025.Markdown
[Zhou et al. "Diffusion Model Predictive Control." Transactions on Machine Learning Research, 2025.](https://mlanthology.org/tmlr/2025/zhou2025tmlr-diffusion/)BibTeX
@article{zhou2025tmlr-diffusion,
title = {{Diffusion Model Predictive Control}},
author = {Zhou, Guangyao and Swaminathan, Sivaramakrishnan and Raju, Rajkumar Vasudeva and Guntupalli, J Swaroop and Lehrach, Wolfgang and Ortiz, Joseph and Dedieu, Antoine and Lazaro-Gredilla, Miguel and Murphy, Kevin Patrick},
journal = {Transactions on Machine Learning Research},
year = {2025},
url = {https://mlanthology.org/tmlr/2025/zhou2025tmlr-diffusion/}
}