Consistency Training with Physical Constraints

We propose a physics-aware Consistency Training (CT) method that accelerates sampling in Diffusion Models with physical constraints. Our approach leverages a two-stage strategy: (1) learning the noise-to-data mapping via CT, and (2) incorporating physics constraints as a regularizer. Experiments on toy examples show that our method generates samples in a single step while adhering to the imposed constraints. This approach has the potential to efficiently solve partial differential equations (PDEs) using deep generative modeling.

@article{chang2025_2502.07636,
  title={ Consistency Training with Physical Constraints },
  author={ Che-Chia Chang and Chen-Yang Dai and Te-Sheng Lin and Ming-Chih Lai and Chieh-Hsin Lai },
  journal={arXiv preprint arXiv:2502.07636},
  year={ 2025 }
}