# PCFM

**Repository Path**: cnwhuwyc/PCFM

## Basic Information

- **Project Name**: PCFM
- **Description**: No description available
- **Primary Language**: Unknown
- **License**: MIT
- **Default Branch**: main
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2026-01-05
- **Last Updated**: 2026-01-05

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# PCFM: Physics-Constrained Flow Matching

[![arXiv](https://img.shields.io/badge/arXiv-2506.04171-b31b1b.svg)](https://arxiv.org/abs/2506.04171)
 • 
[Project Page](https://caipengfei.me/pcfm)
 • 
[Julia Version](https://github.com/utkarsh530/PCFM.jl)

  
[**Utkarsh\***](https://www.linkedin.com/in/utkarsh530/) •
[**Pengfei Cai\***](https://www.linkedin.com/in/pengfei-cai/) •
Alan Edelman •
Rafael Gómez-Bombarelli •
Christopher Rackauckas  

<em>*To appear at <a href="https://neurips.cc/virtual/2025/poster/117071">NeurIPS 2025</a>.*</em>
</div>

This repo implements *Physics-Constrained Flow Matching (PCFM)* -- a framework that enforces physical constraints during sampling of flow-based generative models. 

<img src="assets/pcfm_figure.png" width="67%" alt="PCFM summary figure">


## License

PCFM is released under the **MIT License**.

This repository includes components derived from [amazon-science/ECI-sampling](https://github.com/amazon-science/ECI-sampling) licensed under the **Apache License 2.0**. See [`LICENSE`](./LICENSE), [`LICENSE-APACHE-2.0`](./LICENSE-APACHE-2.0), and [`NOTICE`](./NOTICE) for details.



## Citation

If you use this repository, please cite:

```bibtex
@article{PCFM2025,
  title={Physics-Constrained Flow Matching: Sampling Generative Models with Hard Constraints},
  author={Utkarsh, Utkarsh and Cai, Pengfei and Edelman, Alan and Gomez-Bombarelli, Rafael and Rackauckas, Christopher Vincent},
  journal={arXiv preprint arXiv:2506.04171},
  year={2025}
}