Guanyu Chen

Zhejiang University

Hangzhou, China

Hi! My name is Guanyu Chen(陈冠宇). I am currently in my third year of PhD at College of IC, Zhejiang University, under the supervision of Professor Dong Ni. I completed my B.S. degree in Mathematics, Zhejiang University.

My research focuses on the modeling, simulation, and prediction of complex physical systems, with an emphasis on scientific computing for multiscale and multiphysics systems. Plasma systems are a key application area. My research directions include generative modeling and neural operator learning. I aim to develop efficient and generalizable deep learning methods to improve numerical simulation and prediction frameworks for complex dynamical systems.

Experience

BWTON (八维通科技有限公司)

Internship

Westlake UniversityJul 2026 – Present

Research Visit · AI for Scientific Simulation and Discovery Lab

Working with Prof. Tailin Wu on accelerating simulations of multiscale, multiphysics plasma systems.

publications

NeurIPS

NAGO: Noise-Aware Generative Operator via Flow Matching

Guanyu Chen, Shengze Xu, Pengwei Liu, Xingyu Ren, Pengkai Wang, Antonios Armaou, and 2 more authors

In Advances in Neural Information Processing Systems, 2026

Paper Bib Code

@inproceedings{chen2026nago,
  title = {{NAGO}: Noise-Aware Generative Operator via Flow Matching},
  author = {Chen, Guanyu and Xu, Shengze and Liu, Pengwei and Ren, Xingyu and Wang, Pengkai and Armaou, Antonios and Zeng, Tieyong and Ni, Dong},
  booktitle = {Advances in Neural Information Processing Systems},
  year = {2026},
  paper = {https://openreview.net/forum?id=opzL6gPOhn},
}

arXiv

Discovering Physical Directions in Weight Space: Composing Neural PDE Experts

Pengkai Wang, Pengwei Liu, Yuanyi Wang, Guanyu Chen, Xingyu Ren, Xiaolong Li, and 4 more authors

arXiv preprint arXiv:2605.14546, 2026

arXiv Bib

@article{wang2026physicaldirections,
  title = {Discovering Physical Directions in Weight Space: Composing Neural {PDE} Experts},
  author = {Wang, Pengkai and Liu, Pengwei and Wang, Yuanyi and Chen, Guanyu and Ren, Xingyu and Li, Xiaolong and Hao, Zhongkai and Kong, Yuting and Zhang, Qixin and Ni, Dong},
  journal = {arXiv preprint arXiv:2605.14546},
  year = {2026},
}

Commun. AI

An Efficient Graph-Transformer Operator for Learning Physical Dynamics with Manifolds Embedding

Pengwei Liu, Xingyu Ren, Pengkai Wang, Hangjie Yuan, Zhongkai Hao, Guanyu Chen, and 3 more authors

Communications AI & Computing, 2026

DOI arXiv Bib

@article{liu2026physgto,
  title = {An Efficient Graph-Transformer Operator for Learning Physical Dynamics with Manifolds Embedding},
  author = {Liu, Pengwei and Ren, Xingyu and Wang, Pengkai and Yuan, Hangjie and Hao, Zhongkai and Chen, Guanyu and Xu, Chao and Ni, Dong and Cai, Shengze},
  journal = {Communications AI \& Computing},
  volume = {1},
  pages = {9},
  year = {2026},
  doi = {10.1038/s44488-026-00008-w},
}

JGR

Reversible Deep Operator Network for Grid-Independent, Multi-Scale Magnetotelluric Inversion and Uncertainty Quantification

Heng Zhang, Guanyu Chen, Ziyu Tang, Bo Yang, and Yixian Xu

Journal of Geophysical Research: Solid Earth, 2026

DOI Bib

@article{zhang2026rdon,
  title = {Reversible Deep Operator Network for Grid-Independent, Multi-Scale Magnetotelluric Inversion and Uncertainty Quantification},
  author = {Zhang, Heng and Chen, Guanyu and Tang, Ziyu and Yang, Bo and Xu, Yixian},
  journal = {Journal of Geophysical Research: Solid Earth},
  volume = {131},
  number = {6},
  pages = {e2025JB033046},
  year = {2026},
  doi = {10.1029/2025JB033046},
}

NeurIPS

Uncertainty-Informed Meta Pseudo Labeling for Surrogate Modeling with Limited Labeled Data

Xingyu Ren, Pengwei Liu, Pengkai Wang, Guanyu Chen, Qinxin Wu, and Dong Ni

In Advances in Neural Information Processing Systems, 2025

DOI Bib HTML

@inproceedings{ren2025umpl,
  title = {Uncertainty-Informed Meta Pseudo Labeling for Surrogate Modeling with Limited Labeled Data},
  author = {Ren, Xingyu and Liu, Pengwei and Wang, Pengkai and Chen, Guanyu and Wu, Qinxin and Ni, Dong},
  booktitle = {Advances in Neural Information Processing Systems},
  volume = {38},
  year = {2025},
  doi = {10.52202/085713-2597},
}

arXiv

DGNN: A Neural PDE Solver Induced by Discontinuous Galerkin Methods

Guanyu Chen, Shengze Xu, Dong Ni, and Tieyong Zeng

arXiv preprint arXiv:2503.10021, 2025

arXiv Bib

@article{chen2025dgnn,
  title = {{DGNN}: A Neural {PDE} Solver Induced by Discontinuous Galerkin Methods},
  author = {Chen, Guanyu and Xu, Shengze and Ni, Dong and Zeng, Tieyong},
  journal = {arXiv preprint arXiv:2503.10021},
  year = {2025},
}