Xinghao Dong（董星浩）

PhD Candidate | MSCS at the University of Wisconsin-Madison.
The Artificial Intelligence for Modeling and Simulation Lab.

Bachelor of Science,
Applied Mathematics + Specialization in Computing,
University of California, Los Angeles.

The creative principle resides in mathematics. - Albert Einstein

Research: My goal is to make foundational advancements in the theory, methods, and algorithms of generative AI and SciML. My current research focuses on (score-based) diffusion models with applications in complex dynamical systems that are multi-scale, multi-physics, and chaotic in nature. Additionally, my work extends to significant involvement and broad interests in data-driven modeling, computational fluid dynamics, and optimization.

Previously: I obtained my Bachelor’s degree in Applied Mathematics and Computing from University of California, Los Angeles (UCLA), where I worked on proving the Morrey’s Conjecture by providing several numerical examples.

news

Sep 14, 2024	I’ll be presenting at the APS DFD Annual Meeting in Salt Lake City, UT, on my work, Learning Stochastic Closures via Conditional Diffusion Model and Neural Operator. This project was selected for DFD-Interact, which features only the most exciting submissions. I’ll be at Session C02: Interact: Machine Learning in Fluids from 10:50 AM - 12:50 PM on November 24. Stop by and say hi!
Sep 06, 2024	I have successfully passed my qualification examinations and advanced to candidacy.

latest posts

Apr 27, 2024	Simulation Videos of the Surrogate Models

selected publications

arXiv
Data-Driven Stochastic Closure Modeling via Conditional Diffusion Model and Neural Operator

Xinghao Dong, Chuanqi Chen, and Jin-Long Wu

Abstract Bib PDF

Closure models are widely used in simulating complex multiscale dynamical systems such as turbulence and the earth system, for which direct numerical simulation that resolves all scales is often too expensive. For those systems without a clear scale separation, deterministic and local closure models often lack enough generalization capability, which limits their performance in many real-world applications. In this work, we propose a data-driven modeling framework for constructing stochastic and non-local closure models via conditional diffusion model and neural operator. Specifically, the Fourier neural operator is incorporated into a score-based diffusion model, which serves as a data-driven stochastic closure model for complex dynamical systems governed by partial differential equations (PDEs). We also demonstrate how accelerated sampling methods can improve the efficiency of the data-driven stochastic closure model. The results show that the proposed methodology provides a systematic approach via generative machine learning techniques to construct data-driven stochastic closure models for multiscale dynamical systems with continuous spatiotemporal fields.
@article{dong2024data, title = {Data-Driven Stochastic Closure Modeling via Conditional Diffusion Model and Neural Operator}, author = {Dong, Xinghao and Chen, Chuanqi and Wu, Jin-Long}, year = {2024}, }