COMPOL: A Unified Neural Operator Framework for Scalable Multi-Physics Simulations

Yifei Sun, Tao Wang, Junqi Qu, Yushun Dong, Hewei Tang, Shibo LiSubmission to ICLR 2026 · Withdrawn

Its authors withdrew this paper, so there is no full text to read or summarize.

Abstract

Multi-physics simulations play an essential role in accurately modeling complex interactions across diverse scientific and engineering domains. Although neural operators, especially the Fourier Neural Operator (FNO), have significantly improved computational efficiency, they often fail to effectively capture intricate correlations inherent in coupled physical processes. To address this limitation, we introduce COMPOL, a novel coupled multi-physics operator learning framework. COMPOL extends conventional operator architectures by incorporating sophisticated recurrent and attention-based aggregation mechanisms, effectively modeling interdependencies among interacting physical processes within latent feature spaces. Our approach is architecture-agnostic and seamlessly integrates into various neural operator frameworks that involve latent space transformations. Extensive experiments on diverse benchmarks—including biological reaction-diffusion systems, pattern-forming chemical reactions, multiphase geological flows, and thermo-hydro-mechanical processes — demonstrate that COMPOL consistently achieves superior predictive accuracy compared to state-of-the-art methods.
BibTeX
@misc{sun2025compol,
  author = {Yifei Sun and Tao Wang and Junqi Qu and Yushun Dong and Hewei Tang and Shibo Li},
  title = {{COMPOL: A Unified Neural Operator Framework for Scalable Multi-Physics Simulations}},
  year = {2025},
  eprint = {2501.17296},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2501.17296},
}

Loading related content...

Figure