Abstract:Modern neural networks have shown promise for solving partial differential equations over surfaces, often by discretizing the surface as a mesh and learning with a mesh-aware graph neural network. However, graph neural networks suffer from oversmoothing, where a node's features become increasingly similar to those of its neighbors. Unitary graph convolutions, which are mathematically constrained to preserve smoothness, have been proposed to address this issue. Despite this, in many physical systems, such as diffusion processes, smoothness naturally increases and unitarity may be overconstraining. In this paper, we systematically study the smoothing effects of different GNNs for dynamics modeling and prove that unitary convolutions hurt performance for such tasks. We propose relaxed unitary convolutions that balance smoothness preservation with the natural smoothing required for physical systems. We also generalize unitary and relaxed unitary convolutions from graphs to meshes. In experiments on PDEs such as the heat and wave equations over complex meshes and on weather forecasting, we find that our method outperforms several strong baselines, including mesh-aware transformers and equivariant neural networks.
| Comments: | Ecstatic to share relaxed unitary mesh convolutions with the community :D! This version contains the camera ready for ICML 2026. Send me an email with your thoughts! I love getting mail :^) |
| Subjects: | Machine Learning (cs.LG); Symplectic Geometry (math.SG) |
| Cite as: | arXiv:2602.05352 [cs.LG] |
| (or arXiv:2602.05352v2 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2602.05352 arXiv-issued DOI via DataCite |
From: Edward Berman [view email]
[v1]
Thu, 5 Feb 2026 06:23:25 UTC (7,410 KB)
[v2]
Tue, 12 May 2026 03:14:17 UTC (7,984 KB)
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。