From Basis to Basis: Gaussian Particle Representation for Interpretable PDE Operators
This work addresses interpretability and efficiency issues in PDE learning for fluids, offering a domain-specific solution with incremental improvements in method design.
The paper tackles the problem of learning PDE dynamics for fluids with neural operators and Transformer-based models, which lack interpretability and struggle with localized structures while having high computational cost, by proposing a Gaussian basis representation and a Gaussian Particle Operator that achieves near-linear complexity and competitive accuracy on benchmarks.
Learning PDE dynamics for fluids increasingly relies on neural operators and Transformer-based models, yet these approaches often lack interpretability and struggle with localized, high-frequency structures while incurring quadratic cost in spatial samples. We propose representing fields with a Gaussian basis, where learned atoms carry explicit geometry (centers, anisotropic scales, weights) and form a compact, mesh-agnostic, directly visualizable state. Building on this representation, we introduce a Gaussian Particle Operator that acts in modal space: learned Gaussian modal windows perform a Petrov-Galerkin measurement, and PG Gaussian Attention enables global cross-scale coupling. This basis-to-basis design is resolution-agnostic and achieves near-linear complexity in N for a fixed modal budget, supporting irregular geometries and seamless 2D-to-3D extension. On standard PDE benchmarks and real datasets, our method attains state-of-the-art competitive accuracy while providing intrinsic interpretability.