From Basis to Basis: Gaussian Particle Representation for Interpretable PDE Operators

Zhihao Li, Yu Feng, Zhilu Lai, Wei Wang · 2026

Read From Basis to Basis: Gaussian Particle Representation for Interpretable PDE Operators by Zhihao Li, Yu Feng, Zhilu Lai, Wei Wang (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-21551.txt.

Description

Zhihao Li and three co-authors offer a Gaussian particle form for PDE operators that can be read. Learning fluid motion now leans on neural operators and Transformer models. But those tools often hide how they work, struggle with small, fast-changing patterns, and cost grows with sample counts squared.

Their design uses a Gaussian basis. Learned atoms carry clear geometry: centers, directional scales, and weights. The result is a compact state that needs no mesh and can be drawn directly. A Gaussian Particle Operator then acts in modal space. Gaussian modal windows take a Petrov-Galerkin measurement, and PG Gaussian Attention links scales across the whole field. The basis-to-basis design works at any resolution and costs near-linear time for a fixed modal budget. It handles rough geometries and extends smoothly from 2D to 3D, with leading accuracy on PDE tests and real data.

Catalog record

CREATORZhihao Li, Yu Feng, Zhilu Lai, Wei Wang
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE52 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2602-21551

Browse the full library · About the Tower of Babel