Adaptive Coordinate Transforms for Neural Operators

Chaoyu Liu, Zhonghao Li, Gaohang Chen, Zakhar Shumaylov, Zhongying Deng, Qian Zhang, Zhonghua Qiao, Carola-Bibiane Schönlieb · 2026

Read Adaptive Coordinate Transforms for Neural Operators by Chaoyu Liu, Zhonghao Li, Gaohang Chen, Zakhar Shumaylov, Zhongying Deng, Qian Zhang, Zhonghua Qiao, Carola-Bibiane Schönlieb (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-2605-06203.txt.

Description

Chaoyu Liu and seven co-authors add adaptive coordinate transforms to neural operators that solve partial differential equations. Most such models work on fixed grid coordinates, while the physical structures they track move and sharpen over time. That mismatch makes the learned mapping needlessly nonlocal and biases results toward smoothness near sharp transitions, the authors argue.

Their Adaptive Coordinate Transform block learns a new coordinate system inside the operator chain. The same input feature is then resampled under the new coordinates, keeping the signal while changing its spatial layout. The block plugs into existing designs without redesign. Across diverse equation benchmarks and several operator designs, the module gave consistent and significant gains in predictive accuracy.

Catalog record

CREATORChaoyu Liu, Zhonghao Li, Gaohang Chen, Zakhar Shumaylov, Zhongying Deng, Qian Zhang, Zhonghua Qiao, Carola-Bibiane Schönlieb
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE94 KB
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.06203

Browse the full library · About the Tower of Babel