HyCOP: Hybrid Composition Operators for Interpretable Learning of PDEs

Jinpai Zhao, Nishant Panda, Yen Ting Lin, Eirik Valseth, Diane Oyen, Clint Dawson · 2026

Read HyCOP: Hybrid Composition Operators for Interpretable Learning of PDEs by Jinpai Zhao, Nishant Panda, Yen Ting Lin, Eirik Valseth, Diane Oyen, Clint Dawson (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-2605-00820.txt.

Description

Jinpai Zhao and five co-authors introduce HyCOP, a modular framework that learns solution operators for a family of partial differential equations. Instead of learning one large monolithic map, HyCOP composes simple modules for advection, diffusion, learned closures, and boundary handling. A policy chooses which module to apply and for how long, based on regime features and state statistics, forming short readable programs.

Modules can be classic solvers or learned parts, so stand-ins can be judged at any time without step-by-step rollout. Across varied benchmarks, HyCOP improves out-of-sample results by an order of magnitude over single-block operators and allows transfer by changing the module list. Theory separates composition error from module error as a check.

Catalog record

CREATORJinpai Zhao, Nishant Panda, Yen Ting Lin, Eirik Valseth, Diane Oyen, Clint Dawson
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE118 KB
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.00820

Browse the full library · About the Tower of Babel