Cross-attention-based bipartite graph neural network for coupled nodal and elemental field prediction in large-deformation sheet material forming Journal: Nuclear Physics B

Yingxue Zhao, Haoran Li, Haosu Zhou, Tobias Pfaff, Nan Li · 2026

Read Cross-attention-based bipartite graph neural network for coupled nodal and elemental field prediction in large-deformation sheet material forming Journal: Nuclear Physics B by Yingxue Zhao, Haoran Li, Haosu Zhou, Tobias Pfaff, Nan Li (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-2605-22845.txt.

Description

Zhao and four colleagues propose CAtt-BiGNN, a learned simulator for large-deformation sheet metal forming. It mirrors finite-element structure in its design. Nodes and elements become distinct graph entities joined by directed edges. Nodal displacement and elemental thinning are thus predicted on their native domains. Each field feeds back into the other during rollout. On dome cold-forming and corner hot-forming benchmarks, it beat node-centered baselines, a hypergraph model, and a graph Transformer on joint accuracy.

Catalog record

CREATORYingxue Zhao, Haoran Li, Haosu Zhou, Tobias Pfaff, Nan Li
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE112464
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.22845

Browse the full library · About the Tower of Babel