SCANO: Restoring Boundary-Condition Completeness in Sampled Meshes with a Sliced-Cross-Attention Neural Operator for 3D Thermal Fields
Abstract
Training neural operators on million-cell meshes can exceed GPU memory limits. Models are therefore trained on subsamples and deployed on whole meshes or selected regions. However, subsampling can distort boundary information, degrading accuracy. We ask when boundary conditioning under sampling stays consistent with whole-mesh training, and how to aggregate and inject boundary data. Representing the solution as an integral of boundary data against type-dependent measures, we show that subsampling distorts this conditioning in three ways: incomplete boundary coverage or a sample-dependent global state, face weighting by count rather than area, and compression of boundary inputs into a low-dimensional vector. This analysis yields four measurable consistency conditions checkable before training. We satisfy them with two backbone-agnostic modules. Boundary Quadrature Attention (BQA) aggregates all boundary faces, typed and area-weighted, into slice tokens that every query point reads at every layer. Anchored Global State (AGS) estimates the global state from anchors independent of the sampled points. Adding both to a scalable message-passing and slice-attention backbone yields SCANO. We evaluate the approach on controlled and public benchmarks and a data-centre thermal simulation dataset (DCTS) containing million-cell meshes. On both benchmarks, BQA improves every backbone, reducing whole-mesh error up to 15-fold. On all three datasets, AGS aligns regional with whole-mesh predictions, cutting regional error at least tenfold for most backbones on the public benchmark. Together, the modules nearly halve the best original backbone's whole-mesh error on both benchmarks. SCANO predicts a room-scale field in under one second, over 100 times faster than whole-room anchor self-attention.
Related
- On the gauge invariant boundary condition in open XXZ chains: SoV bases, elementary blocks and overlaps
- Bulk Boundary Condition for Semi-Infinite Surfaces in Density Functional Theory
- Solutions for neutron stars in General Relativity from a complexity structure scalar boundary condition
- Generalization of Rayleigh's high-frequency theory for the 2D Helmholtz equation in a half-space subject to a radiation condition at infinity and a Dirichlet condition on a 1D periodically-uneven boundary
- A note on the lower boundary condition for Ekman flows
BibTeX
@misc{li2026scano,
title = {{SCANO: Restoring Boundary-Condition Completeness in Sampled Meshes with a Sliced-Cross-Attention Neural Operator for 3D Thermal Fields}},
author = {Tianyu Li and Zhiwei Cao and Qingang Zhang and Ruihang Wang and Binyang Song and Yonggang Wen},
howpublished = {arXiv},
year = {2026},
month = nov,
doi = {10.48550/arxiv.2611.04603},
eprint = {2611.04603},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2611.04603},
}
RIS
TY - UNPB
TI - SCANO: Restoring Boundary-Condition Completeness in Sampled Meshes with a Sliced-Cross-Attention Neural Operator for 3D Thermal Fields
AU - Tianyu Li
AU - Zhiwei Cao
AU - Qingang Zhang
AU - Ruihang Wang
AU - Binyang Song
AU - Yonggang Wen
PY - 2026
DA - 2026/11/03
PB - arXiv
DO - 10.48550/arxiv.2611.04603
UR - https://arxiv.org/abs/2611.04603
AB - Training neural operators on million-cell meshes can exceed GPU memory limits. Models are therefore trained on subsamples and deployed on whole meshes or selected regions. However, subsampling can distort boundary information, degrading accuracy. We ask when boundary conditioning under sampling stays consistent with whole-mesh training, and how to aggregate and inject boundary data. Representing the solution as an integral of boundary data against type-dependent measures, we show that subsampling distorts this conditioning in three ways: incomplete boundary coverage or a sample-dependent global state, face weighting by count rather than area, and compression of boundary inputs into a low-dimensional vector. This analysis yields four measurable consistency conditions checkable before training. We satisfy them with two backbone-agnostic modules. Boundary Quadrature Attention (BQA) aggregates all boundary faces, typed and area-weighted, into slice tokens that every query point reads at every layer. Anchored Global State (AGS) estimates the global state from anchors independent of the sampled points. Adding both to a scalable message-passing and slice-attention backbone yields SCANO. We evaluate the approach on controlled and public benchmarks and a data-centre thermal simulation dataset (DCTS) containing million-cell meshes. On both benchmarks, BQA improves every backbone, reducing whole-mesh error up to 15-fold. On all three datasets, AGS aligns regional with whole-mesh predictions, cutting regional error at least tenfold for most backbones on the public benchmark. Together, the modules nearly halve the best original backbone's whole-mesh error on both benchmarks. SCANO predicts a room-scale field in under one second, over 100 times faster than whole-room anchor self-attention.
ER -