AI Paper Atlas

Research map / Graph learning / GNN architectures

HGATSolver: A Heterogeneous Graph Attention Solver for Fluid–Structure Interaction

Qin-Yi Zhang, Hong Wang, Siyao Liu, Haichuan Lin, Linying Cao, Xiao-Hu Zhou, Chen Chen, Shuang-Yi Wang, Zeng-Guang Hou

AAAI 2026 · Graph learning › GNN architectures · approach: Physics-inspired GNNs · 5 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add physics-conditioned gating (PCGM) (ablation)HGATSolver · FI-Valve (fluid)relative L2 error (%): 3.276 -> 2.649Improved
add physics-conditioned gating (PCGM) (ablation)HGATSolver · SI-Vessel (solid)relative L2 error (%): 0.851 -> 0.652Improved
add IGBL uncertainty-weighted loss (ablation)HGATSolver · FI-Valve (solid)relative L2 error (%): 0.292 -> 0.250Improved
add time embedding (ablation)HGATSolver · FI-Valve (solid)relative L2 error (%): 0.315 -> 0.250Improved
add physics parameters as input (ablation)HGATSolver · SI-Vessel (solid)relative L2 error (%): 0.735 -> 0.652Improved
add learnable aggregation (ablation)HGATSolver · FI-Valve (fluid)relative L2 error (%): 3.054 -> 2.649Improved
IGBL vs fixed manual fluid:solid loss weights (ablation)HGATSolver · FI-Valve and SI-Vesselfluid and solid errorImproved
heterogeneous graph vs homogeneous GATHGATSolver · FI-Valve (fluid)relative L2 error (%): 4.125 -> 2.649Improved
HGATSolver vs AMG at Re=4000, 25 samplesHGATSolver · NS+EW Re=4000, 25 train samplesrelative L2 error (%): 0.490 -> 0.475No clear effect
HGATSolver few-shot on NS+EWHGATSolver · NS+EW Re=400, 5 train samplesrelative L2 error (%): 0.279 -> 0.237Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy