AI Paper Atlas

Research map / Graph learning / GNN architectures

Fast and Distributed Equivariant Graph Neural Networks by Virtual Node Learning

Yuelin Zhang, Jiacheng Cen, Jiaqi Han, Wenbing Huang

TPAMI 2026 · Graph learning › GNN architectures · approach: Structural GNN variants · 6 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 virtual nodes (FastEGNN-<3, 0.00>)EGNN · Water-3DMSE (x1e-4): 6.00 -> 2.58Improved
add virtual nodes (FastEGNN-<3, 0.00>)EGNN · N-bodyMSE (x1e-2): 1.41 -> 1.06Improved
add MMD loss on virtual nodes (ablation)FastEGNN · Protein DynamicsMSE: 1.857 -> 1.821Improved
model virtual nodes as ordered set instead of global nodes (ablation)FastEGNN · Protein DynamicsMSE: 1.950 -> 1.821Improved
increase virtual nodes from 3 to 10 (ablation)FastEGNN · Water-3DMSE (x1e-4): 2.58 -> 2.66Hurt
drop all edges with 3 virtual nodes (ablation)FastEGNN · Water-3DMSE (x1e-4): 2.58 -> 3.60Mixed
add edge dropping 0.75 with 10 virtual nodes (ablation)FastEGNN · N-bodyRelative inference time: 1.36 -> 0.53Improved
add virtual nodes to TFNTFN · N-body (no edge dropping)MSE (x1e-2): 1.62 -> 3.42Hurt
add virtual nodes to SchNetSchNet · Protein DynamicsMSE: 2.56 -> 1.98Improved
distribute with virtual nodes (DistEGNN-3) over 8 devices instead of EGNNEGNN · Water-3D, 8 devicesMSE (x1e-4): 11.05 -> 3.27Improved
use METIS partitioning instead of random (ablation)DistEGNN-3 · Water-3DMSE (x1e-4): 3.46 -> 3.04No clear effect
dynamically enlarge cutoff radius (ablation)DistEGNN-1 · Water-3D, 8 devicesMSE (x1e-4): 3.85 -> 3.27Improved

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