AI Paper Atlas

Research map / Graph learning / GNN architectures

S$^3$GNN: Efficient Global Mixing and Local Message Passing for Long-Range Graph Learning

Dai Shi, Luke Thompson, Linhan Luo, Lequan Lin, Andi Han, Junbin Gao, José Miguel Hernández-Lobato

ICML 2026 · Graph learning › GNN architectures · approach: Graph rewiring and scalable message passing · 0 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 spatial propagation and linear feature transform to low-rank global mixing (S3GNN vs diagonal filter) (ablation)diagonal spectral filtering dynamic (Eq. 7) · peptides-structMAE: 0.6824 -> 0.2429Improved
replace S2GNN spectral dynamics with S3GNNS2GNN · Graph Property Prediction Eccentricitylog10(MSE): -1.4622 -> -2.4801Improved
replace S2GNN with S3GNN (no eigendecomposition, no PE)S2GNN · peptides-funcAP: 72.75 -> 73.20Improved
replace S2GNN with S3GNNS2GNN · peptides-structMAE: 0.2487 -> 0.2429No clear effect
replace Stable-ChebNet with S3GNNStable-ChebNet · ogbn-arxivaccuracy: 75.42 -> 76.38Improved
replace Stable-ChebNet with S3GNN on Barbell graph (N=50)Stable-ChebNet · Barbell graph MSEMSE: 0.17 -> 0.005Improved
use S3GNN in the reasoning GNN of KGQAReasoning GNN · WebQSP multi-hop (no RAG)F1: 58.8 -> 69.6Improved
use S3GNN in reasoning GNN with RAGReasoning GNN-RAG · CWQ multi-entityF1: 64.8 -> 64.9No clear effect
replace GCN backbone with S3GNN in topological Schrodinger bridge matchingTSB-BM · Brain signals2-Wasserstein distance: 7.51 -> 7.25Improved
replace processor blocks of MeshGraphNet with S3GNN layersMGN · Cylinder Flowmin test loss: 0.173 -> 0.102Improved
use antisymmetric constraint on W versus free W (ablation)S3GNN · Barbell graphMSEImproved
use orthogonal constraint on W instead of antisymmetric (ablation)S3GNN · Barbell graphMSEHurt

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