AI Paper Atlas

Research map / Graph learning / GNN architectures

Message Passing on the Edge: Towards Scalable and Expressive GNNs

Pablo Barcelo, Fabian Jogl, Alexander Kozachinskiy, Matthias Lanzinger, Stefan Neumann, Cristobal Rojas

ICML 2026 · Graph learning › GNN architectures · approach: Expressive equivariant 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
edge-based message passing over edges and triangles instead of node-based MPNNEB-GNN · CSLaccuracy: 10% -> 20%Improved
add triangle counts as node features to MPNN (ablation)GIN · BRECdistinguishable pairs: 0 -> 0 on Basic, Regular, Extension, CFINo clear effect
use EB-GNN instead of NC-GNNedge-based GNN · BREC Extensiondistinguishable pairs: 0 -> 60Improved
use EB-GNN instead of NC-GNNedge-based GNN · BREC Regular and CFIdistinguishable pairs: Regular 48 vs 48; CFI 0 vs 0No clear effect
EB-GNN vs D-MPNN on QMD edge-level tasksEB-GNN · QMD edge-level (4 tasks)MAE: Bond Index 6.65 -> 4.42Improved
EB-GNN vs D-MPNN on QMD graph-level tasksEB-GNN · QMD graph-level (4 tasks)MAE: IP 4.29 -> 5.49; Dipole Moment 4.59 -> 4.33Mixed
EB-GNN vs 5-lGIN on QM9EB-GNN · QM9 (12 tasks)normalized MAE: better on 4 tasks, comparable on 7Mixed
EB-GNN vs MPNN, GPS, HyMN on large graphsEB-GNN · MalNet-Tinyaccuracy: 91.10 -> 93.30Improved
EB-GNN instead of 5-lGNN for runtimeEB-GNN · QM9runtime: approximately 4 times fasterImproved
EB-GNN on CFI graphsEB-GNN · BREC CFIdistinguishable pairs: 0No clear effect

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