AI Paper Atlas

Research map / Graph learning / GNN architectures

Breaking the Limits of Message Passing Graph Neural Networks

Muhammet Balcilar, Pierre Heroux, Benoit Gauzere, Pascal Vasseur, Sebastien Adam, Paul Honeine

ICML 2021 · Graph learning › GNN architectures · approach: Subgraph and distance-aware GNNs · 26 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
Use spectral-domain GNNML3 instead of 1-WL MPNNsgraph classification on graph8c · graph8cundistinguished graph pairs: 386 -> 0Improved
Use spectral-domain GNNML3 instead of Chebnetgraph discrimination · EXPundistinguished graph pairs: 71 -> 0Improved
Use GNNML3 instead of 1-WL MPNNsEXP classification · EXP-classifyaccuracy: 50\% -> 100\%Improved
Use GNNML3 instead of 1-WL MPNNsgraphlet counting · triangletest MSE: 2.06E-1 -> 4.44E-4Improved
Use GNNML3 instead of 1-WL MPNNsgraphlet counting · tailed-tritest MSE: 1.18E-1 -> 3.18E-4Improved
Use GNNML3 instead of 1-WL MPNNsgraphlet counting · 4-cyclestest MSE: 1.14E-1 -> 6.62E-4Improved
Use GNNML3 instead of Chebnetspectral graph classification · planar graph frequency classesaccuracy: 98.2\% -> 97.8\%Hurt
Use GNNML3 instead of PPGNZinc12K graph regression · Zinc12KMSE: 0.1589 $\pm$ 0.007 -> 0.1612 $\pm$ 0.006Hurt
Use GNNML3 instead of ChebnetMNIST-75 image graph classification · MNIST-75accuracy: 92.08\% $\pm$ 0.22 -> 91.98\% $\pm$ 0.18No clear effect
Use spectral-domain filters rather than standard MPNN propagationGNNML3 · Band-Pass node regressionR^2: 0.0629 -> 0.8189Improved
Use GNNML1 rather than GNNML3graphlet counting · triangletest MSE: 4.44E-4 -> 2.45E-1Hurt
Set Chebnet maximum eigenvalue to a fixed valueChebnet on EXP · EXPundistinguished graph pairsHurt

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