AI Paper Atlas

Research map / Graph learning / GNN architectures

Graph Neural Networks Are Not Continuous Across Graph Resolutions

Christian Koke, Yuesong Shen, Abhishek Saroha, Marvin Eisenberger, Bastian Rieck, Michael M. Bronstein, Daniel Cremers

ICML 2026 · Graph learning › GNN architectures · approach: GNN generalization theory · 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
replace standard propagation with Laplace-transform (exponential) propagation, MPNNGCN-style MPNN · QM7 train High infer LowMAE (kcal/mol): 136.7 -> 17.4Improved
replace standard propagation with Laplace-transform (resolvent) propagation, MPNNGCN-style MPNN · QM7 train Low infer HighMAE (kcal/mol): 138.1 -> 17.6Improved
use spectral Laplace-transform (exponential) propagation instead of ChebNetChebNet · QM7 train High infer LowMAE (kcal/mol): 447.8 -> 15.9Improved
use Laplace-transform propagation instead of Lanczos multi-scale propagationLanczos · QM7 train Low infer HighMAE (kcal/mol): 658.6 -> 16.0Improved
use Laplace-transform MPNN (exp) instead of GCN for latent embeddingGCN · QM7 embedding difference High to Lowembedding difference: 20.3 -> 0.2Improved
use Laplace-transform MPNN (exp) instead of GIN for latent embeddingGIN · QM7 embedding difference Low to Highembedding difference: 2875.8 -> 0.2Improved
test GNNs cross-resolution instead of same-resolution (train High, infer Low)GCN · QM7MAE (kcal/mol): 63.6 -> 136.7Hurt
test GNNs cross-resolution instead of same-resolution (train High, infer Low)Lanczos · QM7MAE (kcal/mol): 9.9 -> 938.4Hurt
increase clique size k when expanding nodesstandard GNNs (GCN) · citation networks node classificationaccuracyHurt
use Laplace-transform propagation under increasing clique size kLaplace-transform GNNs · citation networks node classificationaccuracyImproved

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