Research map / Graph learning / GNN architectures
Towards Quantifying Long-Range Interactions in Graph Machine Learning: a Large Graph Dataset and a Measurement
Huidong Liang, Haitz Sáez de Ocáriz Borde, Baskaran Sripathmanathan, Michael M. Bronstein, Xiaowen Dong
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 changed | Where | Measured | Effect |
|---|---|---|---|
| add graph message passing instead of node features alone (MLP to GraphSAGE) | MLP · City-Networks Paris | test accuracy: 25.5 -> 54.6 | Improved |
| add geographic coordinates to node features (ablation) | Exphormer · City-Networks Paris | test accuracy: 53.5 -> 55.1 | Improved |
| add geographic coordinates to node features (ablation) | GraphSAGE · City-Networks Shanghai | test accuracy: 66.5 -> 68.3 | Improved |
| add the other node features and graph to coordinates (ablation) | MLP coordinates only · City-Networks Paris | test accuracy: 12.5 -> 25.5 | Improved |
| increase number of message-passing layers from 2 to 16 | GNNs and GTs · City-Networks | test accuracy | Improved |
| increase number of message-passing layers | GNNs and GTs · Cora and ogbn-arxiv | test accuracy | Hurt |
| increase number of message-passing layers | GNNs and GTs · Amazon-Ratings and PascalVOC-SP | test accuracy | No clear effect |
| raise label locality k from 8 to 16 or 32 (ablation) | GCN · City-Networks Paris | layers needed to plateau: saturates at num_layers 8 -> plateaus at num_layers 16 | Improved |
| raise layers from 16 to 32 at k=32 (ablation) | GCN · City-Networks Paris | test accuracy | No clear effect |
| use Exphormer instead of GCN | GCN · City-Networks Shanghai | test accuracy: 62.1 -> 70.2 | Improved |
| add GraphGPS global attention on London | GraphGPS · City-Networks London | test accuracy: OOM | Hurt |
Closest papers by meaning
- On Measuring Long-Range Interactions in Graph Neural Networks (ICML 2025 · similarity 0.93)
- Learning Long Range Dependencies on Graphs via Random Walks (ICLR 2025 · similarity 0.89)
- Sketch-Augmented Features Improve Learning Long-Range Dependencies in Graph Neural Networks (NeurIPS 2025 · similarity 0.88)
- Get a GRIP, this will be a long TRIP: A Quantifiable Long-Range Framework for Verifying Over-squashing (NeurIPS 2026 · similarity 0.88)
- Long Range Propagation on Continuous-Time Dynamic Graphs (ICML 2024 · similarity 0.87)
- Long-range Meta-path Search on Large-scale Heterogeneous Graphs (NeurIPS 2024 · similarity 0.87)
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.
