Research map / Graph learning / GNN architectures
Position-aware Graph Neural Networks
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Use shared anchor-set distances to encode node positions | P-GNN · Grid | ROC AUC: 0.566 ± 0.052 -> 0.940 ± 0.027 | Helped |
| Use shared anchor-set distances to encode node positions | P-GNN · Communities (pairwise node classification) | ROC AUC: 0.620 ± 0.022 -> 1.0 ± 0.001 | Helped |
| Use shared anchor-set distances to encode node positions | P-GNN · Email | ROC AUC: 0.545 ± 0.012 -> 0.640 ± 0.029 | Helped |
| Use shared anchor-set distances to encode node positions | P-GNN · Protein | ROC AUC: 0.528 ± 0.011 -> 0.631 ± 0.175 | Helped |
| Use truncated 2-hop rather than exact shortest-path distance (ablation) | P-GNN-E-2L · Grid | ROC AUC: 0.940 ± 0.027 -> 0.694 ± 0.066 | Hurt |
| Use truncated 2-hop rather than exact shortest-path distance (ablation) | P-GNN-E-2L · Communities | ROC AUC: 0.985 ± 0.008 -> 0.991 ± 0.003 | Helped |
| Use truncated 2-hop rather than exact shortest-path distance (ablation) | P-GNN-E-2L · PPI | ROC AUC: 0.808 ± 0.003 -> 0.805 ± 0.003 | No clear effect |
| Use truncated 2-hop rather than exact shortest-path distance (ablation) | P-GNN-E-2L · Protein | ROC AUC: 0.631 ± 0.175 -> 0.729 ± 0.176 | Helped |
| Increase depth from 1 layer to 2 layers (ablation) | P-GNN-E · Grid | ROC AUC: 0.879 ± 0.039 -> 0.940 ± 0.027 | Helped |
| Increase depth from 1 layer to 2 layers (ablation) | P-GNN-E · Communities | ROC AUC: 0.985 ± 0.005 -> 0.985 ± 0.008 | No clear effect |
| Increase depth from 1 layer to 2 layers (ablation) | P-GNN-E · Protein | ROC AUC: 0.507 ± 0.006 -> 0.631 ± 0.175 | Mixed |
| Simplify anchor-set aggregation to use only the closest node in each set (ablation) | P-GNN · Experiments (overall) | computational complexity: O(n^2 log^2n) -> O(n log^2n) | Helped |
Closest papers by meaning
- Learning Efficient Positional Encodings with Graph Neural Networks (ICLR 2025 · similarity 0.90)
- GPEN: Global Position Encoding Network for Enhanced Subgraph Representation Learning (ICML 2025 · similarity 0.89)
- N2GON: Neural Networks for Graph-of-Net with Position Awareness (ICML 2025 · similarity 0.89)
- GNNs Meet Sequence Models Along the Shortest-Path: an Expressive Method for Link Prediction (NeurIPS 2026 · similarity 0.88)
- Graph Positional and Structural Encoder (ICML 2024 · similarity 0.88)
- Towards characterizing the value of edge embeddings in Graph Neural Networks (ICML 2025 · 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.