Research map / Graph learning / Graph self-supervised learning
On the Equivalence between Positional Node Embeddings and Structural Graph Representations
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 |
|---|---|---|---|
| Increase CGNN embedding samples from 1 to 20 (ablation) | CGNN structural representations · Cora node classification | Micro F1: 0.468(0.026) -> 0.726(0.024) | Helped |
| Increase SVD embedding samples from 1 to 20 (ablation) | MC-SVD structural representations · Pubmed node classification | Micro F1: 0.825(0.007) -> 0.855(0.010) | Helped |
| Increase SVD embedding samples from 1 to 20 (ablation) | MC-SVD joint representations · Pubmed link prediction | Micro F1: 0.818(0.032) -> 0.868(0.029) | Helped |
| Increase SVD embedding samples from 1 to 20 (ablation) | MC-SVD joint representations · Cora link prediction | Micro F1: 0.583(0.020) -> 0.580(0.021) | No clear effect |
| Increase SVD embedding samples from 1 to 20 (ablation) | MC-SVD joint representations · PPI link prediction | Micro F1: 0.755(0.008) -> 0.762(0.010) | No clear effect |
| Increase CGNN embedding samples from 1 to 20 (ablation) | CGNN joint representations · Cora link prediction | Micro F1: 0.682(0.026) -> 0.712(0.041) | Helped |
| Increase SVD embedding samples from 1 to 20 (ablation) | MC-SVD joint representations · Pubmed triad prediction | Micro F1: 0.518(0.065) -> 0.546(0.029) | Helped |
| Increase SVD embedding samples from 1 to 20 (ablation) | MC-SVD joint representations · Cora triad prediction | Micro F1: 0.304(0.034) -> 0.300(0.033) | No clear effect |
| Use SVD until convergence rather than one optimization step (ablation) | MC-SVD · Pubmed link prediction | Micro F1: 0.818(0.032) -> 0.807(0.024) | Hurt |
| Use node representations alone instead of joint representations | GIN and RP-GIN · Link prediction | Micro F1 | Hurt |
| Use 1-2-3 GNN as the tuple-representation approach | 1-2-3 GNN · Node classification, link prediction, and triad prediction | Micro F1 | Hurt |
Closest papers by meaning
- Learning Representations for Hierarchies with Minimal Support (NeurIPS 2024 · similarity 0.90)
- Dissecting the Failure of Invariant Learning on Graphs (NeurIPS 2024 · similarity 0.88)
- Learning Efficient Positional Encodings with Graph Neural Networks (ICLR 2025 · similarity 0.88)
- Graph Positional and Structural Encoder (ICML 2024 · similarity 0.88)
- Transitivity-Preserving Graph Representation Learning for Bridging Local Connectivity and Role-Based Similarity (AAAI 2024 · similarity 0.88)
- Node Identifiers: Compact, Discrete Representations for Efficient Graph Learning (ICLR 2025 · similarity 0.88)
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.