Research map / Graph learning / Graph algorithms and clustering
A Provable Framework of Learning Graph Embeddings via Summarization
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 |
|---|---|---|---|
| Learn embeddings on a configuration-based summary graph and restore with the derived matrix | DeepWalk · Cora, Citeseer, Pubmed, Flickr node classification | accuracy: 72.11 -> 74.67; 46.24 -> 49.08; 72.35 -> 74.05; 53.19 -> 53.36 | Mixed |
| Learn embeddings on a configuration-based summary graph and restore with the derived matrix | LINE · Cora, Citeseer, Pubmed, Flickr node classification | accuracy: 68.27 -> 68.79; 41.36 -> 45.15; 68.51 -> 71.40; 51.47 -> 52.42 | Mixed |
| Use the theoretically derived restoration rather than empirical restoration (ablation) | GELSUMM-GCN · Cora, Citeseer, Pubmed, Flickr, Reddit | accuracy: 80.64 -> 80.90; 69.94 -> 70.46; 76.88 -> 77.38; 48.75 -> 50.15; 92.55 -> 93.62 | Helped |
| Summarize the graph and use GELSUMM-GCN at r=0.6 | GCN · Cora, Citeseer, Pubmed, Flickr, Reddit | accuracy: 80.20 -> 80.90; 69.70 -> 70.46; 78.02 -> 77.38; 52.89 -> 50.15; 94.55 -> 93.62 | Mixed |
| Reduce the summary graph ratio from r=0.6 to r=0.2 (ablation) | GELSUMM-GCN · Cora, Pubmed, Flickr, Reddit node classification | accuracy: 80.90 -> 75.22; 77.38 -> 73.92; 50.15 -> 49.51; 93.62 -> 91.90 | Hurt |
| Reduce the summary graph ratio from r=0.6 to r=0.2 (ablation) | GELSUMM-GCN · Citeseer node classification | accuracy: 70.46 -> 67.64 | Hurt |
Closest papers by meaning
- GraphZoom: A Multi-level Spectral Approach for Accurate and Scalable Graph Embedding (ICLR 2020 · similarity 0.91)
- Watch Your Step: Learning Node Embeddings via Graph Attention (NeurIPS 2018 · similarity 0.89)
- DeepWalking Backwards: From Embeddings Back to Graphs (ICML 2021 · similarity 0.88)
- Deep Gaussian Embedding of Graphs: Unsupervised Inductive Learning via Ranking (ICLR 2018 · similarity 0.88)
- Fisher Information Embedding for Node and Graph Learning (ICML 2023 · similarity 0.87)
- Learning Representations for Hierarchies with Minimal Support (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.