Research map / Graph learning / Graph self-supervised learning
Unsupervised Attributed Multiplex Network Embedding
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 |
|---|---|---|---|
| add attention over relation types | DMGI · ACM, DBLP, and Amazon | NMI: ACM: 0.687 -> 0.702; DBLP: 0.409 -> 0.554; Amazon: 0.425 -> 0.412 | Helped |
| add attention over relation types | DMGI · IMDB | Sim@5: 0.605 -> 0.586 | Hurt |
| remove the second term of consensus regularization (ablation) | DMGI attention · DBLP | NMI: 0.554 -> 0.448 | Hurt |
| use separate discriminator scoring matrices for relations (ablation) | DMGI attention · DBLP | NMI: 0.554 -> 0.076 | Hurt |
| remove node attributes and use adjacency as attributes (ablation) | DMGI attention · DBLP | NMI: 0.554 -> 0.053 | Hurt |
| shuffle adjacency rather than attributes to form negative samples (ablation) | DMGI attention · DBLP | Sim@5: 0.798 -> 0.504 | Hurt |
| add a semi-supervised classification module (ablation) | DMGI attention · DBLP | MaF1: 0.778 -> 0.791 | Helped |
| replace average readout with another readout function (ablation) | DMGI attention · DBLP | NMI: 0.554; Random sample 0.555; Maxpool 0.552; Linear projection 0.565; SAGPool 0.563 | No clear effect |
| remove relation PATAP | DMGI attention · DBLP, PAP+PPP versus all relations | NMI: 0.554 -> 0.565 | Helped |
Closest papers by meaning
- Deep Gaussian Embedding of Graphs: Unsupervised Inductive Learning via Ranking (ICLR 2018 · similarity 0.90)
- Beyond Redundancy: Information-aware Unsupervised Multiplex Graph Structure Learning (NeurIPS 2024 · similarity 0.89)
- Watch Your Step: Learning Node Embeddings via Graph Attention (NeurIPS 2018 · similarity 0.89)
- Multiplex Graph Representation Learning with Homophily and Consistency (AAAI 2025 · similarity 0.88)
- Node Identifiers: Compact, Discrete Representations for Efficient Graph Learning (ICLR 2025 · similarity 0.87)
- Scalable Attribute-Missing Graph Clustering via Neighborhood Differentiation (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.