Research map / Graph learning / GNN architectures
Beyond Homophily: Reconstructing Structure for Graph-agnostic Clustering
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 DGCN instead of the strongest listed baseline | Node clustering on heterophilic graphs · Texas, Cornell, Wisconsin, Washington, Twitch, and Squirrel | ACC | Helped |
| Use DGCN instead of the strongest listed baseline | Node clustering on heterophilic graphs · Texas, Cornell, Wisconsin, Washington, Twitch, and Squirrel | NMI | Helped |
| Use DGCN instead of the strongest listed baseline | Node clustering on homophilic graphs · Cora, Citeseer, ACM, AMAP, and EAT | ACC | Mixed |
| Use DGCN instead of the strongest listed baseline | Node clustering on homophilic graphs · Citeseer, ACM, and EAT | NMI | Mixed |
| Add high-frequency filtering to low-frequency filtering (ablation) | DGCN mixed filter · Cora and Texas | clustering accuracy | Helped |
| Apply graph filtering rather than no filtering (ablation) | DGCN · Cora and Texas | clustering accuracy | Helped |
| Use reconstructed graphs instead of the raw graph for filtering (ablation) | DGCN · Cora and Texas | clustering accuracy | Helped |
| Vary the SCE sharpening parameter from β=1 to β=2 (ablation) | DGCN · Heterophilic and homophilic benchmarks | ACC and NMI | Mixed |
| Construct a homophilic graph S from the original graph (ablation) | Graph reconstruction · Most of the 11 datasets | homophily ratio | Helped |
| Construct a heterophilic graph H from the original graph (ablation) | Graph reconstruction · Constructed H graphs | homophily ratio | Mixed |
Closest papers by meaning
- Homophily-Related: Adaptive Hybrid Graph Filter for Multi-View Graph Clustering (AAAI 2024 · similarity 0.94)
- What Is Missing For Graph Homophily? Disentangling Graph Homophily For Graph Neural Networks (NeurIPS 2024 · similarity 0.91)
- PC-Conv: Unifying Homophily and Heterophily with Two-Fold Filtering (AAAI 2024 · similarity 0.89)
- HeroFilter: Adaptive Spectral Graph Filter for Varying Heterophilic Relations (NeurIPS 2025 · similarity 0.89)
- Understanding Heterophily for Graph Neural Networks (ICML 2024 · similarity 0.89)
- Refining Latent Homophilic Structures over Heterophilic Graphs for Robust Graph Convolution Networks (AAAI 2024 · similarity 0.89)
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.