Research map / Graph learning / GNN architectures
Diffusion Probabilistic Models for Structured Node Classification
Hyosoon Jang, Seonghyun Park, Sangwoo Mo, Sungsoo Ahn
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 DPM-SNC on top of GCN | GCN · Pubmed transductive | node-level accuracy: 79.7 -> 83.0 | Improved |
| add DPM-SNC on top of GCNII | GCNII · Cora transductive | node-level accuracy: 84.0 -> 85.3 | Improved |
| add DPM-SNC on top of GCNII | GCNII · Citeseer transductive | node-level accuracy: 72.9 -> 74.1 | Improved |
| add DPM-SNC on top of GCN | GCN · Computer transductive | subgraph-level accuracy: 27.0 -> 29.7 | Improved |
| use DPM-SNC instead of LPA on GAT | GAT · Computer transductive | node-level accuracy: 84.4 -> 84.2 | Hurt |
| use DPM-SNC instead of SPN on GCN | GCN · Pubmed inductive | node-level accuracy: 80.78 -> 80.58 | No clear effect |
| use DPM-SNC instead of SPN on GAT | GAT · Pubmed inductive | graph-level accuracy: 54.82 -> 54.26 | No clear effect |
| add DPM-SNC on top of GCN | GCN · PPI inductive | micro-F1: 99.15 -> 99.46 | Improved |
| use DPM-SNC with GAT-sep backbone instead of GAT-sep alone | GAT-sep · Empire | accuracy: 88.75 -> 89.52 | Improved |
| use DPM-SNC with GAT-sep backbone instead of GAT-sep alone | GAT-sep · Rating | accuracy: 52.70 -> 54.66 | Improved |
| use DPM-SNC instead of IREM | algorithmic reasoning · Shortest path 10 nodes | Same-MSE: 0.0274 -> 0.0138 | Improved |
| use DPM-SNC instead of IREM on larger graphs | algorithmic reasoning · Edge copy 15 nodes | Large-MSE: 0.0019 -> 0.0038 | Hurt |
Closest papers by meaning
- Neural Structured Prediction for Inductive Node Classification (ICLR 2022 · similarity 0.91)
- Diffusion-Convolutional Neural Networks (NeurIPS 2016 · similarity 0.90)
- Beyond Node Sequences: Relational Diffusion for Unified Graph Learning (NeurIPS 2026 · similarity 0.90)
- SDMG: Smoothing Your Diffusion Models for Powerful Graph Representation Learning (ICML 2025 · similarity 0.88)
- Label-free Node Classification on Graphs with Large Language Models (LLMs) (ICLR 2024 · similarity 0.88)
- Unifying Generation and Prediction on Graphs with Latent Graph Diffusion (NeurIPS 2024 · 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.
