Research map / Graph learning / GNN architectures
Self-Organization Preserved Graph Structure Learning with Principle of Relevant Information
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 PRI-GSL instead of IDGL | 2-layer GCN node classification · Squirrel | accuracy (%): 29.13±2.94 -> 33.87±2.08 | Helped |
| Use PRI-GSL instead of IDGL | 2-layer GCN node classification · Actor | accuracy (%): 27.44±5.80 -> 28.55±2.04 | Helped |
| Use PRI-GSL instead of IDGL | 2-layer GCN node classification · Chameleon | accuracy (%): 49.80±4.80 -> 51.83±2.44 | Helped |
| Use PRI-GSL instead of SDRF | 2-layer GCN node classification · CiteSeer | accuracy (%): 69.97±0.28 -> 69.34±2.64 | Hurt |
| Use PRI-GSL instead of Graph-PRI | 2-layer GCN node classification · PubMed | accuracy (%): 76.25±1.44 -> 76.77±3.20 | Helped |
| Use PRI-GSL instead of IDGL | 2-layer GCN node classification · Cora | accuracy (%): 83.23±0.62 -> 83.67±2.09 | Helped |
| Use PRI-GSL instead of NeuralSparse | 2-layer GCN node classification · Photo | accuracy (%): 89.40±1.85 -> 92.40±1.15 | Helped |
| Add the PRI loss (ablation) | PRI-GSL · Node classification (5 random splits; datasets shown in Figure) | accuracy | Helped |
| Add structural role encodings (ablation) | PRI-GSL · Node classification (5 random splits; datasets shown in Figure) | accuracy | Helped |
| Add both the PRI loss and structural role encodings (ablation) | GSL · Node classification (5 random splits; datasets shown in Figure) | accuracy | Helped |
| Use PRI-GSL rather than GCN or Graph-PRI under synthetic edge noise | Cora graph denoising · Cora with 25%, 50%, and 75% edge additions/deletions | accuracy | Helped |
Closest papers by meaning
- Structural Entropy Based Graph Structure Learning for Node Classification (AAAI 2024 · similarity 0.89)
- OpenGSL: A Comprehensive Benchmark for Graph Structure Learning (NeurIPS 2023 · similarity 0.88)
- Graph Structure Inference with BAM: Neural Dependency Processing via Bilinear Attention (NeurIPS 2024 · similarity 0.86)
- From Theory to Practice: Rethinking Green and Martin Kernels for Unleashing Graph Transformers (ICML 2025 · similarity 0.86)
- Transitivity-Preserving Graph Representation Learning for Bridging Local Connectivity and Role-Based Similarity (AAAI 2024 · similarity 0.85)
- On the Interplay between Graph Structure and Learning Algorithms in Graph Neural Networks (ICML 2025 · similarity 0.85)
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.