Research map / Graph learning / Graph self-supervised learning
Bridging Structure and Semantics: Uncertainty-Modulated Dual-Path Diffusion for Robust Text-Attributed Graph Learning
Zhizhi Yu, Jiachen Liu, Qingyu Li, Dongxiao He, Di Jin
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 cross-path interaction between semantic and structural branches (ablation) | UDPD · Citeseer | accuracy (%): 82.72 -> 85.89 | Improved |
| add semantic diffusion (instead of static semantic input) (ablation) | UDPD · OGBN-Arxiv | accuracy (%): 74.88 -> 78.55 | Improved |
| add structural diffusion (instead of static structural input) (ablation) | UDPD · Cora | accuracy (%): 88.03 -> 92.73 | Improved |
| add uncertainty-modulated interaction strength (instead of fixed) (ablation) | UDPD · OGBN-Arxiv | accuracy (%): 76.16 -> 78.55 | Improved |
| add LLM keyword embedding (instead of raw text encoding) (ablation) | UDPD · PubMed | accuracy (%): 88.56 -> 94.93 | Improved |
| add LLM keyword embedding (instead of raw text encoding) (ablation) | UDPD · OGBN-Arxiv | accuracy (%): 71.21 -> 78.55 | Improved |
| add interaction between paths (ablation) | UDPD · Cora | accuracy (%): 89.82 -> 92.73 | Improved |
| add uncertainty modulation (ablation) | UDPD · Cora | accuracy (%): 88.50 -> 92.73 | Improved |
| add semantic diffusion (ablation) | UDPD · PubMed | accuracy (%): 89.09 -> 94.93 | Improved |
| add structural diffusion (ablation) | UDPD · ArXiv-2023 | accuracy (%): 80.94 -> 83.34 | Improved |
| add interaction between paths (ablation) | UDPD · PubMed | accuracy (%): 90.01 -> 94.93 | Improved |
| add uncertainty modulation (ablation) | UDPD · ArXiv-2023 | accuracy (%): 82.71 -> 83.34 | Improved |
Closest papers by meaning
- SSTAG: Structure-Aware Self-Supervised Learning Method for Text-Attributed Graphs (NeurIPS 2025 · similarity 0.88)
- Cross-Domain Graph Data Scaling: A Showcase with Diffusion Models (NeurIPS 2025 · similarity 0.88)
- Unifying Text Semantics and Graph Structures for Temporal Text-attributed Graphs with Large Language Models (NeurIPS 2025 · similarity 0.88)
- DTGB: A Comprehensive Benchmark for Dynamic Text-Attributed Graphs (NeurIPS 2024 · similarity 0.87)
- HeterSEED: Semantics–Structure Decoupling for Heterogeneous Graph Learning under Heterophily (NeurIPS 2026 · similarity 0.87)
- Generalization Principles for Inference over Text-Attributed Graphs with Large Language Models (ICML 2025 · similarity 0.86)
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.
