Research map / Graph learning / Graph self-supervised learning
Global-Recent Semantic Reasoning on Dynamic Text-Attributed Graphs with Large Language Models
Yunan Wang, Jianxin Li, Ziwei Zhang
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 |
|---|---|---|---|
| use node-centric implicit reasoning with sliding window for recent semantics (ablation) | DyGRASP recent semantic reasoning · GDELT | token consumption: 1.68B -> 0.10B | Improved |
| use node-centric implicit reasoning with sliding window for recent semantics (ablation) | DyGRASP recent semantic reasoning · Enron | token consumption: 3.62B -> 0.86B | Improved |
| use node-centric implicit reasoning with sliding window for recent semantics (ablation) | DyGRASP recent semantic reasoning · Googlemap | token consumption: 2.42B -> 0.35B | Improved |
| use node-centric implicit reasoning with sliding window for recent semantics (ablation) | DyGRASP recent semantic reasoning · Stack_elec | token consumption: 7.00B -> 0.96B | Improved |
| add DyGRASP recent-global semantic reasoning to TGAT | TGAT · GDELT destination retrieval (transductive) | Hit@10 (%): 89.97 -> 91.96 | Improved |
| add DyGRASP recent-global semantic reasoning to TGAT | TGAT · Enron destination retrieval (transductive) | Hit@10 (%): 89.12 -> 96.25 | Improved |
| add DyGRASP recent-global semantic reasoning to TGAT | TGAT · Googlemap destination retrieval (transductive) | Hit@10 (%): 63.53 -> 82.79 | Improved |
| add DyGRASP recent-global semantic reasoning to TGAT | TGAT · Stack_elec destination retrieval (transductive) | Hit@10 (%): 93.74 -> 99.74 | Improved |
| add DyGRASP recent-global semantic reasoning to DyGFormer | DyGFormer · GDELT destination retrieval (transductive) | Hit@10 (%): 91.64 -> 93.24 | Improved |
| add DyGRASP recent-global semantic reasoning to DyGFormer | DyGFormer · Enron destination retrieval (transductive) | Hit@10 (%): 92.04 -> 99.40 | Improved |
| add DyGRASP recent-global semantic reasoning to DyGFormer | DyGFormer · Googlemap destination retrieval (transductive) | Hit@10 (%): 51.32 -> 85.88 | Improved |
| add DyGRASP to DyGFormer on Googlemap (inductive) | DyGFormer · Googlemap destination retrieval (inductive) | Hit@10 (%): 43.84 -> 81.14 | Improved |
Closest papers by meaning
- Unifying Text Semantics and Graph Structures for Temporal Text-attributed Graphs with Large Language Models (NeurIPS 2025 · similarity 0.92)
- DTGB: A Comprehensive Benchmark for Dynamic Text-Attributed Graphs (NeurIPS 2024 · similarity 0.92)
- Generalization Principles for Inference over Text-Attributed Graphs with Large Language Models (ICML 2025 · similarity 0.90)
- SSTAG: Structure-Aware Self-Supervised Learning Method for Text-Attributed Graphs (NeurIPS 2025 · similarity 0.88)
- Talk like a Graph: Encoding Graphs for Large Language Models (ICLR 2024 · similarity 0.88)
- TGB 2.0: A Benchmark for Learning on Temporal Knowledge Graphs and Heterogeneous Graphs (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.
