Research map / Graph learning / Graph self-supervised learning
THGB: A Comprehensive Benchmark for Text-attributed Heterogeneous Graphs
Lixin Zhou, Zemin Liu, Yuan Fang, Dan Niu, Jing Ying
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 HGNN (SimpleHGN) instead of homogeneous GCN | GNN baselines · IMDB-TS few-shot | Micro-F1: 0.6390 -> 0.6877 | Improved |
| use HGNN (SimpleHGN) instead of homogeneous GCN | GNN baselines · Patent-TA few-shot | Micro-F1: 0.5757 -> 0.6847 | Improved |
| use PLM-based THLM instead of best GNN | node classification · DBLP-TA few-shot | Micro-F1: 0.7989 -> 0.8416 | Improved |
| use LLM-based HiGPT instead of PLM-based THLM | node classification · IMDB-TS few-shot | Micro-F1: 0.7092 -> 0.7843 | Improved |
| use LLM-based OFA instead of THLM | node classification · DBLP-TA few-shot | Micro-F1: 0.8416 -> 0.8823 | Improved |
| use heterogeneous models HGT/RGCN instead of simple GCN/GAT | link prediction · IMDB-TS | ROC-AUC: GCN 0.7763 vs HGT 0.5373 | Hurt |
| use PLM/LLM-based methods instead of GNNs | link prediction · ACM-TA | ROC-AUC: 0.7190 -> 0.9696 | Improved |
| remove features of non-target node types (use only target node text) (ablation) | TAPE · IMDB-TS few-shot | accuracy / F1 | Hurt |
| use only titles or only storylines instead of both (ablation) | feature variants · IMDB-TS few-shot | accuracy / F1 | Hurt |
| add text features of other node types to target text (ablation) | feature variants · IMDB-TS few-shot | accuracy / F1 | Improved |
Closest papers by meaning
- DTGB: A Comprehensive Benchmark for Dynamic Text-Attributed Graphs (NeurIPS 2024 · similarity 0.91)
- TEG-DB: A Comprehensive Dataset and Benchmark of Textual-Edge Graphs (NeurIPS 2024 · similarity 0.91)
- Heterogeneous Graph Neural Network on Semantic Tree (AAAI 2025 · similarity 0.91)
- HeterSEED: Semantics–Structure Decoupling for Heterogeneous Graph Learning under Heterophily (NeurIPS 2026 · similarity 0.90)
- Unifying Text Semantics and Graph Structures for Temporal Text-attributed Graphs with Large Language Models (NeurIPS 2025 · similarity 0.89)
- Bootstrapping Heterogeneous Graph Representation Learning via Large Language Models: A Generalized Approach (AAAI 2025 · 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.
