Research map / Graph learning / Graph self-supervised learning
A Comprehensive Study on Text-attributed Graphs: Benchmarking and Rethinking
Hao Yan, Chaozhuo Li, Ruosong Long, Chao Yan, Jianan Zhao, Wenwen Zhuang, Jun Yin, Peiyan Zhang, Weihao Han, Hao Sun, Weiwei Deng, Qi Zhang and 3 more
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 topological contrastive pre-training (TCL) to the PLM (ablation) | BERT-Base · ogbn-arxiv-TA | accuracy: 72.96 -> 74.87 | Improved |
| add topological masked LM pre-training (TMLM) (ablation) | BERT-Base · ogbn-arxiv-TA | accuracy: 72.96 -> 73.97 | Improved |
| add Deepwalk-alignment pre-training (TDK) (ablation) | BERT-Base · ogbn-arxiv-TA | accuracy: 72.96 -> 74.23 | Improved |
| combine TMLM then TCL+TDK (TMDC) instead of single task (ablation) | BERT-Base · ogbn-arxiv-TA | accuracy: 74.87 -> 76.11 | Improved |
| add TMDC pre-training | BERT-Base · Photo | accuracy: 77.53 -> 84.09 (best GNN 84.46) | Mixed |
| add TMDC pre-training | BERT-Base · Children | accuracy: 59.91 -> 61.43 | Improved |
| co-train BERT-Tiny with GCN instead of static features | GCN · ogbn-arxiv-TA | accuracy: 72.03 -> 69.22 | Hurt |
| co-train BERT-Tiny with GraphSAGE instead of static features | GraphSAGE · Photo | accuracy: 83.25 -> 86.64 | Improved |
| co-train BERT-Tiny with GCN instead of static features | GCN · Computers | accuracy: 87.43 -> 83.93 | Hurt |
| increase sampled neighbors per layer in co-training (ablation) | GCN(T)/SAGE(T) · ogbn-arxiv-TA, History, Computers, Photo | accuracy | Improved |
| use DeBERTa instead of BERT/RoBERTa as static feature encoder | RevGAT · ogbn-arxiv-TA | accuracy: 74.82 -> 68.26 (Base RoBERTa vs DeBERTa) | Hurt |
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)
- Generalization Principles for Inference over Text-Attributed Graphs with Large Language Models (ICML 2025 · similarity 0.91)
- SSTAG: Structure-Aware Self-Supervised Learning Method for Text-Attributed Graphs (NeurIPS 2025 · similarity 0.91)
- Unifying Text Semantics and Graph Structures for Temporal Text-attributed Graphs with Large Language Models (NeurIPS 2025 · similarity 0.90)
- Harnessing Explanations: LLM-to-LM Interpreter for Enhanced Text-Attributed Graph Representation Learning (ICLR 2024 · similarity 0.90)
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.
