GSTBench: A Benchmark Study on the Transferability of Graph Self-Supervised Learning
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2025
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866908525324664832 |
|---|---|
| author | Song, Yu Hua, Zhigang Xie, Yan Liu, Jingzhe Long, Bo Liu, Hui |
| author_facet | Song, Yu Hua, Zhigang Xie, Yan Liu, Jingzhe Long, Bo Liu, Hui |
| contents | Self-supervised learning (SSL) has shown great promise in graph representation learning. However, most existing graph SSL methods are developed and evaluated under a single-dataset setting, leaving their cross-dataset transferability largely unexplored and limiting their ability to leverage knowledge transfer and large-scale pretraining, factors that are critical for developing generalized intelligence beyond fitting training data. To address this gap and advance foundation model research for graphs, we present GSTBench, the first systematic benchmark for evaluating the transferability of graph SSL methods. We conduct large-scale pretraining on ogbn-papers100M and evaluate five representative SSL methods across a diverse set of target graphs. Our standardized experimental setup decouples confounding factors such as model architecture, dataset characteristics, and adaptation protocols, enabling rigorous comparisons focused solely on pretraining objectives. Surprisingly, we observe that most graph SSL methods struggle to generalize, with some performing worse than random initialization. In contrast, GraphMAE, a masked autoencoder approach, consistently improves transfer performance. We analyze the underlying factors that drive these differences and offer insights to guide future research on transferable graph SSL, laying a solid foundation for the "pretrain-then-transfer" paradigm in graph learning. Our code is available at https://github.com/SongYYYY/GSTBench. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_06975 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | GSTBench: A Benchmark Study on the Transferability of Graph Self-Supervised Learning Song, Yu Hua, Zhigang Xie, Yan Liu, Jingzhe Long, Bo Liu, Hui Machine Learning Artificial Intelligence Self-supervised learning (SSL) has shown great promise in graph representation learning. However, most existing graph SSL methods are developed and evaluated under a single-dataset setting, leaving their cross-dataset transferability largely unexplored and limiting their ability to leverage knowledge transfer and large-scale pretraining, factors that are critical for developing generalized intelligence beyond fitting training data. To address this gap and advance foundation model research for graphs, we present GSTBench, the first systematic benchmark for evaluating the transferability of graph SSL methods. We conduct large-scale pretraining on ogbn-papers100M and evaluate five representative SSL methods across a diverse set of target graphs. Our standardized experimental setup decouples confounding factors such as model architecture, dataset characteristics, and adaptation protocols, enabling rigorous comparisons focused solely on pretraining objectives. Surprisingly, we observe that most graph SSL methods struggle to generalize, with some performing worse than random initialization. In contrast, GraphMAE, a masked autoencoder approach, consistently improves transfer performance. We analyze the underlying factors that drive these differences and offer insights to guide future research on transferable graph SSL, laying a solid foundation for the "pretrain-then-transfer" paradigm in graph learning. Our code is available at https://github.com/SongYYYY/GSTBench. |
| title | GSTBench: A Benchmark Study on the Transferability of Graph Self-Supervised Learning |
| topic | Machine Learning Artificial Intelligence |
| url | https://arxiv.org/abs/2509.06975 |