Saved in:
| Main Authors: | , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2604.00399 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866914435998679040 |
|---|---|
| author | Chen, Yaqi Huang, Shixun Twemlow, Ryan Wang, Lei Le, John Wang, Sheng Susilo, Willy Yan, Jun Shen, Jun |
| author_facet | Chen, Yaqi Huang, Shixun Twemlow, Ryan Wang, Lei Le, John Wang, Sheng Susilo, Willy Yan, Jun Shen, Jun |
| contents | GNN prompting aims to adapt models across tasks and graphs without requiring extensive retraining. However, most existing graph prompt methods still require task-specific parameter updates and face the issue of generalizing across graphs, limiting their performance and undermining the core promise of prompting. In this work, we introduce a Cross-graph Tuning-free Prompting Framework (CTP), which supports both homogeneous and heterogeneous graphs, can be directly deployed to unseen graphs without further parameter tuning, and thus enables a plug-and-play GNN inference engine. Extensive experiments on few-shot prediction tasks show that, compared to SOTAs, CTP achieves an average accuracy gain of 30.8% and a maximum gain of 54%, confirming its effectiveness and offering a new perspective on graph prompt learning. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_00399 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | A Cross-graph Tuning-free GNN Prompting Framework Chen, Yaqi Huang, Shixun Twemlow, Ryan Wang, Lei Le, John Wang, Sheng Susilo, Willy Yan, Jun Shen, Jun Machine Learning GNN prompting aims to adapt models across tasks and graphs without requiring extensive retraining. However, most existing graph prompt methods still require task-specific parameter updates and face the issue of generalizing across graphs, limiting their performance and undermining the core promise of prompting. In this work, we introduce a Cross-graph Tuning-free Prompting Framework (CTP), which supports both homogeneous and heterogeneous graphs, can be directly deployed to unseen graphs without further parameter tuning, and thus enables a plug-and-play GNN inference engine. Extensive experiments on few-shot prediction tasks show that, compared to SOTAs, CTP achieves an average accuracy gain of 30.8% and a maximum gain of 54%, confirming its effectiveness and offering a new perspective on graph prompt learning. |
| title | A Cross-graph Tuning-free GNN Prompting Framework |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2604.00399 |