How Much Can Transfer? BRIDGE: Bounded Multi-Domain Graph Foundation Model with Generalization Guarantees

Haonan Yuan, Qingyun Sun, Junhua Shi, Xingcheng Fu, Bryan Hooi, Jianxin Li, Philip S. Yu
Proceedings of the 42nd International Conference on Machine Learning, PMLR 267:73604-73644, 2025.

Abstract

Graph Foundation Models hold significant potential for advancing multi-domain graph learning, yet their full capabilities remain largely untapped. Existing works show promising task performance with the “pretrain-then-prompt” paradigm, which lacks theoretical foundations to understand why it works and how much knowledge can be transferred from source domains to the target. In this paper, we introduce BRIDGE, a bounded graph foundation model pre-trained on multi-domains with Generalization guarantees. To learn discriminative source knowledge, we align multi-domain graph features with domain-invariant aligners during pre-training. Then, a lightweight Mixture of Experts (MoE) network is proposed to facilitate downstream prompting through self-supervised selective knowledge assembly and transfer. Further, to determine the maximum amount of transferable knowledge, we derive an optimizable generalization error upper bound from a graph spectral perspective given the Lipschitz continuity. Extensive experiments demonstrate the superiority of BRIDGE on both node and graph classification compared with 15 state-of-the-art baselines.

Cite this Paper

BibTeX
@InProceedings{pmlr-v267-yuan25h, title = {How Much Can Transfer? {BRIDGE}: Bounded Multi-Domain Graph Foundation Model with Generalization Guarantees}, author = {Yuan, Haonan and Sun, Qingyun and Shi, Junhua and Fu, Xingcheng and Hooi, Bryan and Li, Jianxin and Yu, Philip S.}, booktitle = {Proceedings of the 42nd International Conference on Machine Learning}, pages = {73604--73644}, year = {2025}, editor = {Singh, Aarti and Fazel, Maryam and Hsu, Daniel and Lacoste-Julien, Simon and Berkenkamp, Felix and Maharaj, Tegan and Wagstaff, Kiri and Zhu, Jerry}, volume = {267}, series = {Proceedings of Machine Learning Research}, month = {13--19 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v267/main/assets/yuan25h/yuan25h.pdf}, url = {https://proceedings.mlr.press/v267/yuan25h.html}, abstract = {Graph Foundation Models hold significant potential for advancing multi-domain graph learning, yet their full capabilities remain largely untapped. Existing works show promising task performance with the “pretrain-then-prompt” paradigm, which lacks theoretical foundations to understand why it works and how much knowledge can be transferred from source domains to the target. In this paper, we introduce BRIDGE, a bounded graph foundation model pre-trained on multi-domains with Generalization guarantees. To learn discriminative source knowledge, we align multi-domain graph features with domain-invariant aligners during pre-training. Then, a lightweight Mixture of Experts (MoE) network is proposed to facilitate downstream prompting through self-supervised selective knowledge assembly and transfer. Further, to determine the maximum amount of transferable knowledge, we derive an optimizable generalization error upper bound from a graph spectral perspective given the Lipschitz continuity. Extensive experiments demonstrate the superiority of BRIDGE on both node and graph classification compared with 15 state-of-the-art baselines.} }
Endnote
%0 Conference Paper %T How Much Can Transfer? BRIDGE: Bounded Multi-Domain Graph Foundation Model with Generalization Guarantees %A Haonan Yuan %A Qingyun Sun %A Junhua Shi %A Xingcheng Fu %A Bryan Hooi %A Jianxin Li %A Philip S. Yu %B Proceedings of the 42nd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2025 %E Aarti Singh %E Maryam Fazel %E Daniel Hsu %E Simon Lacoste-Julien %E Felix Berkenkamp %E Tegan Maharaj %E Kiri Wagstaff %E Jerry Zhu %F pmlr-v267-yuan25h %I PMLR %P 73604--73644 %U https://proceedings.mlr.press/v267/yuan25h.html %V 267 %X Graph Foundation Models hold significant potential for advancing multi-domain graph learning, yet their full capabilities remain largely untapped. Existing works show promising task performance with the “pretrain-then-prompt” paradigm, which lacks theoretical foundations to understand why it works and how much knowledge can be transferred from source domains to the target. In this paper, we introduce BRIDGE, a bounded graph foundation model pre-trained on multi-domains with Generalization guarantees. To learn discriminative source knowledge, we align multi-domain graph features with domain-invariant aligners during pre-training. Then, a lightweight Mixture of Experts (MoE) network is proposed to facilitate downstream prompting through self-supervised selective knowledge assembly and transfer. Further, to determine the maximum amount of transferable knowledge, we derive an optimizable generalization error upper bound from a graph spectral perspective given the Lipschitz continuity. Extensive experiments demonstrate the superiority of BRIDGE on both node and graph classification compared with 15 state-of-the-art baselines.
APA
Yuan, H., Sun, Q., Shi, J., Fu, X., Hooi, B., Li, J. & Yu, P.S.. (2025). How Much Can Transfer? BRIDGE: Bounded Multi-Domain Graph Foundation Model with Generalization Guarantees. Proceedings of the 42nd International Conference on Machine Learning, in Proceedings of Machine Learning Research 267:73604-73644 Available from https://proceedings.mlr.press/v267/yuan25h.html.

Related Material