PALA: Class-imbalanced Graph Domain Adaptation via Prototype-anchored Learning and Alignment

Ma, Xin; Wang, Yifan; Yi, Siyu; Ju, Wei; Wu, Bei; Qiao, Ziyue; Tang, Chenwei; Lv, Jiancheng

doi:10.24963/ijcai.2025/356

PALA: Class-imbalanced Graph Domain Adaptation via Prototype-anchored Learning and Alignment

Xin Ma, Yifan Wang, Siyu Yi, Wei Ju, Bei Wu, Ziyue Qiao, Chenwei Tang, Jiancheng Lv

Download BibTeX Cite

Proceedings of the Thirty-Fourth International Joint Conference on Artificial Intelligence. Main Track. Pages 3198-3207.

https://doi.org/10.24963/ijcai.2025/356

Keywords

Data Mining: DM: Class imbalance and unequal cost
Data Mining: DM: Mining graphs

Abstract

Graph domain adaptation is a key subfield of graph transfer learning that aims to bridge domain gaps by transferring knowledge from a label-rich source graph to an unlabeled target graph. However, most existing methods assume balanced labels in the source graph, which often fails in practice and leads to biased knowledge transfer. To address this, in this paper, we propose a prototype-anchored learning and alignment framework for class-imbalanced graph domain adaptation. Specifically, we incorporate pointwise node mutual information into the graph encoder to capture high-order topological proximity and learn generalized node representations. Leveraging this, we then introduce categorical prototypes with adversarial proto-instances for prototype-anchored learning and recalibration to represent the source graph under an imbalanced class distribution. Finally, we introduce a weighted prototype contrastive adaptation strategy that aligns target pseudo-labels with source prototypes to handle class imbalance during adaptation. Extensive experiments show that our PALA outperforms the state-of-the-art methods. Our code is available at https://github.com/maxin88scu/PALA.

Xin Ma, Yifan Wang, Siyu Yi, Wei Ju, Bei Wu, Ziyue Qiao, Chenwei Tang, and Jiancheng Lv. 2025. PALA: Class-imbalanced Graph Domain Adaptation via Prototype-anchored Learning and Alignment. In Proceedings of the Thirty-Fourth International Joint Conference on Artificial Intelligence, IJCAI-25. IJCAI, 3198–3207. https://doi.org/10.24963/ijcai.2025/356

@inproceedings{ijcai2025p356,
  title     = {PALA: Class-imbalanced Graph Domain Adaptation via Prototype-anchored Learning and Alignment},
  author    = {Ma, Xin and Wang, Yifan and Yi, Siyu and Ju, Wei and Wu, Bei and Qiao, Ziyue and Tang, Chenwei and Lv, Jiancheng},
  booktitle = {Proceedings of the Thirty-Fourth International Joint Conference on
               Artificial Intelligence, {IJCAI-25}},
  publisher = {International Joint Conferences on Artificial Intelligence Organization},
  editor    = {James Kwok},
  pages     = {3198--3207},
  year      = {2025},
  month     = {8},
  note      = {Main Track},
  doi       = {10.24963/ijcai.2025/356},
  url       = {https://doi.org/10.24963/ijcai.2025/356},
}

TY  - CPAPER
TI  - PALA: Class-imbalanced Graph Domain Adaptation via Prototype-anchored Learning and Alignment
AU  - Ma, Xin
AU  - Wang, Yifan
AU  - Yi, Siyu
AU  - Ju, Wei
AU  - Wu, Bei
AU  - Qiao, Ziyue
AU  - Tang, Chenwei
AU  - Lv, Jiancheng
T2  - Proceedings of the Thirty-Fourth International Joint Conference on Artificial Intelligence, IJCAI-25
PB  - International Joint Conferences on Artificial Intelligence Organization
PY  - 2025
SP  - 3198
EP  - 3207
DO  - 10.24963/ijcai.2025/356
UR  - https://doi.org/10.24963/ijcai.2025/356
ER  -