Yaxin Xiao

ASTAPLE | EEE, HK PolyU

Yaxin Xiao (right)

Yaxin Xiao is a Research Assistant Professor in the ASTAPLE group, Department of Electrical and Electronic Engineering (EEE), at the Hong Kong Polytechnic University (HK PolyU). Her research lies in privacy and security of machine learning, especially model extraction, membership inference attacks, machine unlearning, watermarking for intellectual property protection, and LLM-agent security.

Previously, she was a Postdoctoral Fellow at HK PolyU, supervised by Prof. Haibo Hu and Prof. Qingqing Ye. She received her Ph.D. from PolyU supervised by Prof. Haibo Hu under the support of the Hong Kong PhD Fellowship Scheme (HKPFS), and earned both her B.Eng. and M.Eng. degrees from Zhejiang University.

Recent venues: NeurIPS, ICCV, AAAI, IEEE TDSC, ACL, and ICDE. A fuller publication list is available on the Publications page, and a CV (PDF) is also available.

selected publications

AAAI / CCF-A

Class-feature Watermark: A Resilient Black-box Watermark Against Model Extraction Attacks

Yaxin Xiao, Qingqing Ye, Zi Liang, Haoyang Li, RongHua Li, Huadi Zheng, and Haibo Hu^*

In Proceedings of the AAAI Conference on Artificial Intelligence, 2026

Bib PDF

@inproceedings{xiao2026classfeature,
  title = {Class-feature Watermark: A Resilient Black-box Watermark Against Model Extraction Attacks},
  author = {Xiao, Yaxin and Ye, Qingqing and Liang, Zi and Li, Haoyang and Li, RongHua and Zheng, Huadi and Hu, Haibo},
  booktitle = {Proceedings of the AAAI Conference on Artificial Intelligence},
  year = {2026},
}

ICCV / CCF-A

Reminiscence Attack on Residuals: Exploiting Approximate Machine Unlearning for Privacy

Yaxin Xiao, Qingqing Ye, Li Hu, Huadi Zheng, Haibo Hu^*, Zi Liang, Haoyang Li, and Yijie Jiao

In IEEE/CVF International Conference on Computer Vision, 2025

Bib PDF

@inproceedings{xiao2025privacyattacks,
  title = {Reminiscence Attack on Residuals: Exploiting Approximate Machine Unlearning for Privacy},
  author = {Xiao, Yaxin and Ye, Qingqing and Hu, Li and Zheng, Huadi and Hu, Haibo and Liang, Zi and Li, Haoyang and Jiao, Yijie},
  booktitle = {IEEE/CVF International Conference on Computer Vision},
  year = {2025},
}

TDSC / CCF-A

Unlocking High-Fidelity Learning: Towards Neuron-Grained Model Extraction

Yaxin Xiao, Haibo Hu^*, Qingqing Ye, Li Tang, Zi Liang, and Huadi Zheng

IEEE Transactions on Dependable and Secure Computing, 2025

Bib PDF

@article{xiao2025unlocking,
  title = {Unlocking High-Fidelity Learning: Towards Neuron-Grained Model Extraction},
  author = {Xiao, Yaxin and Hu, Haibo and Ye, Qingqing and Tang, Li and Liang, Zi and Zheng, Huadi},
  journal = {IEEE Transactions on Dependable and Secure Computing},
  year = {2025},
}

NeurIPS / CCF-A

MExMI: Pool-based Active Model Extraction Crossover Membership Inference

Yaxin Xiao, Qingqing Ye, Haibo Hu^*, Huadi Zheng, Chengfang Fang, and Jie Shi

In Advances in Neural Information Processing Systems, 2022

Bib PDF

@inproceedings{xiao2022mexmi,
  title = {MExMI: Pool-based Active Model Extraction Crossover Membership Inference},
  author = {Xiao, Yaxin and Ye, Qingqing and Hu, Haibo and Zheng, Huadi and Fang, Chengfang and Shi, Jie},
  booktitle = {Advances in Neural Information Processing Systems},
  year = {2022},
}