Biography

Chenguang Xiao(肖晨光) is a PhD graduate from School of Computer Science, University of Birmingham. Currently, His research interests lie in the conjunction between Federated learning and class imbalance problem under the supervision of Dr. Shuo Wang and Prof. Lee Chapman. Previously, he received a B.Eng. degree in Electrical Engineering and Automation from University of Electronic Science and Technology of China.

Research Interests

• Federated Learning resampling: distributed data resampling/augmentation techniques under privacy constraints
• Model bias reduction: local model bias reduction techniques with heterogeneous data sources
• Federated Learning optimization: advanced optimization algorithms for distributed learning

Publications

1. ArXiv

   Rethinking the Initialization of Momentum in Federated Learning with Heterogeneous Data

   Chenguang Xiao, Shuo Wang* (*Corresponding authors)

   (ArXiv), 2025.

   PDF Code BibTex Preprint


2. FC

   Benchmarking Federated Machine Unlearning Methods for Tabular Data

   Chenguang Xiao, Han Wu, Shuo Wang, Diederick van Thiel, Abhirup Ghosh* (*Corresponding authors)

   Financial Cryptography and Data Security 2025, (FC), 2025.

   PDF Code BibTex Under Review


3. AAAI

   FedGA: Federated Learning with Gradient Alignment for Error Asymmetry Mitigation

   Chenguang Xiao, Zheming Zuo*, Shuo Wang (*Corresponding authors)

   Annual AAAI Conference on Artificial Intelligence, (AAAI FLUID), 2025.

   PDF Code BibTex Accepted


4. ECML-PKDD

   Triplets Oversampling for Class Imbalanced Federated Datasets

   Chenguang Xiao, Shuo Wang* (*Corresponding authors)

   Joint European Conference on Machine Learning and Knowledge Discovery in Databases, (ECML-PKDD), 2023.

   PDF Code BibTex Oral Presentation


5. SSCI

   An Experimental Study of Class Imbalance in Federated Learning

   Chenguang Xiao, Shuo Wang* (*Corresponding authors)

   IEEE Symposium Series on Computational Intelligence, (SSCI), 2021.

   PDF BibTex Oral Presentation

Open Source Projects

Contributing

• LaTeX-Workshop
• cv_resume

Maintaining

• FedMind
• HDF5-FEMNIST
• jobSched

Services

Conference Reviewers

• Joint European Conference on Machine Learning and Knowledge Discovery in Databases (ECML-PKDD), 2023-2024
• Springer Machine Learning, 2023-2024