Personalized Bayesian Federated Learning with Wasserstein Barycenter Aggregation

Yifan Sun 孙怡帆

中国人民大学

孙怡帆，中国人民大学统计学院教授，博士生导师，人文社科部副部长、数理统计系系主任，教育部人文社会科学重点研究基地应用统计研究中心研究员，全国工业统计学教学研究会常务理事、中国统计教育学会理事，目前主要从事联邦学习和隐私计算领域研究，在NIPS、ICML、AAAI等高人工智能会议和期刊发表学术论文40余篇，主持国家自然科学基金，获教学标兵、北京市高等教育教学成果一等奖等教学奖励。

# Organizer

Jiang Hu 户将

# Time

Tuesday, 10:00-11:00 am

Dec 16, 2025

# Venue

B654, Shuangqing Complex Building

#Abstract

Personalized Bayesian federated learning (PBFL) handles non-i.i.d. client data and quantifies uncertainty by combining personalization with Bayesian inference. However, existing PBFL methods face two limitations: restrictive parametric assumptions in client posterior inference and naive parameter averaging for server aggregation. To overcome these issues, we propose FedWBA, a novel PBFL method that enhances both local inference and global aggregation. At the client level, we use particle-based variational inference for nonparametric posterior representation. At the server level, we introduce particle-based Wasserstein barycenter aggregation, offering a more geometrically meaningful approach. Theoretically, we provide local and global convergence guarantees for FedWBA. Locally, we prove a KL divergence decrease lower bound per iteration for variational inference convergence. Globally, we show that the Wasserstein barycenter converges to the true parameter as the client data size increases. Empirically, experiments show that FedWBA outperforms baselines in prediction accuracy, uncertainty calibration, and convergence rate, with ablation studies confirming its robustness.