Open Access

ARTICLE

-FedVAE: Detached Posterior-Confidence Gating for Dimension-Wise KL Regularization in Personalized Federated Collaborative Filtering

Jincheng Cai¹, Li Feng^1,*, Ni Zhao²

1 The School of Computer Science and Engineering, Macau University of Science and Technology, Macau, China
2 The School of Finance and Economics, Shenzhen University of Information Technology, Shenzhen, China

* Corresponding Author: Li Feng. Email:

Computers, Materials & Continua 2026, 89(2), 85 https://doi.org/10.32604/cmc.2026.086183

Received 26 May 2026; Accepted 10 August 2026; Issue published 15 September 2026

Abstract

Federated Variational Autoencoders (VAEs) keep interaction data local, but existing federated VAE recommenders typically apply uniform KL regularization and do not adapt dimension-wise penalties to unreliable posteriors in sparse interaction scenarios. We propose $\alpha$-FedVAE, which uses a detached, clipped normalized signal-to-noise ratio as a local confidence gate for each KL dimension of a fused user posterior, without extra communication. Across MovieLens-100K, MovieLens-1M, and Amazon Video, $\alpha$-FedVAE improves mean HR@20 by 7.8%–55.3% and NDCG@20 by 8.0%–65.8% over FedDAE. These results indicate that $\alpha$-FedVAE improves personalized recommendation under sparse and decentralized settings while preserving the communication footprint.

---

Keywords

---