Bilkent University
Department of Computer Engineering
CS 590/690 SEMINAR
HADES-FL: Privacy-Preserving Federated Learning via Selective Feature Encryption and Hybrid Model Fusion
Ergün Batuhan Kaynak
Master Student
(Supervisor: Asst.Prof.Sinem Sav)
Computer Engineering Department
Bilkent University
Abstract: Data privacy has become a critical challenge as the growing demand for accurate models requires handling more sensitive data, which in turn raises significant privacy risks. We address the challenge of privacy-preserving training in federated learning (FL) by introducing HADES-FL, a novel framework that selectively encrypts only the most privacy-sensitive features while leaving the remaining data and the model portion unencrypted. This system ensures both privacy protection and computational efficiency. Unlike fully encrypted FL training pipelines, which suffer from high computational overhead, HADES-FL integrates an encrypted and non-encrypted training pipeline via a fusion mechanism, enabling seamless interaction between encrypted and plaintext representations. To achieve this, we first apply principal component analysis (PCA) to identify and select the most privacy-sensitive features for encryption. Our analysis demonstrates that training a network with only these selected features significantly reduces the success rate of reconstruction attacks in FL. Building on this insight, we design a hybrid FL system that trains an end-to-end encrypted network via multiparty homomorphic encryption (MHE) on the selected features while simultaneously training a plaintext network on the remaining features. These two networks are then integrated using a fusion mechanism, ensuring both privacy preservation and computational efficiency. Finally, we demonstrate that HADES-FL achieves accuracy on par with vanilla FL while preserving privacy.
DATE: April 14, Thursday @ 15:00 Place: EA 502