A Dropout-Resilient and Privacy-Preserving Framework for Federated Learning via Lightweight Masking

TY - GEN

T1 - A Dropout-Resilient and Privacy-Preserving Framework for Federated Learning via Lightweight Masking

AU - Jiang, Yufeng

AU - Liu, Jianghua

AU - Xu, Chenhao

AU - Zuo, Cong

AU - Xu, Lei

AU - Lei, Jian

N1 - Publisher Copyright: © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2026.

PY - 2026

Y1 - 2026

N2 - Federated Learning (FL) allows models to be trained across decentralized clients without exchanging raw data, enhancing privacy. Despite this advantage, privacy concerns persist because local gradients shared with an untrusted aggregation server could potentially leak sensitive information. To address this, we introduce a dropout-resilient and privacy-preserving framework for federated learning via lightweight masking (DRPFed), which leverages secure masking and a trusted third party (TTP). The approach utilizes the Diffie-Hellman key exchange protocol to create shared secret keys between clients and the TTP, which are then used to generate masks for obscuring local gradients before they are sent to the server. To guarantee accurate aggregation, the TTP provides the final client with a compensatory mask, ensuring that the combined masks cancel out. Additionally, if a client disconnects, the TTP reallocates the missing mask among the remaining active clients to preserve aggregation correctness. Experimental evaluations demonstrate that, unlike the standard FedAvg, our method maintains model accuracy while effectively handling client dropouts. The proposed solution successfully protects gradient privacy against honest-but-curious servers and malicious clients, all while upholding the reliability of federated model training.

AB - Federated Learning (FL) allows models to be trained across decentralized clients without exchanging raw data, enhancing privacy. Despite this advantage, privacy concerns persist because local gradients shared with an untrusted aggregation server could potentially leak sensitive information. To address this, we introduce a dropout-resilient and privacy-preserving framework for federated learning via lightweight masking (DRPFed), which leverages secure masking and a trusted third party (TTP). The approach utilizes the Diffie-Hellman key exchange protocol to create shared secret keys between clients and the TTP, which are then used to generate masks for obscuring local gradients before they are sent to the server. To guarantee accurate aggregation, the TTP provides the final client with a compensatory mask, ensuring that the combined masks cancel out. Additionally, if a client disconnects, the TTP reallocates the missing mask among the remaining active clients to preserve aggregation correctness. Experimental evaluations demonstrate that, unlike the standard FedAvg, our method maintains model accuracy while effectively handling client dropouts. The proposed solution successfully protects gradient privacy against honest-but-curious servers and malicious clients, all while upholding the reliability of federated model training.

KW - Federated Learning

KW - Privacy-Preserving

KW - Single-Mask Encryption

UR - https://www.scopus.com/pages/publications/105022070705

U2 - 10.1007/978-981-95-3543-9_16

DO - 10.1007/978-981-95-3543-9_16

M3 - Conference contribution

AN - SCOPUS:105022070705

SN - 9789819535422

T3 - Lecture Notes in Computer Science

SP - 293

EP - 312

BT - Information and Communications Security - 27th International Conference, ICICS 2025, Proceedings

A2 - Han, Jinguang

A2 - Chen, Liquan

A2 - Cheng, Guang

A2 - Xiang, Yang

A2 - Susilo, Willy

PB - Springer Science and Business Media Deutschland GmbH

T2 - 27th International Conference on Information and Communications Security, ICICS 2025

Y2 - 29 October 2025 through 31 October 2025

ER -