EvaFL: An Efficient Verifiable Privacy-Preserving Federated Learning Against Malicious Servers
- 27 April 2026
- journal article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Dependable and Secure Computing
- Vol. 23 (4) , 8717-8732
Abstract
Federated Learning (FL) preserves client data privacy by distributing model training but remains vulnerable to inference attacks (e.g., gradient inversion). Existing secure aggregation schemes mitigate basic privacy threats, but most of them are under the semi-honest server assumption. Malicious servers can corrupt the global model through forging aggregation results. Moreover, the high interaction rounds and communication complexity of the existing schemes still constrain their feasibility in large-scale distributed deployment scenarios. To tackle these challenges, we propose EvaFL, an efficient verifiable privacy-preserving federated learning against malicious servers, which reduces the communication overhead and privacy threats from malicious severs. We propose the system model of EvaFL and give the concrete protocol. We leverage the linear homomorphism property of Shamir secret sharing under discrete logarithm assumption to reuse the mask seed shares, which avoids the communication overhead caused by share distribution in multiple rounds of iterations. In addition, by integrating consistency checking into the unmasking step, we further reduce one round interaction. To resist malicious servers, we adopt linear homomorphic hash to realize the correctness verification of the aggregation results. Finally, we implement and evaluate our EvaFL based on MNIST and CIFAR10 datasets to show its feasibility for privacy training. The single round aggregation completion time of EvaFL is reduced by 69% compared to BBGLR (CCS 2020) and by 11% compared to Flamingo (S&P 2023).Keywords
Funding Information
- National Cryptologic Science Fund of China (2025NCSF02025)
- National Natural Science Foundation of China (U24B20149)
- National Natural Science Foundation of China (62272385)
- National Natural Science Foundation of China (U23A20302)
- National Natural Science Foundation of China (62202375)
- Key Research and Development Projects of Shaanxi Province (2024GX-ZDCYL-01-09)
- Key Research and Development Projects of Shaanxi Province (2024GX-ZDCYL- 01-15)
- Natural Science Basic Research Program of Shaanxi (2025JCYBQN- 814)
- Scientific Research Program Funded by Education Department of Shaanxi Provincial Government (24JK0653)
This publication has 28 references indexed in Scilit:
- Starlit: Privacy-Preserving Federated Learning to Enhance Financial Fraud Detection2025
- Federated learning with differential privacy for breast cancer diagnosis enabling secure data sharing and model integrityScientific Reports, 2025
- Ride-hailing pick-up area recommendation in a vehicle-cloud collaborative environment: a feature-aware personalized clustering federated learning approachCluster Computing, 2024
- Enhancing Security: Federated Learning against Man-In-The-Middle Threats with Gradient Boosting Machines and LSTM2024
- Mitigating Data Injection Attacks on Federated Learning2024
- Federated learning for medical image analysis: A surveyPattern Recognition, 2024
- Flamingo: Multi-Round Single-Server Secure Aggregation with Applications to Private Federated Learning2023
- Prio+: Privacy Preserving Aggregate Statistics via Boolean Shares2022
- Membership Inference Attacks From First Principles2022
- A practical scheme for non-interactive verifiable secret sharing1987