Privacy-Preserving Federated Learning from Partial Decryption Verifiable Threshold Multi-Client Functional Encryption

2025-11-16

AI Summary

In federated learning, gradient leakage poses serious privacy risks and enables poisoning attacks, while existing threshold aggregation schemes lack verifiability of aggregation results. To address this, we propose VTSAFL—a Verifiable Threshold Secure Aggregation protocol for FL—built upon a novel threshold multi-client functional encryption scheme with publicly verifiable partial decryption. VTSAFL is the first to achieve efficient verifiability of aggregated gradients while maintaining constant-size functional keys and partial decryption outputs. By integrating threshold cryptography, multi-client functional encryption, and verifiable computation, it ensures end-to-end privacy preservation and aggregation integrity. Experiments on MNIST demonstrate that VTSAFL matches baseline model accuracy while reducing total training time by over 40% and cutting communication overhead by 50%, significantly enhancing deployment efficiency on resource-constrained IoT devices.

Technology Category

• Machine Learning: Distributed Machine Learning & Federated Learning
• Computer Vision: Adversarial Attacks & Robustness
• Multiagent Systems: Multiagent Learning

Application Category

• Security and Privacy: Privacy-enhancing technologies
• Systems and Infrastructure for Web, Mobile and WoT: Federated Web and WoT systems, including distributed, federated and edge-based data processing
• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization

Abstract

In federated learning, multiple parties can cooperate to train the model without directly exchanging their own private data, but the gradient leakage problem still threatens the privacy security and model integrity. Although the existing scheme uses threshold cryptography to mitigate the inference attack, it can not guarantee the verifiability of the aggregation results, making the system vulnerable to the threat of poisoning attack. We construct a partial decryption verifiable threshold multi client function encryption scheme, and apply it to Federated learning to implement the federated learning verifiable threshold security aggregation protocol (VTSAFL). VTSAFL empowers clients to verify aggregation results, concurrently minimizing both computational and communication overhead. The size of the functional key and partial decryption results of the scheme are constant, which provides efficiency guarantee for large-scale deployment. The experimental results on MNIST dataset show that vtsafl can achieve the same accuracy as the existing scheme, while reducing the total training time by more than 40%, and reducing the communication overhead by up to 50%. This efficiency is critical for overcoming the resource constraints inherent in Internet of Things (IoT) devices.

Problem

• Addressing gradient leakage threats in federated learning privacy
• Ensuring verifiable aggregation results against poisoning attacks
• Reducing computational and communication overhead for IoT scalability

Innovation

• Partial decryption verifiable threshold encryption scheme
• Constant-size functional keys and decryption results
• Reduces training time and communication overhead significantly