FHAUC: Privacy Preserving AUC Calculation for Federated Learning using Fully Homomorphic Encryption

2024-03-21

arXiv.org

AI Summary

In federated learning, AUC evaluation is vulnerable to malicious aggregator attacks, and existing differential privacy (DP) schemes fail to mitigate such threats while suffering severe accuracy degradation on small datasets. To address this, we propose the first fully homomorphic encryption (FHE)-based distributed AUC computation framework for horizontal federated learning—requiring no trusted central party. Our method integrates secure multi-party computation with optimized encrypted sorting and counting circuits, enabling end-to-end privacy-preserving AUC estimation under ciphertext-only operation. Unlike DP-based approaches, our solution eliminates both accuracy loss and dataset-size dependency. Experiments demonstrate that, with 100 participants, our framework achieves 99.93% AUC accuracy in just 0.68 seconds; notably, computational efficiency and estimation accuracy remain stable even as local sample sizes decrease.

Technology Category

• Machine Learning: Distributed Machine Learning & Federated Learning
• Search and Optimization: Distributed Search
• Humans and AI: Human-in-the-loop Machine Learning

Application Category

• Security and Privacy: Data transparency and provenance
• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Ensuring data privacy is a significant challenge for machine learning applications, not only during model training but also during evaluation. Federated learning has gained significant research interest in recent years as a result. Current research on federated learning primarily focuses on preserving privacy during the training phase. However, model evaluation has not been adequately addressed, despite the potential for significant privacy leaks during this phase as well. In this paper, we demonstrate that the state-of-the-art AUC computation method for federated learning systems, which utilizes differential privacy, still leaks sensitive information about the test data while also requiring a trusted central entity to perform the computations. More importantly, we show that the performance of this method becomes completely unusable as the data size decreases. In this context, we propose an efficient, accurate, robust, and more secure evaluation algorithm capable of computing the AUC in horizontal federated learning systems. Our approach not only enhances security compared to the current state-of-the-art but also surpasses the state-of-the-art AUC computation method in both approximation performance and computational robustness, as demonstrated by experimental results. To illustrate, our approach can efficiently calculate the AUC of a federated learning system involving 100 parties, achieving 99.93% accuracy in just 0.68 seconds, regardless of data size, while providing complete data privacy.

Problem

• Addressing privacy risks in federated learning evaluation phase
• Overcoming limitations of differential privacy in AUC computation
• Ensuring secure and accurate model evaluation via homomorphic encryption

Innovation

• Uses fully homomorphic encryption for privacy
• Computes AUC securely without plaintext data
• Detects and prevents malicious aggregator manipulations