Efficient Privacy-Preserving Recommendation on Sparse Data using Fully Homomorphic Encryption

2025-09-03

AI Summary

To address the inefficiency in processing large-scale sparse rating matrices and high communication overhead in recommender systems under fully homomorphic encryption (FHE), this paper pioneers the deep integration of compressed sparse row (CSR) representation with FHE, enabling compact encrypted-domain storage and efficient sparse matrix computation. The method combines FHE with matrix factorization to achieve privacy-preserving recommendations without exposing raw user data. Its core contribution is a novel CSR-FHE co-encoding mechanism that substantially reduces ciphertext transmission volume and computational complexity. Theoretical analysis and empirical evaluation on benchmark datasets (e.g., MovieLens) demonstrate that the proposed scheme achieves recommendation accuracy comparable to plaintext models—RMSE degradation is less than 0.02—while reducing communication overhead by up to 63%. These results validate the feasibility and practicality of privacy-preserving sparse recommendation under FHE.

Technology Category

• Machine Learning: Privacy
• Data Mining & Knowledge Management: Recommender Systems
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization

Application Category

• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Security and Privacy: Large-scale security measurements
• Responsible Web: Data and user privacy-enhancing technologies for the Web

Abstract

In today's data-driven world, recommendation systems personalize user experiences across industries but rely on sensitive data, raising privacy concerns. Fully homomorphic encryption (FHE) can secure these systems, but a significant challenge in applying FHE to recommendation systems is efficiently handling the inherently large and sparse user-item rating matrices. FHE operations are computationally intensive, and naively processing various sparse matrices in recommendation systems would be prohibitively expensive. Additionally, the communication overhead between parties remains a critical concern in encrypted domains. We propose a novel approach combining Compressed Sparse Row (CSR) representation with FHE-based matrix factorization that efficiently handles matrix sparsity in the encrypted domain while minimizing communication costs. Our experimental results demonstrate high recommendation accuracy with encrypted data while achieving the lowest communication costs, effectively preserving user privacy.

Problem

• Handling sparse user-item matrices efficiently with FHE
• Reducing communication overhead in encrypted recommendation systems
• Preserving privacy while maintaining high recommendation accuracy

Innovation

• Fully Homomorphic Encryption secures recommendation systems
• Compressed Sparse Row handles matrix sparsity efficiently
• Matrix factorization minimizes encrypted communication costs