🤖 AI Summary
Existing privacy-preserving approximate nearest neighbor (PP-ANN) methods struggle to balance security and efficiency and often overlook privacy risks during database construction, leaving them vulnerable to embedding inversion and membership inference attacks. This work proposes PPPQ-ANN, the first framework to synergistically integrate fully homomorphic encryption (FHE) and trusted execution environments (TEEs) across the entire ANN pipeline—spanning both database construction and query processing. By combining product quantization with ciphertext packing optimizations, the framework substantially reduces FHE computational overhead. Evaluated on million-scale datasets, the system achieves database construction in under two hours and sequential search throughput exceeding 50 queries per second, demonstrating practical performance while providing strong privacy guarantees.
📝 Abstract
A nearest-neighbor framework is a fundamental tool for various applications involving Large Language Models (LLMs) and Visual Language Models (VLMs). Vectors used for nearest-neighbor searches have richer information for similarity searches. This information leads to security risks, such as embedding inversion and membership attacks. Therefore, Privacy-Preserving Approximate Nearest-Neighbor (PP-ANN) approaches are necessary for highly confidential data. However, conventional PP-ANN approaches based on a Trusted Execution Environment (TEE) or Fully Homomorphic Encryption (FHE) do not achieve practical security or performance. Additionally, conventional approaches focus on the search process rather than database generation for nearest-neighbor. To address these issues, we propose a Privacy-Preserving Product-Quantization Approximate Nearest Neighbor (PPPQ-ANN) framework. PPPQ-ANN provides a multi-layered security structure for vectors based on a hybrid of FHE and TEE. Additionally, PPPQ-ANN minimizes FHE ciphertext computations by combining Product-Quantization (PQ) with optimized data packing. We demonstrate the performance of PPPQ-ANN on million-scale datasets. As a result, PPPQ-ANN achieves database generation in less than 2 hours and more than 50 QPS in a sequential search while preserving privacy. Therefore, PPPQ-ANN optimizes the trade-off between security and performance by utilizing a hybrid of FHE and TEE, achieving practical performance while preserving privacy.