Compacting privacy-preserving k-nearest neighbor search using logic synthesis

Abstract

This paper introduces the first efficient, scalable, and practical method for privacy-preserving k-nearest neighbors (k-NN) search. The approach enables performing the widely used k-NN search in sensitive scenarios where none of the parties reveal their information while they can still cooperatively find the nearest matches. The privacy preservation is based on the Yao's garbled circuit (GC) protocol. In contrast with the existing GC approaches that only accept function descriptions as combinational circuits, we suggest using sequential circuits. This work introduces novel transformations, such that the sequential description can be evaluated by interfacing with the existing GC schemes that only accept combinational circuits. We demonstrate a great efficiency in the memory required for realizing the secure k-NN search. The first-of-a-kind implementation of privacy preserving k-NN, utilizing the Synopsys Design Compiler on a conventional Intel processor demonstrates the applicability, efficiency, and scalability of the suggested methods.

The Death and Rebirth of Privacy-Preserving WiFi Fingerprint Localization with Paillier Encryption

“…, π k of the k smallest d i and it can calculate its location similarly as in PriWFL. Songhori et al [28] presented a memory-efficient sequential garbled gate for k-nearest neighbors search and their circuit can be used for our purpose. The communication overhead of transferringx andỹ grows linearly with M and is in the magnitude of some kBs for M = 505 and κ = 128.…”

PILOT: Practical Privacy-Preserving Indoor Localization Using OuTsourcing

“…On the other hand, Bitonic Sort becomes more efficient with increasing k and would be more efficient than k-NN when k is greater than 40 or 50 because its performance is independent of k. However, for IL k is usually small, e.g., k=3 or k=4. Our k-NN protocol achieves a 1.4× to 1.6× run-time improvement over [97].…”

“…Songhori et al [97] propose a circuit for k-Nearest Neighbors (k-NN) with size M (2k − 1)(2 + log 2 M ) AND gates, where M is the number of elements (here, the number of Reference Points) and is their bit-length.…”

“…Optimized k-Nearest Neighbors: We improve the k-NN circuit of [97] by using more efficient conditional swap circuits of [65]. This results in Mk(2 + log 2 M ) AND gates which for k=3 is an improvement by factor 1.6× over [97]. Our modified k-NN implementation is given in Alg.…”

“…An alternative approach for computing k-NN is to use Oblivious Sorting. This improves the efficiency of the non-constant round SMPC protocols such as GMW because the k-NN algorithm of [97] has high depth of O (Mk ). Instead, we use the Bitonic Sort algorithm of [98] that is summarized in Alg.…”