Efficient Privacy-Preserving Counting Method with Homomorphic Encryption

Jae-yun Kim, Jieun Yun, Jung Hee Cheon, Saerom Park

Abstract: Homomorphic Encryption (HE) is a promising tool for privacy-preserving analytics on sensitive data. However, fundamental operations like counting remain challenging due to HE’s restricted operations. In this work, we present a novel and efficient encrypted counting algorithm for CKKS scheme, utilizing a simultaneous equality checking method with decomposed indices. This algorithm facilitates secure statistical data analysis, which is applicable to data with a large domain size and serves as a foundation for various privacy-preserving tasks. We implement counting, term frequency acquisition, n-gram extraction, and information retrieval for CKKS scheme using our novel methodology. Through the experiments, we demonstrated that our algorithm takes 5.2 s to count the frequency of each word in a document containing 256 words, with a vocabulary size of 256. After counting, the information retrieval algorithm takes 35.6 s to find the relevant document among 256 documents from the Amazon fine food review dataset.

External IDs:dblp:conf/icisc/KimYCP24