Jaehyung Kim

Jaehyung Kim

Auto-generated profile, based on publicly available information

Jaehyung Kim

Stanford University, Stanford, CA, USA

Research Focus & Metrics

Academic Career Length

5Years (2022-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
8 May 2024
Verified
Book Chapter
Closed access
PDF file
YBYoungjin Bae
Youngjin Bae
JCJung Hee Cheon
Jung Hee Cheon
JKJaehyung Kim
Jaehyung Kim
DSDamien Stehlé
Damien Stehlé

2 December 2024
Verified
Conference Paper
Open access
JKJaehyung Kim
Jaehyung Kim
JSJinyeong Seo
Jinyeong Seo
YSYongsoo Song
Yongsoo Song
Bootstrapping is currently the only known method for constructing fully homomorphic encryptions. In the BFV scheme specifically, bootstrapping aims to reduce the error of a ciphertext while preserving the encrypted plaintext. The existing BFV bootstrapping methods follow the same pipeline, relying on the evaluation of a digit extraction polynomial to annihilate the error located in the least significant digits. However, due to its strong dependence on performance, bootstrapping could only utilize a limited form of plaintext modulus, such as a power of a small prime number. In this paper, we present a novel approach to instantiate BFV bootstrapping, distinct from the previous digit extraction-based method. The core idea of our bootstrapping is to utilize CKKS bootstrapping as a subroutine, so the performance of our method mainly depends on the underlying CKKS bootstrapping rather than the plaintext modulus. We implement our method at a proof-of-concept level to provide concrete benchmark results. When performing the bootstrapping operation for a 51-bits plaintext modulus, our method improves the previous digit extraction-based method by a factor of 37.9 in latency and 29.4 in throughput. Additionally, we achieve viable bootstrapping performance for large plaintext moduli, such as 144-bits and 234-bits, which has never been measured before.

10 December 2024
Verified
Book Chapter
Closed access
PDF file
YBYoungjin Bae
Youngjin Bae
JKJaehyung Kim
Jaehyung Kim
DSDamien Stehlé
Damien Stehlé
ESElias Suvanto
Elias Suvanto

15 November 2023
Verified
Conference Paper
Open access
JCJung Hee Cheon
Jung Hee Cheon
WCWonhee Cho
Wonhee Cho
JKJaehyung Kim
Jaehyung Kim
DSDamien Stehlé
Damien Stehlé
Homomorphic Encryption (HE) schemes such as BGV, BFV, and CKKS consume some ciphertext modulus for each multiplication. Bootstrapping (BTS) restores the modulus and allows homomorphic computation to continue, but it is time-consuming and requires a significant amount of modulus. For these reasons, decreasing modulus consumption is crucial topic for BGV, BFV and CKKS, on which numerous studies have been conducted. We propose a novel method, called Mult2, to perform ciphertext multiplication in the CKKS scheme with lower modulus consumption. Mult2 relies an a new decomposition of a ciphertext into a pair of ciphertexts that homomorphically performs a weak form of Euclidean division. It multiplies two ciphertexts in decomposed formats with homomorphic double precision multiplication, and its result approximately decrypts to the same value as does the ordinary CKKS multiplication. Mult2 can perform homomorphic multiplication by consuming almost half of the modulus. We extend it to Multt for any t≥ 2, which relies on the decomposition of a ciphertext into t components. All other CKKS operations can be equally performed on pair/tuple formats, leading to the double-CKKS (resp. tuple-CKKS) scheme enabling homomorphic double (resp. multiple) precision arithmetic. As a result, when the ciphertext modulus and dimension are fixed, the proposed algorithms enable the evaluation of deeper circuits without bootstrapping, or allow to reduce the number of bootstrappings required for the evaluation of the same circuits. Furthermore, they can be used to increase the precision without increasing the parameters. For example, Mult2 enables 8 sequential multiplications with 100 bit scaling factor with a ciphertext modulus of only 680 bits, which is impossible with the ordinary CKKS multiplication algorithm.

1 January 2022
Verified
Book Chapter
Closed access
PDF file
SKSeonghak Kim
Seonghak Kim
MPMinji Park
Minji Park
JKJaehyung Kim
Jaehyung Kim
TKTaekyung Kim
Taekyung Kim
CMChohong Min
Chohong Min

7 July 2025 in IACR Communications in Cryptology
Verified
Journal Article
Open access
PDF file
JKJaehyung Kim
Jaehyung Kim
TNTaeyeong Noh
Taeyeong Noh
The Cheon–Kim–Kim–Song (CKKS) scheme is renowned for its efficiency in encrypted computing over real numbers. However, it lacks an important functionality that most exact schemes have, an efficient modular reduction. This derives from the fundamental difference in encoding structure. The CKKS scheme encodes messages to the least significant bits, while the other schemes encode to the most significant bits (or in an equivalent manner). As a result, CKKS could enjoy an efficient rescaling but lost the ability to modular reduce inherently. Instead of homomorphically approximating the modular reduction function, we suggest to use the inherent modular reduction over ℤ q [ X ] / ( X N + 1 ) . We construct a novel homomorphic modular reduction algorithm using the discrete bootstrapping from Bae et al. [Asiacrypt'24] and a new discretization algorithm from modulus switching. One of the key advantages of our modular reduction is that its computational complexity grows sublinearly ( O ( log k ) ) as we increase the input range [ 0 , k ) , which is asymptotically better than the state-of-the-art with ≥ O ( k ) . We checked our algorithms with concrete experiments. Notably, our modulo 1 function for input range [ 0 , 2 20 ) takes only 44.9 seconds with 13.3 bits of (mean) precision, in a single-threaded CPU. Recall that modular reduction over such a large range was almost infeasible in the previous works, as they need to evaluate a polynomial of degree > 2 20 (or equivalent). As an application of our method, we compared a bit decomposition based on our framework with the state-of-the-art method from Drucker et al. [J.Cryptol'24]. Our method is 7.1 × faster while reducing the failure probability by more than two orders of magnitude.