International Association for Cryptologic Research

International Association
for Cryptologic Research

CryptoDB

Charlotte Bonte

Publications

Year
Venue
Title
2022
ASIACRYPT
FINAL: Faster FHE instantiated with NTRU and LWE 📺
The NTRU problem is a promising candidate to build efficient Fully Homomorphic Encryption (FHE).However, all the existing proposals (e.g. LTV, YASHE) need so-called `overstretched' parameters of NTRU to enable homomorphic operations. It was shown by Albrecht~et~al. (CRYPTO~2016) that these parameters are vulnerable against subfield lattice attacks. Based on a recent, more detailed analysis of the overstretched NTRU assumption by Ducas and van Woerden (ASIACRYPT~2021), we construct two FHE schemes whose NTRU parameters lie outside the overstretched range.The first scheme is based solely on NTRU and demonstrates competitive performance against the state-of-the-art FHE schemes including TFHE. Our second scheme, which is based on both the NTRU and LWE assumptions, outperforms TFHE with a 28\% faster bootstrapping and 45\% smaller bootstrapping and key-switching keys.
2017
CHES
Faster Homomorphic Function Evaluation Using Non-integral Base Encoding
In this paper we present an encoding method for real numbers tailored for homomorphic function evaluation. The choice of the degree of the polynomial modulus used in all popular somewhat homomorphic encryption schemes is dominated by security considerations, while with the current encoding techniques the correctness requirement allows for much smaller values. We introduce a generic encoding method using expansions with respect to a non-integral base, which exploits this large degree at the benefit of reducing the growth of the coefficients when performing homomorphic operations. This allows one to choose a smaller plaintext coefficient modulus which results in a significant reduction of the running time. We illustrate our approach by applying this encoding in the setting of homomorphic electricity load forecasting for the smart grid which results in a speed-up by a factor 13 compared to previous work, where encoding was done using balanced ternary expansions.