Paper 2018/309
Error Estimation of Practical Convolution Discrete Gaussian Sampling with Rejection Sampling
Zhongxiang Zheng, Xiaoyun Wang, Guangwu Xu, and Chunhuan Zhao
Abstract
Discrete Gaussian Sampling is a fundamental tool in lattice cryptography which has been used in digital signatures, identify-based encryption, attribute-based encryption, zero-knowledge proof and fully homomorphic cryptosystem. As a subroutine of lattice-based scheme, a high precision sampling usually leads to a high security level and also brings large time and space complexity. In order to optimize security and efficiency, how to achieve a higher security level with a lower precision becomes a widely studied open question. A popular method for addressing this question is
to use different metrics other than statistical distance to measure errors. The proposed metrics include KL-divergence, Rényi-divergence, and Max-log distance, and these techniques are supposed to achieve
Note: Update the analysis about delta_RD
Metadata
- Available format(s)
- -- withdrawn --
- Publication info
- Preprint. MINOR revision.
- Keywords
- Discrete Gaussian Samplingconvolution theoremlatticeerror estimation
- Contact author(s)
- zhengzx13 @ mails tsinghua edu cn
- History
- 2019-02-01: withdrawn
- 2018-04-03: received
- See all versions
- Short URL
- https://ia.cr/2018/309
- License
-
CC BY