The Learning with Errors (LWE) problem is a fundamental hardness assumption for post-quantum lattice-based cryptography. Standard LWE schemes rely on discrete Gaussian error distributions, which incur high sampling overhead. This paper explores the feasibility of replacing the Gaussian distribution with a novel non-Gaussian alternative—the Step Distribution—within Regev’s classic LWE public-key encryption scheme. The Step Distribution partitions the error range into equal-width steps, with probabilities proportional to Gaussian weights at step midpoints. The distribution is formally defined, its asymptotic convergence to the discrete Gaussian is proved, and conditions under which the decryption error probability becomes negligible are derived. The Step Distribution offers a tunable parameter—the number of steps—that allows a trade-off between security (closeness to the Gaussian) and computational efficiency. This work provides a flexible error distribution for LWE-based cryptosystems, particularly suitable for resource-constrained environments.
In this paper we propose the first provably secure public key encryption scheme based on the Learning with Errors (LWE) problem, in which secrets and errors are sampled uniformly at random from a relatively small set rather than from the commonly used discrete Gaussian distribution.Using a uniform distribution, instead of a Gaussian, has the potential of improving computational efficiency a great deal due to its simplicity, thus making the scheme attractive for use in practice.At the same time our scheme features the strong security guarantee of being based on the hardness of worst-case lattice problems.After presenting the construction of our scheme we prove its security and propose asymptotic parameters.Finally, we compare our scheme on several measures to one of the most efficient LWE-based encryption schemes with Gaussian noise.We show that the expected efficiency improvement is debunked, due to the large blow-up of the parameter sets involved.
Abstract. In this paper we propose the first provably secure public key encryption scheme based on the Learning with Errors (LWE) problem, in which secrets and errors are sampled uniformly at random from a relatively small set rather than from the commonly used discrete Gaussian distribution. Using a uniform distribution, instead of a Gaussian, has the potential of improving computational efficiency a great deal due to its simplicity, thus making the scheme attractive for use in practice. At the same time our scheme features the strong security guarantee of being based on the hardness of worst-case lattice problems. After presenting the construction of our scheme we prove its security and propose asymptotic parameters. Finally, we compare our scheme on several measures to one of the most efficient LWE-based encryption schemes with Gaussian noise. We show that the expected efficiency improvement is debunked, due to the large blow-up of the parameter sets involved.
The Module Learning With Errors (
M-LWE ) problem is a core computational assumption of lattice-based cryptography which offers an interesting trade-off between guaranteed security and concrete efficiency. The problem is parameterized by a secret distribution as well as an error distribution. There is a gap between the choices of those distributions for theoretical hardness results (standard formulation of
M-LWE , i.e., uniform secret modulo q and Gaussian error) and practical schemes (small bounded secret and error). In this work, we make progress toward narrowing this gap. More precisely, we prove that
M-LWE with uniform
η -bounded secret for any
1≤η≪q and Gaussian error, in both its search and decision variants, is at least as hard as the standard formulation of
M-LWE , provided that the module rank d is at least logarithmic in the ring degree n. We also prove that the search version of
M-LWE with large uniform secret and uniform
η -bounded error is at least as hard as the standard
M-LWE problem, if the number of samples m is close to the module rank d and with further restrictions on
η . The latter result can be extended to provide the hardness of search
M-LWE with uniform
η -bounded secret and error under specific parameter conditions. Overall, the results apply to all cyclotomic fields, but most of the intermediate results are proven in more general number fields.
In this paper we propose the first provably secure public key encryption scheme based on the Learning with Errors (LWE) problem, in which secrets and errors are sampled uniformly at random from a relatively small set rather than from the commonly used discrete Gaussian distribution. Using a uniform distribution, instead of a Gaussian, has the potential of improving computational efficiency a great deal due to its simplicity, thus making the scheme attractive for use in practice. At the same time our scheme features the strong security guarantee of being based on the hardness of worst-case lattice problems. After presenting the construction of our scheme we prove its security and propose asymptotic parameters. Finally, we compare our scheme on several measures to one of the most efficient LWE-based encryption schemes with Gaussian noise. We show that the expected efficiency improvement is debunked, due to the large blow-up of the parameter sets involved.
Since its proposal by Regev in 2005, the Learning With Errors (LWE) problem was used as the underlying problem for a great variety of schemes. Its applications are many-fold, reaching from basic and highly practical primitives like key exchange, public-key encryption, and signature schemes to very advanced solutions like fully homomorphic encryption, group signatures, and identity based encryption. One of the underlying reasons for this fertility is the flexibility with that LWE can be instantiated. Unfortunately, this comes at a cost: It makes selecting parameters for cryptographic applications complicated. When selecting parameters for a new LWE-based primitive, a researcher has to take the influence of several parameters on the efficiency of the scheme and the runtime of a variety of attacks into consideration. In fact, the missing trust in the concrete hardness of LWE is one of the main problems to overcome to bring LWE-based schemes to practice. This thesis aims at closing the gap between the theoretical knowledge of the hardness of LWE, and the concrete problem of selecting parameters for an LWE-based scheme. To this end, we analyze the existing methods to estimate the hardness of LWE, and introduce new estimation techniques where necessary. Afterwards, we show how to transfer this knowledge into instantiations that are at the same time secure and efficient. We show this process on three examples: - A highly optimized public-key encryption scheme for embedded devices that is based on a variant of Ring-LWE. - A practical signature scheme that served as the foundation of one of the best lattice-based signature schemes based on standard lattices. - An advanced public-key encryption scheme that enjoys the unique property of natural double hardness based on LWE instances similar to those used for fully homomorphic encryption.
The samples of Learning With Errors (LWE) are in the form (A, b = As + e), where e admits a discrete Gaussian that resides over
Zm, and
b∈Zqm. To reduce the size of b, this paper presents a ciphertext-efficient primitive called Learning With Sampling (LWS). Let
Q⊂Zm be a sublattice, the samples of LWS are (
A,b=As+eS), where
eS admits a discrete Gaussian that resides over the coset Q-As. Conditioned on the randomness of As, we show that the quantization error
eS amounts to the discrete Gaussian error e used in LWE, as long as the sampling width tailored to exceed the smoothing parameter of the lattice Q. Consequently, this ensures negligible difference between e and
eS. We provide computational and information-theoretic proofs to establish the hardness of search and decision variants of LWS. We demonstrate that LWS can effectively replace LWE in public key encryption (PKE), moderately reducing ciphertext size while maintaining security.