An improved integral distinguisher scheme based on neural networks

Zahednejad, Behnam; Lyu, Lijun

doi:10.1002/int.22895

“…With regards to the cryptanalysis of Speck and Simon, other works followed. Zahednejhad et al [25] applied the DL methodology proposed by Gohr to construct a neural-based integral distinguisher scheme for several block ciphers, including Speck32, Present, RECT-ANGLE, and LBlock. The neural-based integral distinguisher increased the number of distinguished rounds of most block ciphers by at least one round, when compared to the state-of-the-art integral distinguishing method.…”

Section: Previous Workmentioning

confidence: 99%

On Pseudorandomness and Deep Learning: A Case Study

Ansaroudi

¹

,

Zaccagnino

²

,

D’Arco

³

2023

Applied Sciences

0

View full text Add to dashboard Cite

Pseudorandomness is a crucial property that the designers of cryptographic primitives aim to achieve. It is also a key requirement in the calls for proposals of new primitives, as in the case of block ciphers. Therefore, the assessment of the property is an important issue to deal with. Currently, an interesting research line is the understanding of how powerful machine learning methods are in distinguishing pseudorandom objects from truly random objects. Moving along such a research line, in this paper a deep learning-based pseudorandom distinguisher is developed and trained for two well-known lightweight ciphers, Speck and Simon. Specifically, the distinguisher exploits a convolutional Siamese network for distinguishing the outputs of these ciphers from random sequences. Experiments with different instances of Speck and Simon show that the proposed distinguisher highly able to distinguish between the two types of sequences, with an average accuracy of 99.5% for Speck and 99.6% for Simon. Hence, the proposed method could significantly impact the security of these cryptographic primitives and of the applications in which they are used.

show abstract

“…The findings show that in the DES cipher, a neural network can distinguish the XOR distribution of a linear expression. Other attacks, such as integral, have also been studied in connection with machine learning [6].…”

Section: Related Work Of Use Machine Learning In Cryptanalysismentioning

confidence: 99%

Automated Deep Learning BLACK-BOX Attack for Multimedia P-BOX Security Assessment

Tolba

¹

,

Derdour

²

,

Ferrag

³

et al. 2022

View full text Add to dashboard Cite

Resistance to differential cryptanalysis is a fundamental security requirement for symmetric block ciphers, and recently, deep learning has attracted the interest of cryptography experts, particularly in the field of block cipher cryptanalysis, where the bulk of these studies are differential distinguisher based black-box attacks. This paper provides a deep learning-based decryptor technique for investigating the permutation primitives used in multimedia block cipher encryption algorithms.We aim to investigate how deep learning can be used to improve on previous classical works by employing cipher text pair aspects to maximize information extraction with low-data constraints by using convolution neural network features to discover the correlation among permutable atoms to extract the plain text from the ciphered text without any P-box expertise. The evaluation of testing methods has been conceptualized as a regression task in which neural networks are supervised using a variety of parameters such as variations between input and output, number of iterations, and P-box generation patterns. On the other hand, the transfer learning skills demonstrated in this study indicate that discovering suitable testing models from the ground is also achievable using our model with optimum prior cryptographic expertise, where we contribute the results of deep learning in the field of deep learning based differential cryptanalysis development.Various experiments were performed on discrete and continuous chaotic and non-chaotic permutation patterns, and the best-performing model had an MSE of 1.8217e −04 and an R 2 of 1, demonstrating the practicality of the suggested technique.

show abstract

Improved integral neural distinguisher model for lightweight cipher PRESENT

Wu,

Guo

2024

Cybersecurity

0

View full text Add to dashboard Cite

PRESENT is a lightweight block cipher, which has attracted many scholars to research its security. In 2022, Zahednejad et al. proposed the integral neural distinguisher on round-reduced PRESENT. In this paper, a new integral neural distinguisher for PRESENT is constructed. In contrast to Zahednejad’s works, the proposed integral neural distinguisher can improve the number of attack rounds in one round. This paper proposes a new data format ($$invP_{0}^{n},invP_{1}^{n},\ldots ,invP_{15}^{n},invS_{0}^{n},invS_{1}^{n},\ldots ,invS_{15}^{n}$$ i n v P 0 n , i n v P 1 n , … , i n v P 15 n , i n v S 0 n , i n v S 1 n , … , i n v S 15 n ), which can exposes more features of PRESENT previous round ciphertext. Simultaneously, this paper incorporates MBConv module into the convolutional layers of DenseNet, which enable the neural network to identify a greater variety of features in ciphertext. The data format of the paper is analysed. The results of the analysis show that the data format in this paper is able to identify more features compared to Zahednejad’s data format. Further to this, experiments are performed on PRESENT using the integral neural distinguisher in this paper. The experimental results show that the changes make to the neural network and data format have improved the accuracy of distinguishers. Finally, key recovery attacks are conducted on the integral neural distinguishers of SmallPresent-(8) to demonstrate the efficacy of the distinguisher proposed in this paper. The results demonstrate that the key recovery success rates for 5-round and 6-round are 98% and 90%, considering error bits within a range of two bits.

show abstract

An improved integral distinguisher scheme based on neural networks

Cited by 6 publications

References 61 publications

On Pseudorandomness and Deep Learning: A Case Study

On Pseudorandomness and Deep Learning: A Case Study

Automated Deep Learning BLACK-BOX Attack for Multimedia P-BOX Security Assessment

Improved integral neural distinguisher model for lightweight cipher PRESENT

Contact Info

Product

Resources

About