Chosen-Plaintext Cryptanalysis of a Clipped-Neural-Network-Based Chaotic Cipher

TL;DR

This paper demonstrates that a chaotic cipher based on a clipped neural network can be compromised through a chosen-plaintext attack, highlighting security vulnerabilities in the cipher's design.

Contribution

It provides the first cryptanalysis of a neural-network-based chaotic cipher, revealing its susceptibility to chosen-plaintext attacks.

Findings

The cipher can be broken with a feasible chosen-plaintext attack.

Experimental results confirm the attack's effectiveness.

Highlights security flaws in neural-network-based chaotic encryption schemes.

Abstract

In ISNN'04, a novel symmetric cipher was proposed, by combining a chaotic signal and a clipped neural network (CNN) for encryption. The present paper analyzes the security of this chaotic cipher against chosen-plaintext attacks, and points out that this cipher can be broken by a chosen-plaintext attack. Experimental analyses are given to support the feasibility of the proposed attack.