The failure of a neural network for solving Boolean equations of algebraic attack

Abstract EN

The neural network represents an important method for solving several problems in many applications.

this paper will prove the failure of neural network with algebraic attack which makes it impossible to be used as a tool for solving the linear Boolean equations that generated after applying linearization on the nonlinear equation that generated from applied algebraic attack on the generators of stream cipher.

here when be used deferent types of neural networks such as perceptron , or Boolean neural network, will take long time without arrived to the correct solution (secret key for cipher system),compare with the mathematical method such as Gaussian elimination that given the correct solution with short time.