Research On Data Encryption Technology Based On Adversarial Neural Network

With the continuous popularization of modern networks and the rapid development of communication technologies,digital technologies are widely used in various industries.However,because digital information is easy to copy and spread,there are many security problems.Therefore,the research on the security and protection technology of digital information has great theoretical significance and application value.Passwords are the basic means of protecting information security.Researchers have proposed a number of specific encryption algorithms,of which block cipher AES is a widely used encryption algorithm.In most cases,in order to make the encryption algorithm feasible,the design phase only considers the calculation of bounded attacking opponents.Under this big assumption,various security algorithms such as semantic security,defense against plaintext attacks,and defense against selective ciphertext attacks are proposed.The development of artificial intelligence has made many significant advances in recent years.In some cases,artificial intelligence can learn on its own and perform some tasks better than humans.This paper studies the security of an encryption algorithm against neural ciphers.By analyzing the idea of anti-neural network and introducing cryptographic anti-attack game theory,an anti-neural network encryption method based on ciphertext attack is proposed.Choosing ciphertext attack is a better attack method in cryptography.This paper introduces an attacker who chooses ciphertext attack mode to improve and improve the existing encryption algorithm against neural network.The simulation results of the algorithm are analyzed.It is found that the algorithm can learn the safe password generation algorithm without the need of artificial knowledge,and it has obvious advantages compared with DES,3DES,AES and other algorithms.