Application of Artificial Neural Network in Cryptography

In today's scenario where everything is just a click away, there is a huge concern for Intelligence Security. In this work, artificial neural network(ANN) has been explored to prevent any unaccredited access of data. Several kinds of cryptographic techniques are already in use to handle this problem. However, each of these techniques have their own advantages and limitations. Some techniques are more secure whereas some are less. In this work ANN is investigated in data security application by analyzing different types of semantic nets along with their architectures and compared. Further, scrutiny is done on the footing of their training algorithms by listing down their merits/demerits. A recurrent network is used to carry out the implementation of a finite state sequential machine which is used for cryptography purpose. Based on the analysis of obtained results, it is concluded that Artificial Neural Network can be effectively utilized as a contemporary technique to encrypt and decrypt a data.