Password Authentication Using Context-Sensitive Associative Memory Neural Networks: A Novel Approach

Passwords are the most widely used form of authentication. In many systems the passwords, on the host itself, are not stored as plain text but are encrypted. However, conventional cryptography based encryption methods are having their own limitations, either in terms of complexity or in terms of efficiency. The conventional verification table approach has significant drawbacks and storing passwords in password table is one of the drawbacks.

In the present paper, we propose a cognitive neural model using Context-Sensitive Associative Memory Model(CSAM) for password authentication, which is derived from cognitive domain and vector logic. According to the model, the product of two vectors is an associative memory(context-dependent) that plays critical role in the neural networks domain. In this model the output (encrypted password) is associated with the Kronecker Product of an input (key) and a context (password). The encrypted password is decoded with key and the context-dependent memory (Krnocker product) to get the original password. The proposed system provides better accuracy and quicker response time to authenticate the password but this model requires more space for holding context-dependent associative memory.