Continuous User Authentication Based on Deep Neural Networks

Abstract

A user authentication method consists of a username, password, or any other related credential. These methods are mostly used only once to validate the user’s identity at the start of session. However, one-time verification of user’s identity is not resilient enough to provide adequate security all over the session. Such authentication methods should be adopted which can continuously verify that only genuine user is using the system resources for entire session. This research work has implemented a true continuous authentication system, based on keystroke dynamics, which tends to validate the user on each action by using the proposed robust recurrent confidence model(R-RCM). Moreover, the recurrent neural network(RNN) has been used to exploit the sequential nature of keystroke data. System has been tested with two experimental approaches and results are reported in mean genuine actions (ANGA) and imposter actions (ANIA).