Biometric and Implicit Authentication schemes on smart devices are gaining considerable attention in the research and industry. Traditional biometric schemes identify and validate a user based on what he is, whereas implicit authentication provides a mechanism to continuously authenticate device users by monitoring their interactions without any interruption or user inputs. A number of these techniques have been proposed by researchers and major tech-giant companies have deployed them as an attractive alternative to legacy password systems. However, the accuracy and efficiency of these traditional biometric and implicit authentication schemes have always been questioned as they demand high computing resources or continuous usage of sensors. In the most recent years, Machine Learning (ML)-based approaches have been proposed to overcome the abovementioned challenges. Through this project, we aim to provide a comprehensive overview of ML techniques proposed for biometric and implicit authentication. The main objectives of the project are:
In this project, student is expected to communicate with the CSE supervisor for the continuous guidance and supervision. There will be weekly meetings on the project progress with a supervisor. This project will provide an opportunity to a student to directly involve in state-of-the-art and emerging field of cybersecurity i.e., user authentication using AI, and learn ways to conduct research for real applications. We also prefer student to have strong technical writing skills to write a survey report.
From this project, student with the help of supervisor is expected to: