An Adaptive Approach for Active Multi-Factor Authentication

Multi-Factor Authentication (MFA) is the current trend to genuinely identify authorized users through the active authentication process using passwords, biometrics, cognitive behavior, etc. As new and improved authentication modalities of various types are becoming available, these are opening up options for security researchers to devise solutions facilitating continuous authentication to online systems. This paper focuses on describing a framework for continuous authentication where authentication modalities are selected adaptively by sensing the users’ operating environment (the device and communication media, and historical data). Empirical studies are conducted with varying environmental parameters and the performance of the adaptive MFA is compared with other selection strategies. The empirical results appear promising, which reflects that such a multi-factor decision support technique can be applied to realworld identity management and authentication systems. Keywords—Active Authentication; Genetic Algorithm; Adaptive Selection; Multi-Modal Framework