BehavAuth platform. The first core component of the BehavAuth system is the cloud platform. This component of the system implements the machine-learning and artificial intelligence models using sensor and device usage data from users’ smartphones. The models are trained initially and then adapt based on the contextual information in which the user is. To ensure the validity of the user at the initialisation process, the BehavAuth platform communicates only with the third- party platform, which has previously established a secure communication with the mobile app. Authentication tokens are used between the third-party platform and mobile app, to ensure the secure communication and validation among the two components. Similar secure communication is established between the two platforms. The main functionalities that the BehavAuth platform will provide are: • User registration. Users are registered at the BehavAuth platform through the third- party platform. The latter uses a pseudonym, an application key and the behavioural data to register a user at the BehavAuth platform. During this registration, an initialisation process is performed, the appropriate entities are created, and the training process of the various machine-learning models is initiated. • User authentication. Given that a user has been registered at the BehavAuth platform, an authentication service will be exposed, that will allow users to authenticate to the third-party platform. The various machine-learning models will receive the behavioural data of the user and will perform validation. A response will be sent to the third-party platform indicating the result i.e. if the user is an imposter or not. In case this is the actual user, the biometric data will be fed to the machine- learning models in order to implement the adaptation based on the contextual data.
