Digital Avatars: A programming framework for personalized human interactions through virtual profiles

Technology is evolving in the direction of making the users a key part in the model, developing services that adapt to their preferences and needs in a seamless way. So, information and knowledge about people has become one of the main actives for the IT industry. However, only a few major companies are presently able to take advantage of the digital footprint everyone leaves while performing their daily activities. Moreover, people cannot even decide when, how, and by whom the contents and information they are producing are used. In this work, we present the Digital Avatars framework for mobile-based collaborative social computing applications as a realization of People as a Service, a mobile computing model that empowers smartphones with the capability of learning virtual profiles or digital avatars of their owners and sharing them as a service in a controlled and seamless way. Digital Avatars offers services for third party applications, including an inference engine and a module for run-time scripts execution for interacting with the digital avatar in the smartphone assuring the privacy and full control of the users over their data. In this paper, we describe the architecture of this framework and the structure of the digital avatars focusing on an independent living scenario for Mild Cognitive Impairment patients.