GeniusPhone: Extracting and Relating Relevant Personal Information João Guerreiro Instituto Superior Técnico joao.p.guerreiro@ist.utl.pt Tiago Guerreiro Instituto Superior Técnico tjvg@vimmi.inesc-id.pt Daniel Gonçalves Instituto Superior Técnico daniel.goncalves@inesc.pt Abstract Mobile devices have become faithful companions that keep track of most of our daily interactions and are always available to interact with. However, considering their capabilities, mobile devices play an insufficient role helping the user in his common daily tasks. The information they have access to is limited, as is the context in which it can be used. We present GeniusPhone, a system that inter-relates the users’ personal information and interactions with others from their computers and mobile devices, using it to gather additional data from online public sources. The information retrieved from the personal devices, due to its personal and trustable character, helps us filter the information retrieved from other less trustable and structured sources. It is able to provide the users with relevant summaries about someone or something, from their point of view, at the time they want or need them. 1. Introduction Mobile phones are widely used and have become essential tools for most of us. Moreover, they contain information about the users, their habits and daily interactions, as no other person or device. However, they are still of little use when the user needs to meaningfully access personal information, created both on the phone and in other devices, in a related, synergistic way. Nowadays, the usage of mobile device acquired knowledge is restricted to word dictionaries, recent contact recall or other somehow basic functionalities. Even projects that try to go beyond these basic functionalities are restricted to a limited set of information (ex: [1, 2]), ignoring the enormous amount of personal and public information possibly available on and from a mobile device and its inter- relation. What all the past and present approaches lack is a usage of personal information, from different sources, taking advantage of all the knowledge the devices have about their owners. Further, little attempt is made in trying to combine such information with that from public sources. In a social environment it is natural to wonder "I know that person, but where from?" or "I had some things to discuss with Jack, but what were they?”. GeniusPhone is able to provide answers to those questions by gathering and inter- relating information from the user’s devices enriched with other public information sources, to offer the user context- and personally-sensitive information when it is needed. All this information is interconnected within a semantic network providing a structured and meaningful knowledge base, dealing with the information from different sources as a coherent whole and being able to recognize the information that defines a person, theme or event from the user’s point of view. 2. GeniusPhone GeniusPhone manages the users’ personal information and their interactions with others in a mobile context, to obtain relevant and timely information about some person, document or subject. All the data existent in mobile devices, together with other personal and public information can provide insights on the user and surrounding context. The personal information can be used to filter data from other less trustable sources thus reducing the search universe and resolving ambiguities. At the same time, it guarantees that the resulting information is gathered from the user context and point of view. The world- wide-web is a “universe” of information, so it is crucial we can find and establish some standards to find where the useful information is. The GeniusPhone platform, is able to collect and interrelate personal information from the user’s devices (mobile devices and personal computers) like documents and their metadata, emails (and the attachments therein), calls, web pages, SMS, agenda, call logs, etc., and inter-relate it as a consistent whole. It can then, at the users’ request, find relevant data about a particular subject or person by looking not only at that personal information, but also at online sources