Knowledge Drive Profiles Dynamic
Inferring user profiles is a research area of interest with applications in, e.g., recommendation systems, and cognitive rehabilitation. User profiles are being produced mostly from data mining and machine learning approaches, which brings the challenge of providing explanations about the creation, representation, and dynamics of user profiles. This project aims to suggest a novel knowledge-based approach for user profiling and its dynamics. Thus, we suggest the use of AGM-based belief revision to dynamic of profiles to (a) provide a formal representation of user profiles; (b) describe the changes in the profiles; (c) identify what caused the changes, and (d) return the sequence of changes that will transform an original profile into a target profile. The suggested approach will be tested in real-world applications. In this view, this project will provide a significant contribution to the Explainable Artificial Intelligence area, with applications in, e.g., cognitive rehabilitation, and human-computer interaction.
In the last two decades, user profiles have been used in several areas of information technology. We can mention its use in recommendation systems or adaptable user interfaces. In this project, we propose formalizing profiles’ creation, representation, and dynamics in a Knowledge-Driven perspective to improve their applications in the mentioned areas.
28/03/2021 - 28/03/2024
Fundação para a Ciência e Tecnologia
University of Madeira
Blending Interaction: Engineering Interactive Systems & Tools