PhD Proposal Presentation
Date: Friday, December 9, 2016 at 10:00AM in Rice 204
Advisor: John A. Stankovic
Committee: Alfred Weaver (Committee Chair), Hongning Wang, Kai-Wei Chang and Laura Barnes (Systems and Information Engineering).
Towards Personalized Conflict Detection from Heterogeneous Health Applications
With the rapid advent of digital health care platforms, an increasing number of people are relying on health interventions/advice generated from heterogeneous sources to manage their health and wellness, such as mobile health apps, health websites, etc. Advice generated from these sources often adversely interact with each other or with activities of daily living based on users’ context, behavior and physiological conditions and thus result in conflicts. If this goes undetected, these conflicts can result in serious health risks. We propose a system to detect conflicts that can occur between (i) two or more textual health interventions and (ii) a textual health intervention and an activity of daily living. Our proposed system consists of novel natural language processing techniques and temporal activity modeling approaches to detect conflicts in health applications in a personalized and explanatory manner. We hypothesize that such a system will aid end users in informed decision making regarding health care. This proposal provides an overview of the conflict detection methods, the key research questions, an evaluation plan, and some preliminary results.