- About
- Courses
- Research
- Computational Social Science
- Critical Data Studies
- Data Science
- Economics and Information
- Education Technology
- Ethics, Law and Policy
- Human-Computer Interaction
- Human-Robot Interaction
- Incentives and Computation
- Infrastructure Studies
- Interface Design and Ubiquitous Computing
- Natural Language Processing
- Network Science
- Social Computing and Computer-supported Cooperative Work
- Technology and Equity
- People
- Career
- Undergraduate
- Info Sci Majors
- BA - Information Science (College of Arts & Sciences)
- BS - Information Science (CALS)
- BS - Information Science, Systems, and Technology
- MPS Early Credit Option
- Independent Research
- CPT Procedures
- Student Associations
- Undergraduate Minor in Info Sci
- Our Students and Alumni
- Graduation Info
- Contact Us
- Info Sci Majors
- Masters
- PHD
- Prospective PhD Students
- Admissions
- Degree Requirements and Curriculum
- Grad Student Orgs
- For Current PhDs
- Diversity and Inclusion
- Our Students and Alumni
- Graduation Info
- Program Contacts and Student Advising
IS PhD student Mashfiqui Rabbi will present on "MyBehavior: Automating Personalized Health Feedback Using a Multi-Arm Bandit Model" for this week's Brown Bag Seminar.
MyBehavior: Automating Personalized Health Feedback Using a Multi-Arm Bandit Model
In this work, we propose MyBehavior, a mobile application with a suggestion engine that learns a user’s physical activity and dietary behavior, and provides finely-tuned personalized suggestions. To our knowledge, MyBehavior is the first smartphone app to provide personalized health suggestions automatically, going beyond commonly used one-size-fits-all prescriptive approaches, or tailored interventions from healthcare professionals. MyBehavior uses an online multi-armed bandit model to automatically generate context-sensitive and personalized activity/food suggestions by learning the user’s actual behavior. The app continually adapts its suggestions by exploiting the most frequent healthy behaviors, while sometimes exploring non-frequent behaviors, in order to maximize the user’s chance of reaching a health goal (e.g. weight loss).
We evaluated MyBehavior with a three-week deployment and found that personalized suggestions were more effective and easier to incorporate in users’ daily lives, compared to its generic, prescriptive counterpart.
N.B., This work blends several ideas from machine learning, decision theory, psychological theories of behavior change, and context-aware computing. However, MyBehavior is currently at a very early formative stage, and your feedback is highly valuable for us to take the work forward.