Multi-Objective GPS Optimization; For You Page or For Corporate Profits: How Recommender Systems are Harmfully Designed

Jayakumar, Pawan, School of Engineering and Applied Science, University of Virginia
Vrugtman, Rosanne, EN-Comp Science Dept, University of Virginia
Wylie, Caitlin, EN-Engineering and Society, University of Virginia

Every major free software service such as search engines and social media are centered
around the production and consumption of user specific data. These services generate value by
collecting user interaction data which can provide personalized experiences such as targeted ads
or media. The manner in which this data is collected has ethical implications. My technical
research problem investigated how we can use personalized user location data in order to reduce
the energy consumed by their phone’s GPS sensor. My STS research problem investigated how
recommender systems that are powered by personalized user data can harm the user. Together,
they show how personal data can be used for both beneficial and harmful purposes.

BS (Bachelor of Science)
Recomender Systems, Multi-Objective Reinforcement Learning, GPS Sensor Optimization

School of Engineering and Applied Science

Bachelor of Science in Computer Science

Technical Advisor: Rosanne Vrugtman

STS Advisor: Caitlin D. Wylie

Technical Team Members: Pawan Jayakumar,

All rights reserved (no additional license for public reuse)
Issued Date: