User Representation

This research area explores how to represent and protect user information in personalized systems while maintaining privacy and transparency. Key work includes developing frameworks for data minimization that learn to limit personal data collection based on performance curves and legal principles like GDPR compliance. The research addresses exposing query identification as a transparency challenge, developing methods to identify which searches expose specific content in ranking results. Additionally, the area encompasses natural language user profiles for recommendation systems, proposing more interpretable and scrutable approaches to representing user preferences that could reduce reliance on noisy implicit observations.