PhD Scholarship - Building an Aussie Information Recommendation System You Can Trust

This PhD scholarship focuses on efficient and scalable multi-group fairness for online information recommendation systems. The project addresses the challenge of privacy in online media platforms where users must share personal data to receive quality recommendation services. While various privacy mechanisms exist, current customizability is limited as users are restricted to predefined sets of sensitive attributes for protection. This research aims to revolutionize privacy mechanisms by granting users full control over their sensitive information. The project proposes a graph-based user representation learning paradigm that allows each user to disclose an incomplete yet distinct set of personal information while still learning expressive user representations for downstream recommendation tasks. This innovative approach meets users' varying privacy demands and responds to increasing public awareness around online privacy.