Public defence in Computer Science, M.Sc. Antonios Matakos
Opponent: Professor Jilles Vreeken, Saarland University, Germany
Custos: Professor Aristides Gionis, Aalto University School of Science, Department of Computer Science
Contact details of the doctoral student: [email protected]
The public defence will be organised on campus.
The thesis is publicly displayed 10 days before the defence in the publication archive Aaltodoc of Aalto University.
Public defence announcement:
Modern-day social media are governed by algorithms optimized towards maximizing user engagement. This makes said algorithms geared towards serving attention-grabbing content and content that conforms to the users' prior beliefs, which can lead to negative outcomes such as filter bubbles and polarization. A filter bubble, a term coined by Eli Pariser, is a state of ideological isolation fueled by social media algorithms. Several recent studies have attempted to detect and characterize the strength of filter bubbles and polarization. In this Thesis, we propose models that capture the presence of these negative phenomena. Additionally, we propose algorithms that aim to reduce polarization, a harmful effect, while on the other hand promoting the diversity of exposure to differing opinions, a beneficial effect. The algorithms focus on the optimization of specific functions that measure polarization, filter bubbles and diversity among others in a social network. We also propose algorithms to foster new connections in a network. We demonstrate the effectiveness our algorithms by carrying out experiments on real-world social network datasets, and evaluating their performance according to the proposed metrics.