How’s your experience? A blockchain-based framework for cookie sharing enabling adversarial diversity

February 22nd, 2022

Online recommender systems (RSs) are an integral element of the web used by around 4 billion people world wide. The ultimate goal of these systems is to help users find content, such as articles, videos, products, social media posts, etc. depending on the respective web service, that is in some form appealing or relevant to them.

Thesis Type Master
Student Michael Kretschmer
Status Finished
Proposal on 02/11/2021 2:00 pm
Proposal room Zoom
Presentation on 31/05/2022 1:00 pm
Presentation room Seminar room I5 6202
Supervisor(s) Ralf Klamma
Stefan Decker
Advisor(s) Alexander Neumann

This is especially helpful and partially required considering the sheer amount of content on the web.
Although RSs are ubiquitous online, the attitudes expressed towards them and the knowledge of them varies greatly among their users.
While the vast majority of users seems to appreciate RSs at least to some extent, some researchers have expressed concerns that personalized recommendations narrow the exposure to diverse content which may have negative side effects.
Thus, we study diversity in recommender systems in this work by developing a system which grants users direct control over the content recommendations they receive.