We want to invite to a lunch seminar with Thomas E. Kolb, PhD candidate from TU Wien (Austria) and member of the CDL-RecSys.

Thomas is visiting MediaFutures for two months (until November 14) as part of an Erasmus traineeship. His research focuses on news recommender systems, particularly on fairness and bias over time.

Friday, 24th of October, Thomas will present his recent work and share insights from his ongoing research in this area.

Bio:

Thomas is conducting research as part of his Ph.D. on the subject of long-term dynamics of bias and fairness in cross-domain recommender Systems. To analyse these dynamics in a real world environment his lab works together with a company within the domain of news, books and lifestyle. The exploration of long-term dynamics in this field has immense potential for the development of fairer recommender systems. He firmly believes in the significance of providing the research community with fresh insights to foster the creation of responsible and fair recommender systems.

Abstract:

Recommender systems have become a key technology in digital media environments, yet their success cannot be measured by accuracy alone. In this talk, Thomas E. Kolb will first provide an overview of the lab’s current research activities across domains such as, e-commerce, fashion, and news. He will then present his past and current work on evaluating and designing recommender systems from a beyond-accuracy perspective, including insights on what makes up a “good reading recommendation” in news contexts based on the lab’s industry collaborations. The talk concludes with an outlook on recent trends in conversational and generative recommender systems, based on insights from his tutorial at the ACM Recommender Systems Conference.