Sunday, July 9th – Afternoon

Room 1.39

14:00 – 14:15 Welcome and workshop intro
14:15 – 14:40 Graph Embedding Based Recommendation Techniques on the Knowledge Graph László Grad-Gyenge, Attila Kiss and Peter Filzmoser
14:40 – 15:05 A Framework for Computational Serendipity Xi Niu and Fakhri Abbas
15:05 – 15:30 Distance- and Rank-based Music Mainstreaminess Measurement Markus Schedl and Christine Bauer
16:00 – 16:25 Introducing Surprise and Opposition by Design in Recommender Systems Christine Bauer and Markus Schedl
16:25 – 16:40 Position Talk: The Media Bubbles Project László Grad-Gyenge
16:40 – 17:20 Keynote: Diverse by Design: Considering the Human in Diversity-Aware Systems  Nava Tintarev
17:20 – 17:30 Discussion and Closing

Keynote by Nava Tintarev

Diverse by Design: Considering the Human in Diversity-Aware Systems

Abstract: Recommender systems find relevant content for us online, including the personalized news that we increasingly receive on Twitter and Facebook. As a consequence of personalization we increasingly see content that agrees with our views and we cease to be exposed to views contrary to our own. Both algorithms and users filter content, and this creates more polarized points of view, so called “filter bubbles”. This talk considers a solution to this problem using interaction paradigms, and methods for generating explanations (both text and graphics). Such explanations can be used to improve trust and transparency, but they can also be used to improve the discovery of novel content and help users identify their own blindspots. This talk introduces a vision of information presentation strategies which help users accept surprising content, and present opposing content in a way that is acceptable to them. In doing so, diversity-aware design solutions can address both user and algorithmic biases, and have a greater potential to decrease polarization of views.

Speaker Bio: Nava is an Assistant Professor and Delft Technology Fellow in the Web Information Systems group, Faculty of Electrical Engineering, Mathematics and Computer Science at TU Delft. She studies how to best present and adapt the presentation of complex data (using both natural language generation, and visualizations) in artificial advice giving systems. Nava was previously an assistant professor at Bournemouth University (UK), a research fellow at Aberdeen University (UK), and a research engineer for Telefonica Research (Spain). She is interested in tackling issues regarding ethics in big data, algorithmic transparency, fake news, and filter bubbles.