Supporting Conference Attendees with Visual Decision Making Interfaces

Abstract

Recent efforts in recommender systems research focus increasingly on human factors affecting recommendation acceptance, such as transparency and user control. In this paper, we present IntersectionExplorer, a scalable visualization to interleave the output of several recommender engines with user-contributed relevance information, such as bookmarks and tags. Two user studies at conferences indicate that this approach is well suited for technical audiences in smaller venues, and allowed the identification of applicability limitations for less technical audiences attending larger events.

Can I believe what I see? Data visualization and trust in the humanities

“…Edwards and Veale (2017, 81) point out that machine learning explanations are conditioned by the type of user: any explanation needs to be usable by its audience. Developing a visualization for recommender-system results that reveals some of the origins of its decisions, Verbert et al (2016) found that users, as one might hope, place greater trust in explained than in unexplained results.…”