Systems that learn from or personalize themselves to users are quickly becoming mainstream yet interaction with these system is limited and often uninformative for the end user. This talk focuses on approaches and challenges to making these systems transparent, controllable and ultimately trustworthy to end users.  In this talk, I will present work conducted as part of the TaskTracer system and ensuing work on how the reasoning of machine learning programs could be explained to end users, how end users could make corrections, and what challenges they face.

Bio:

Simone Stumpf received a PhD in Computer Science in 2001 and a BSc in Computer Science with Cognitive Science in 1996, both from University College London. She joined City University London as Lecturer in 2009. Previously, she worked at Oregon State University and University College London as a Research Manager/Fellow. Her research interests focus on intelligent user interfaces, end-user programming and personal information management. She is currently Co-PI on the NSF-funded "End-user debugging of machine-learned programs" project.