Skip to Content

Search: {{$root.lsaSearchQuery.q}}, Page {{$root.page}}

Statistics Department Seminar Series: Joseph Jay Williams, Assistant Professor in Computer Science, University of Toronto

"Combining Reinforcement Learning, Psychology & Human Computation for Randomized A/B Experimentation: Perpetually Enhancing and Personalizing User Interfaces"
Tuesday, April 23, 2019
1:00-2:00 PM
340 West Hall Map
How can we transform everyday technology interfaces into intelligent, self-improving systems?Examples in education and health include: enhancing the explanations students receive in online quizzes, personalizing text messages to get people to exercise, and improving the reflective questions an app poses to help people manage stress and mental health. Combining human-computer interaction with psychology and statistical machine learning, my group conducts randomized "A/B" experiments and uses reinforcement learning algorithms to discover which conditions work best for different subgroups of people, automatically enhancing and personalizing future interfaces. I present an example system, which crowd sourced explanations for how to solve math problems from students and teachers, simultaneously conducting an A/B experiment to identify which explanations other students rated as being helpful. Modeling this as a multi-armed bandit where the arms were constantly increasing (every time a new explanation was crowd sourced) we used Thompson Sampling to do real-time analysis data from the experiment, providing higher rated explanations to future students (LAS 2016, CHI 2018). This generated explanations that helped learning as much as those of a real instructor. I consider future collaborative work with statistics and machine learning researchers, where we provide a testbed for evaluating a range of algorithms (e.g. for bandit and reinforcement learning problems) that can tackle exploration-exploitation tradeoffs in dynamically adapting randomized experiments to balance discovery against rapid enhancement and personalization of technology.

www.josephjaywilliams.com has links to slides and a recording of a related talk.

Bio:
Joseph Jay Williams is an Assistant Professor in Computer Science at the University of Toronto. He was previously an Assistant Professor at the National University of Singapore's School of Computing in the department of Information Systems & Analytics, a Research Fellow at Harvard's Office of the Vice Provost for Advances in Learning, and a member of the Intelligent Interactive Systems Group in Computer Science. He completed a postdoc at Stanford University in Summer 2014, working with the Office of the Vice Provost for Online Learning and the Open Learning Initiative. He received his PhD from UC Berkeley in Computational Cognitive Science, where he combined experimental psychology with Bayesian statistics and machine learning to investigate how people learn and reason. He received his B.Sc. from University of Toronto in Cognitive Science, Artificial Intelligence and Mathematics, and is originally from Trinidad and Tobago.
Building: West Hall
Website:
Event Type: Lecture / Discussion
Tags: Electrical Engineering and Computer Science, Lecture, Mathematics, Research, statistics
Source: Happening @ Michigan from Department of Statistics