Designing adaptive user technologies through automating collaborative experimentation: Coordinating practitioners and researchers in HCI, psychology, education, and statistical machine learning

Abstract:

Understanding people’s complex real-world thinking is a challenge for the behavioral sciences, while computational scientists aim to build systems that can behave intelligently in the real-world. This talk presents a framework for redesigning the everyday websites people interact with to function as: (1) Micro-laboratories for psychological experimentation and data collection, (2) Intelligent adaptive agents that implement machine learning algorithms to dynamically discover how to optimize and personalize people’s learning and reasoning. I present an example of how this framework is used to create “MOOClets” that embed randomized experiments into real-world online educational contexts – like learning to solve math problems. Explanations (and experimental conditions) are crowdsourced from learners, teachers and scientists. Dynamically changing randomized experiments compare the learning benefits of these explanations in vivo with users, continually adding new conditions as new explanations are contributed. Reinforcement learning algorithms are used for real-time analysis of the effect of explanations on users’ learning, and optimization of the policies for delivering explanations to provide the explanations that are best for different learners. The framework enables a broad range of algorithms for multi-armed bandits to discover how to optimize and personalize users’ behavior, and dynamically adapt technology components to trade off experimentation (exploration) with helping users (exploitation).

The talk can be viewed online (or a recording seen afterwards) at the URL tiny.cc/williamstalk

Bio:

Joseph Jay Williams investigates intelligent adaptive technologies for personalized learning. His research bridges human-computer interaction and computational cognitive science: drawing on statistical machine learning, psychology, and education. He designs and uses real world online lessons to enable randomized experiments to discover how to personalize learning, and embeds machine learning/AI algorithms for real-time causal discovery and personalization.

He is a Research Fellow at Harvard VPAL Research, the office for online learning research and development. He is also a member of the Intelligent Interactive Systems Group in Harvard Computer Science, and leading the advisory board for an NSF Cyberinfrastructure grant to Neil Heffernan at WPI to enable psychology, education, and machine learning researchers to embed experiments in the ASSISTments online mathematics platform. He completed a postdoc at Stanford University in the Graduate School of Education working with the Office of the Vice Provost for Online Learning and Candace Thille's Open Learning

Initiative. He received his PhD in 2013 doing Computational Cognitive Science in UC Berkeley's Psychology Department. As part of the Concepts and Cognition Lab he investigated why prompting people to explain "why?" helps reasoning, and in the Computational Cognitive Science Lab developed models of reasoning, decision-making and learning using Bayesian statistics and machine learning. He is originally from Trinidad and Tobago. More information about his research and papers is at www.josephjaywilliams.com/research-overview.