Joseph Jay Williams's Page

Profile Information

About Me: My research designs adaptive systems for online content, like intelligent lessons that emulate your favorite teacher in personalizing explanations. This computational cognitive science approach requires integrating my research in psychology & education, human-computer interaction, and statistical machine learning.
TEDxPortofSpain: Is the internet replacing teachers?

To make any static website become intelligently adaptive, I use powerful systems for randomized A/B experiments. Examples range from adaptive explanations for how to solve math problems, to self-personalizing emails to change people's behavior. My systems continually crowdsource new 'A' and 'B' designs from psychological scientists, using randomized comparisons to evaluate how helpful these alternative designs are, for people with different profiles. Algorithms from statistical machine learning use this data for real-time enhancement and personalization, by changing which designs are presented to future users. By enabling scientific discoveries from each user to directly impact engineering improvements for the next, my work provides intelligent content that never stops improving.

My TEDxPortofSpain talk at bit.ly/tedxwilliams explains how I use this approach in education, using MOOClets to intelligently adapt explanations for how to solve math problems.