Tuesday, April 9, 2024
10:00am-12:00pm
RS208, Rosebrugh Building,
164 College St.
Ilya Musabirov, University of Toronto
Mock Lecture (50 Mins): A Technical Introduction to Large Language Models
Technical Talk (35 Mins) & Q&A (15 Mins): Platform-based Adaptive Experimental Research in Education: Lessons from the XPRIZE Digital Learning Challenge
In this presentation, I share insights and lessons from the design and multi-experimental evaluation of an adaptive experimentation platform conducted within the XPRIZE Digital Learning Challenge framework. The discussion includes cross-platform software design that supports rapid integration and deployment of adaptive experiments, alongside five systematic replications achieved within a 30-day period. I will outline key scenarios where platform-supported adaptive experiments are applicable and share reflections on lessons learned from this two-year project, aimed at assisting researchers and practitioners in integrating adaptive experiments into real-world courses. Additionally, I will explore system design considerations, the role of cloud computing in accelerating platform-based adaptive educational research, the challenges of learning from heterogeneous real-world data, and the potential for integrating online and offline machine learning models.
Bio: Ilya Musabirov is a PhD student in Computer Science at the University of Toronto (UofT), where he also teaches courses in Data Science and Computing.
He is broadly interested in approaches to teaching computer and data science, making the best use of blended learning environments.
Before coming to Toronto, he designed data science and computer science courses and programs at HSE University, St. Petersburg, Russia. His goal was to help them build solid research and analytical skills. He proudly has graduates who combine their domain expertise and data science skills across various areas in academia and industry — from Digital Humanities to Data Engineering — and countries, from Norway to Australia.
At UofT, his work focuses on building platforms that make educational and behavioural interventions adaptive, aiming to make learning more engaging and effective for students and leveraging these platforms in the large-scale improvement of learning experience in the real world. He was a leading graduate student in the Adaptive Experimentation Accelerator team that won the Grand Prize of the XPRIZE Digital Learning Challenge in 2021-23 and a vital member of the QuickTA project supported by DARPA AI Tools for Adult Learning Transformation Prize in 2022-24.
In this work, he relies on bringing to Learning Engineering and HCI best practices from field behavioural interventions. His research focuses on developing tools for data-driven decision-making using computational interaction methods: Bayesian models, multi-armed bandits, data visualization, and optimization.
Evaluation Questions link: https://forms.office.com/r/yJwZ3HWg2R
