Toward Inclusive and Personalized Learning with Large Language Models


Large Language Models (LLMs) have surfaced as a pivotal technology in enhancing educational experiences, marking a significant step forward in the field of digital learning. Despite their considerable advancements, these models exhibit a critical shortfall in inclusivity, predominantly displaying an anglo-centric bias that limits their capacity for truly personalized learning. In this presentation, I will delineate the prevailing landscape of leading LLMs, placing a particular emphasis on their fundamental shortcomings. Specifically, I aim to shed light on the obstacles that preclude their comprehensive incorporation into educational systems designed to universalize access to knowledge and customize learning trajectories to meet individual learner requirements. Following this, I will introduce several research initiatives undertaken by my team, which are directed towards mitigating these impediments, thereby contributing to the broader effort of making LLMs more inclusive and capable of supporting a diverse array of learning needs.

Wednesday, May 15, 2024

11:00 – 11:30 AM ADT

Dr. Muhammad Abdul-Mageed
Co-Lead, Computational Modelling Theme
Associate Professor, School of Information and Department of Linguistics
University of British Columbia