Scaling Expertise via Language Models: With Applications to Education

Abstract‏‏: Access to expertise is essential for fostering effective interactions in many areas of society. For example, in education, experienced teachers teach students through effective interactions and train novices. However, access to expertise is often limited, undermining the training of novice educators and student outcomes. While language models offer the promise of democratizing access, they often mimic surface-level patterns and lack the human touch to keep students engaged when learning becomes difficult. In this talk, I will present two works that address these challenges by embedding expert-like thinking into language models and empowering human novices to perform at expert level in real-time interactions. First, I will discuss Bridge, an adaptation method that extracts latent expert reasoning to improve language models in complex interactions. Then, I will introduce Tutor CoPilot, a novel Human-AI approach that leverages a model of expert thinking to provide expert-like guidance to tutors in real time. In the first randomized controlled trial of a Human-AI system for live tutoring, Tutor CoPilot significantly improves the quality of learning interactions for 1,800 K-12 students and 900 tutors.

Bio: Rose E. Wang is a Computer Science PhD candidate at Stanford University. She develops machine learning and natural language processing methods to tackle challenges in real-world interactions, with a focus on Education. Her work is deployed in industry and directly improves the education of under-served students through partnerships she has cultivated during her Ph.D., including Title I school districts and several education companies, impacting 200,000+ students, 1,700+ teachers, 16,100+ tutors, in millions of tutoring sessions across the U.S., UK and India. Her work is recognized by NSF Graduate Research Fellowship, CogSci Best Paper Award, NeurIPS Cooperative AI Best Paper Award, ICLR Oral, Rising Star in Data Science, Building Educational Applications Ambassador Paper Award, and the Learning Engineering Tools Competition Award.

Zoom: https://mit.zoom.us/j/99311665529