Classroom observations play a central role in education evaluation and improvement efforts. However, human-based classroom observations are costly, difficult to scale, and prone to a range of measurement challenges. The rapid development of Natural Language Processing (NLP) techniques provides a potentially transformative approach for instructional measurement and feedback. This talk will introduce M-Powering Teachers (MPT), an innovative system that uses NLP to measure high-leverage teaching practices and deliver high-quality, automated feedback to teachers. Through a range of randomized controlled trials, MPT has demonstrated promising benefits for instructional quality and student outcomes in both virtual and in-person learning contexts. The talk will also discuss several major technical and practical challenges that need to be addressed before MPT can achieve its full potential.
Dr. Jing Liu is an Assistant Professor in Education Policy at the University of Maryland, where he also holds affiliations with the Value-Centered Artificial Intelligence and the Human-Computer Interaction Lab. He earned his Ph.D. in the Economics of Education from Stanford University and also finished a two-year postdoc at Brown University. His research focuses on the development, application, and evaluation of innovative AI tools to support educators and enhance learning opportunities for historically marginalized students. He has published in leading peer-reviewed outlets across disciplines, including economics, public policy, education, and computer science. His research has been funded by the National Science Foundation, the Gates Foundation, and the Spencer Foundation, among others, and covered by national media outlets such as the New York Times, NBC News, and Education Week.