Title: How are systems of knowledge built?

Presenter: Assistant Professor, Psychology and Neuroscience, University of Colorado Boulder
https://www.colorado.edu/lab/del/

Abstract: In both cognitive development and education, the end goal is to obtain structured, relational representations that afford broad generalization. But how do we get there? How effective is direct teaching of underlying principles? How much can we get from associative learning that is often criticized to be slow, incremental, and limited in generalizability? Focusing on children’s learning of the multi-digit number symbol system (but borrowing tools and inspiration from many other domains), I will show that, prior to formal education, many preschoolers are building early, implicit understanding of the symbol system, and that this early knowledge is highly predictive to later math learning. Using machine learning techniques to characterize the structure and growth of this knowledge shows regularities in age-related acquisitions, but also errors and idiosyncrasies that are not easily explained following a rule-learning trajectory. Further training studies and deep learning models confirm that associative processes, operating on highly structured input data, can lead to fast and far generalization with limited data. Results also highlight the crucial role that attention plays and the need to study the moment-by-moment processes that make learning happen. I will show an on-going head-mounted eye-tracker study that aims to elucidate how prior knowledge with number names can guide attention and memory during formal learning of explicit principles. I will end with future directions on how to use machine learning and other multi-sensory technologies (e.g., eye-tracking, VR headset) to further understand the mechanisms of human learning with implications for education.

Bio: Dr. Lei Yuan is an Assistant Professor in the Department of Psychology & Neuroscience at the University of Colorado Boulder. She received her Ph.D. in Cognitive Psychology at Northwestern University in 2016, followed by an NIH NSRA funded Postdoctoral fellowship at Indiana University. She joined CU Boulder in the Fall of 2020. She is the director of the DEL (Development, Education, and Learning) lab. https://www.colorado.edu/lab/del/ Her research examines the structure of input data in children's early learning environment, the processes and mechanisms through which children learn from this data, and how this learning creates hidden deficits or competencies for later school learning. To answer these questions, her lab combines large-scale cross-sectional and longitudinal studies, training experiments, computational modeling, high-density behavioral data collection (e.g., eye tracking), and translational research in schools.