Justin N Wood, Associate Professor of Informatics at Indiana University Bloomington
When: Thursday, March 13, 2025, 2:00 PM - 3:00 PM
Where: Teams
Speaker: Justin N Wood who is an Associate Professor of Informatics, Indiana University
Title:
Digital Twin Studies Reveal Universal Learning Principles in Brains and Machines
Abstract:
What are the core learning mechanisms in brains? To address this, we introduce digital twin studies, in which newborn animals and artificial agents are raised in the same environments and tested on the same tasks, allowing direct comparisons of their learning abilities. Our findings reveal that generic learning models, such as transformers, acquire object and social skills similar to animals when reared (trained) under comparable developmental conditions. We propose that learning in humans, animals, and machines can be unified under a space-time fitting principle, where generic learning mechanisms adapt to the spatiotemporal structure of their environment. This framework offers a powerful lens for understanding the origins of intelligence across biological and artificial systems.
Bio:
Justin Wood is an Associate Professor of Informatics at Indiana University Bloomington. He is affiliated with programs in Animal Behavior, Neuroscience, Cognitive Science, and Psychological & Brain Sciences. He received his B.A. from University of Virginia and M.S. & Ph.D. from Harvard University. Dr. Wood has studied the psychological abilities of a range of populations, including human adults, babies, chimpanzees, wild monkeys, starlings, and newborn chicks. He started his lab at University of Southern California in 2008, before moving to the engineering school at Indiana University Bloomington in 2019.
Dr. Wood works at the intersection of four fields—developmental psychology, neuroscience, virtual reality, and artificial intelligence—to study the core learning mechanisms in brains. His lab created Newborn Embodied Turing Tests (NETTs), in which newborn animals and artificial agents are raised in the same environments and tested with the same tasks. NETTs are the first formalized benchmarks for testing computational models of newborn brains. His goal is to use NETTs to answer one of the great unsolved mysteries in science (What are the origins and computational basis of intelligence?) by linking the study of biological and artificial intelligence.
Microsoft Teams Need help?
Meeting ID: 366 973 870 403
Passcode: FU2Q8LC3
For organizers: Meeting options