QLS Seminar - John Murray

Learning and Generalization through Neural Representations

John Murray, Yale University 
Tuesday November 29, 12-1pm
Zoom Link: https://mcgill.zoom.us/j/86855481591

Abstract: Efficient generalization is essential for learning in a complex and dynamic environment. In theories of neural computation and learning, neural representations adapt to task demands, yet in turn, representations shape inductive biases for generalization during task learning. In this talk I will present neural network modeling and behavioral analysis that explores this interplay between neural representations, inductive bias, and learning behavior.