Deep Learning and the Brain

Andrew Saxe

Stanford University, Stanford, CA, USA

Abstract

Deep learning systems rely on many layers of processing to perform sensory processing tasks like visual object recognition, speech recognition, and natural language processing (Bengio & LeCun, 2007). By learning simpler features in lower layers, and composing these into more complex features in higher layers, deep networks take advantage of the compositional structure of many real world domains and have realized impressive performance in a range of engineering applications, from visual object classification (Krizhevsky, Sutskever, & Hinton 2013; Ciresan, Meier, & Schmidhuber, 2012) to speech recognition (Mohamed, Dahl, & Hinton, 2012) and natural language processing (Collobert & Weston, 2008). How might deep learning models illuminate phenomena of interest to cognitive scientists such as perceptual learning, language acquisition, cognitive development, and category formation? How does depth impact both the dynamics of learning in a neural network, and the content of what is learned? This workshop will explore the implications of deep learning for our understanding of the brain and mind through a series of invited talks from established researchers in the field.

Files

Deep Learning and the Brain (55 KB)