FYS 30 The Optimized Life: Deep Principles for Human Learning
Even though we do it every day, recognizing a face isn’t as easy as it seems. Different hairstyles or a sideways angle fooled our best algorithms for decades. But now, sufficiently large neural networks learn to recognize faces at timescales that starkly contrast the millennia evolution took in sculpting our brains. Indeed, deep learning has demonstrated that many sophisticated behaviors and neural codes actually emerge naturally from a limited set of principles. Students in this course will uncover these principles through reading literature and example code, and draw on them to generate hypotheses about optimizing learning in humans. No experience is required to succeed in this course but a strong interest in cognitive science, computer science, mathematics and/or statistics will be beneficial.
Prerequisite
Open only to first year frosh