Abstract: 

Representation learning in neural nets continues to play a fundamental role in advancing our understanding of deep learning algorithms and our ability to extend successful applications. In this session we will explore how the information bottleneck analysis of deep learning algorithms sheds insight into how these algorithms learn and patterns across layers of learned representations. We conclude with discussion of how this analysis sheds a more practical light on theoretical concepts in deep learning research such as nuisance insensitivity and disentanglement.

Bio: 

Mike serves as Chief ML Scientist and Head of Machine Learning for SIG, UC Berkeley Data Science faculty, and Director of Phronesis ML Labs. He has led teams of Data Scientists in the bay area as Head of Data Science at Uber ATG, Chief Data Scientist for InterTrust and Takt, Director of Data Science for MetaScale/Sears, and CSO for Galvanize where he founded the galvanizeU-UNH accredited Masters in Data Science degree and oversaw the company's transformation from co-working space to Data Science organization. Mike began his career in academia serving as a mathematics teaching fellow for Columbia University before teaching at the University of Pittsburgh.