Math Colloquium: Computer Science Job Talk - Mark Hopkins, Reed College

The Disruptive Power of Deep Learning - 

In 2012, an image recognition system from the University of Toronto startled the artificial intelligence community with unprecedented performance gains over then-state-of-the-art systems. This watershed event created a surge of interest in deep neural networks, a field of study known as "deep learning." In this talk, I will give a brief history of deep learning, provide a simple explanation of neural networks, and discuss how its sudden prominence has affected research in natural language processing. In particular, I will discuss how my research agenda has adapted to the new possibilities presented by deep learning.