A grad student at the University of Wisconsin, interested in optimization and passionate about skiing!
Optimization, adaptive sampling theory, learning theory, large scale machine learning, data science
What if we have to collect human responses for some machine learning task? We’d like to pay for as little as possible, so we’d like this to be active. Connecting human responses to active learning algorithms is difficult in part because model updates have to run quickly. How quickly can these updates be performed with a large computer equipped with much memory and many cores?
Also, there’s a great talk by Prof. Jarvis Haupt on adaptive sampling that illustrates some of my interest. There’s also a Wired article about compressed sensing and images having many areas of the same color. How do you best “fill in the blanks?”
My current “course of life” or CV is available here.
University of Wisconsin-Madison
University of Minnesota
Major: Electrical engineering.
2010-2015 (graduated cum laude).
I worked with Prof. Jarvis Haupt to develop several projects, including an iPhone app, thermal camera and sign language recognition system.