I am interested in machine learning and it's application to computer vision. My work focuses mainly on image representations, broadly defined as problems where the goal is to produce an image (or image-like object) as output. The difficulty of this style of problem in current practice lies in the inherently generative nature of the problem and the inherently discriminative nature of deep learning based machine learning.
I recently defended my dissertation, Learning and Evaluating Image Representations at the University of Illinois at Urbana Champaign for work done with David Forsyth. Prior to Illinois I received my BS from Rensselaer Polytechnic Institute where I worked with Chuck Stewart (RPI) and Scott Gallager (WHOI).