Welcome! I'm a (n-2010+1)'th year PhD Student in the Machine Learning Department at CMU, working with Alyosha Efros. I graduated from CMU in 2010 with a B.S. in computer science and cognitive science, with a minor in neural computation.
I'm interested in computer vision and all the learning problems that are
associated with it. In particular, I'm interested in transfer learning
(or something like it; I'm not entirely sure that term applies here). What I mean is, in
vision, some types of labels come cheaply: for example, GPS tags or
structure-from-motion and segmentation cues in some videos. For most of
the tasks we're interested in, though, labels are expensive. Actions or
objects, for example, must be annotated manually. Can we use the cheap
labels to get better generalization in vision problems where
labels are scarce?