Research
Interests
I am interested in just about any machine learning problem that is somewhat theoretical in nature. While I find probability and statistics fascinating in their own right, my research thus far has focused on applying tools and methods from these fields to get a better understanding of what we can guarantee about machine learning algorithms.
Current Projects
Currently I am working on algorithms for active and semisupervised learning.
Publications
Conference and Journal Papers

A New Perspective on Learning Linear Separators with Large \( L_q L_p \) Margins paper supp
M.F. Balcan and C. Berlind
Conference on Artificial Intelligence and Statistics (AISTATS), 2014, to appear.

Efficient Semisupervised and Active Learning of Disjunctions paper supp
M.F. Balcan, C. Berlind, S. Ehrlich, and Y. Liang
International Conference on Machine Learning (ICML), 2013.
Workshop Contributions

On Learning Linear Separators with Large \( L_\infty L_1 \) Margins poster
M.F. Balcan and C. Berlind
Workshop on Learning Faster from Easy Data
Advances in Neural Information Processing Systems (NIPS), 2013.