Professor Tony Jebara helps develop a framework combining generative probabilistic models with discriminative Support Vector Machines, reconciling a key tension in machine learning.
