Abstract:

Statistical machine learning is a fast growing area, focused on 
automated detection of meaningful patterns in large and complex data sets.
Theoretical analysis has played a major role in some of the most prominent
practical successes in this field. However, our mainstream machine learning
theory assumes some strong simplifying assumptions which are often
unrealistic. In the past decade, the practice of machine learning has 
led to the development of various heuristic paradigms that answer the needs 
of a vastly growing range of applications. Many useful such paradigms fall 
beyond the scope of the currently available analysis, raising the need for 
major extensions of the common theoretical models.

In this talk, I will survey some of these application-motivated challenges.
In particular, I will discuss recent developments in the theoretical
analysis of semi-supervised learning, multi-task learning, "learning to
learn", privacy-preserving learning and more.