Andrej Risteski works in the intersection of machine learning and theoretical computer science, with the primary goal of designing provable and practical algorithms for problems arising in machine learning. Broadly, this includes tasks like clustering, maximum likelihood estimation, inference, and learning generative models. All of these tend to be non-convex in nature and intractable in general. However, in practice, a plethora of heuristics like gradient descent, alternating minimization, convex relaxations, and variational methods work reasonably well. In his research, Risteski tries to understand what the realistic conditions are under which we can give guarantees of the performance of these algorithms, as well as devise new, more efficient methods.
- Foundations of Machine Learning, Spring 2017. Visiting Graduate Student.