Nonsmooth Optimization on a Finer Scale - Distinguished Seminar in Optimization and Data

Explore the intricacies of nonsmooth optimization in this distinguished seminar featuring Jelena Diakonikolas from the University of Wisconsin-Madison. Delve into the challenges and opportunities presented by nonsmooth optimization problems, which are prevalent in industrial and machine learning applications. Learn about a novel approach to characterizing the complexity of these problems through the concept of local bounded variation of the (sub)gradient. Discover how this new perspective challenges previous beliefs about the parallelization of nonsmooth optimization problems, particularly for piecewise linear objectives with polynomially many pieces. Gain insights into the speaker's research on algorithm design and complexity analysis in large-scale continuous optimization, with a focus on machine learning applications. The seminar, titled "Nonsmooth Optimization on a Finer Scale," offers a comprehensive look at cutting-edge developments in the field of mathematical optimization.

Distinguished Seminar in Optimization and Data: Jelena Diakonikolas, University of Wisconsin-Madison