ARC Colloquium/ML Seminar series: Elad Hazan - Princeton University
(Please note that the talk will be held in MiRC 102 A & B and the refreshments will be served at the talk)
Projection-free Optimization and Online Learning
Modern large data sets prohibit any super-linear time operations. This motivates the study of iterative optimization algorithms with low complexity per iteration. The computational bottleneck in applying state-of-the-art iterative methods is many times the so-called "projection step".
We consider projection-free optimization/learning that replaces projections by more efficient linear optimization steps. We describe the first linearly-converging algorithm of this type for polyhedral sets and how it gives rise to optimal-rate stochastic optimization and online learning algorithms.