Continuous Optimization Team Seminar (Talk by Andi Han, University of Sydney).

This is an online seminar. Registration is required.
【Continuous Optimization Team】
【Date】2023/May/19(Fri) 9:30-10:30(JST)
*【Speaker】Andi Han, University of Sydney *

Title: Towards unified frameworks and generic acceleration for Riemannian optimization
Abstract:
Learning over smooth nonlinear spaces has found wide applications in fields such as statistics, machine learning, engineering and neuroscience. A principled approach for addressing such problems is to endow the search space a Riemannian manifold geometry and numerical optimization can be performed intrinsically. Recent years have seen a surge of interest in leveraging Riemannian optimization for nonlinearly-constrained problems. In this talk, I will present two of my works with a goal towards unified frameworks and generic strategies for analyzing and improving Riemannian optimization. The first half of the talk explores how the choice of Riemannian geometries systematically impacts the algorithm convergence, with a focus on the positive definite matrices. The second half will introduce a generic acceleration technique for optimization on manifolds.