Speaker: Lindon Roberts (University of Sydney)

Title: Large-scale derivative-free optimization using random subspace methods

Synopsis: Many standard optimization algorithms require being able to cheaply and accurately compute derivatives for the objective and/or constraint functions. However, in the presence of noise, or computationally expensive or black-box procedures, derivative information may be inaccurate or impractical to compute. Derivative-Free Optimization (DFO) encompasses a variety of techniques for nonlinear optimization in the absence of derivatives. However, such techniques can struggle on large-scale problems for reasons including high linear algebra costs and strong dimension-dependency of worst-case complexity bounds. In this talk, I will discuss model-based and direct search DFO algorithms based on iterative searches in randomly drawn subspaces and show how these methods can be used to improve the scalability of DFO. This is joint work with Coralia Cartis (Oxford) and Clément Royer (Paris Dauphine-PSL).

Bio: Lindon Roberts is a lecturer at the University of Sydney’s School of Mathematics and Statistics. He received his DPhil in mathematics from the University of Oxford in 2019, and from 2019-2022 was an MSI Fellow at the Australian National University. His research interests are primarily in nonlinear optimization, particularly derivative-free optimization and applications in image analysis and machine learning. In 2021 he received the 2021 Leslie Fox Prize for Numerical Analysis.

WED 27 JULY 4PM – 5PM AEST MELBOURNE

ZOOM MEETING ID: 873 1557 5255; PASSWORD: 778635