Seminar: Optimization in the Presence of Noise
Department of Industrial Engineering
Optimization in the Presence of Noise
We consider optimization problems with noisy objective and/or constraint functions. Such problems arise in a variety of applications including machine learning. We present a few examples with different sources and models of noise. As solution methods, we focus on extensions of gradient-based algorithms (such as gradient descent and sequential quadratic programming) that employ approximate derivatives. We discuss the effect of the noise in function and derivative evaluations on the solution processes and convergence results of those algorithms. Throughout the talk, we provide a brief overview of relevant recent research and open questions.
Figen Oztoprak received her Ph.D. in Industrial Engineering from Sabanci University in 2011. She worked as a predoc and then as a postdoc in projects on nonlinear programming, optimization for machine learning, and parallel computing. She was an assistant professor at Istanbul Bilgi University between 2014-2016, and has been a member of the Knitro solver development team since 2016.
Date: Friday, May 20, 2022
Seminar will take place onsite and online, simultaneously.
Place: Engineering Building, South Campus, VYKM 2
Online Seminar Link:
Meeting ID: 914 9694 2452