Knowing The What But Not The Where In Bayesian Optimization
Vu Nguyen
Dr Vu Nguyen is currently a Senior Research Associate at the Machine Learning Research Group, University of Oxford. He is working with Professor Michael Osborne and Professor Andrew Briggs on a machine learning project for tuning quantum devices using Bayesian optimisation and deep reinforcement learning. Previously he was working as a Research Scientist at a Credit AI in Melbourne and was a postdoctoral researcher at Deakin University where he obtained his PhD in 2015. He was the recipient of ACML 2016 best paper award, IEEE ICDM 2017 best papers and one of the 200 young researchers world-wide for attending Heidelberg Laureate Forum 2015.
Bayesian optimization has demonstrated impressive success in finding the optimum input and output of the black-box function. In some applications, the optimum output of the function is known in advance and the goal is to find the corresponding optimum input. Existing work in Bayesian optimization (BO) has not effectively exploited the knowledge of the optimum output for optimization. In this paper, we consider a new setting in BO in which the knowledge of the optimum output is available, such as in tuning hyperparameters for deep reinforcement learning and supervised learning. Our goal is to exploit such knowledge to find the optimum input efficiently.
Related seminars
Tim Baldwin
MBZUAI, The University of Melbourne
Alessio Del Bue
Italian Institute of Technology (IIT)
Dr. Xiaoming Liu
Michigan State University
Dr Lan Du
Monash University