Speaker: Wenjie Xu EPFL
Time: 2024-08-01 10:00-2024-08-01 11:00
Venue: FIT 1-222

Abstract:

Performance tuning is ubiquitous in engineering, where a set of parameters are tuned to optimize the performance of a system subject to a set of constraints. However, the mappings from the parameters to the performance/constraints are typically black-box and each evaluation of these functions can be very expensive. Bayesian optimization (BO), as a sample-efficient black-box optimization method, has found successful applications in many performance-tuning tasks. However, vanilla BO algorithms can not manage the constraint violations well, suffer from the curse of dimensionality, and assume access to the absolute black-box function value. In this talk, I will present our recent works on constrained, grey-box, and preferential Bayesian optimization algorithms, in an attempt to address the violation management, scalability, and human feedback issues.

Short Bio:

Wenjie Xu is a Ph.D. candidate in Electrical Engineering at EPFL. He received the B.Eng. degree in Electronic Engineering from Tsinghua University in 2018 and M.Phil. degree in Information Engineering from The Chinese University of Hong Kong in 2020. His current research interests are in the data-driven optimization, with applications to cyber-physical-human systems. He was a recipient of the Chinese government award for outstanding self-financed students abroad (2023). He received the ASME Energy System Technical Committee Best Paper Award at 2022 American Control Conference.