Bayesian Optimization: Theory and Practice Using Python

© The Author(s), under exclusive license to APress Media, LLC, part of Springer Nature 2023

P. LiuBayesian Optimizationhttps://doi.org/10.1007/978-1-4842-9063-7_6

6. Knowledge Gradient: Nested Optimization vs. One-Shot Learning

Peng Liu¹  

(1)

Singapore, Singapore

 

In the previous chapter, we learned the inner workings of the optimization procedure using BoTorch, highlighting the auto-differentiation mechanism and modular design of the framework. This paves the way for many new acquisition functions we can plug in and test. In this chapter, we will extend our toolkit of acquisition functions to the knowledge gradient (KG), a nonmyopic acquisition function that performs better than expected improvement (EI) in many cases.

Although empirically superior, ...