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_7

7. Case Study: Tuning CNN Learning Rate with BoTorch

Peng Liu¹  

(1)

Singapore, Singapore

 

By now, we have established a good foundation regarding the theoretical inner workings of a typical Bayesian optimization process: a surrogate function that approximates the underlying true function and gets updated as new data arrives and an acquisition function that guides the sequential search under uncertainty. We have covered popular choices of acquisition function, including expected improvement (EI, with its closed-form expression derived in Chapter 3), the more general Monte Carlo acquisition ...