This repo contains experimental code at the moment.
This package formulates a prototypical Bayesian optimization algorithm using abstract building blocks. Currently it supports gradient free optimization only.
It comes with a basic OptimizationHelper
that provides utilities for unconstrained problem definition and logging.
The purpose of this little framework is to reuse code and to expose Bayesian optimization algorithms via a unified interface.
The mathematical framework behind Bayesian optimization algorithms that motivates us to write such a generic Julia code has been described in Bayesian Optimization Book by Roman Garnett available at bayesoptbook.com/.