This is an (messy) implementation of the method for automating (model selection in) Bayesian optimization described in
Malkomes, Gustavo; Garnett, Roman. Automating Bayesian optimization with Bayesian optimization (2018). Advances in Neural Information Processing Systems 31 (NeurIPS 2018)
This version of the code is not using the active_gp_learning package,
so it's still a little bit messy.
Feel free to send me an email at gustavo@sigopt.com; Perhaps we can collaborate in a new version of this code. ;-)