Adaptively expanding subspaces (BAxUS)

Adaptively expanding subspaces (BAxUS)#

About#

This is an implementation of Bayesian Optimization with adaptively expanding subspaces (BAxUS) as described in [Papenmeier et al., 2022]. This implementation uses the original open source implementation and Python package developed by the authors.

How to run#

Warning

This solver runs in a different conda environment than base.

You can find a conda environment where this solver can run here.

If you have cloned poli-baselines locally:

conda env create --file src/poli_baselines/solvers/bayesian_optimization/baxus/environment.baxus.yml
conda activate poli__baxus

import numpy as np

from poli import objective_factory
from poli_baselines.solvers.bayesian_optimization.baxus import BAxUS

problem = objective_factory.create(
    name="toy_continuous_problem",
    function_name="ackley_function_01",
    n_dimensions=10,
)
black_box, x0 = problem.black_box, problem.x0

x0 = np.random.uniform(-1, 1, size=10).reshape(1, 10)
y0 = black_box(x0)

solver = BAxUS(black_box, x0, y0, bounds=(-3.0, 3.0), noise_std=0.0, n_init=2)

solver.solve(max_iter=3)

See more#

The original reference: Increasing the Scope as You Learn: Adaptive Bayesian Optimization in Nested Subspaces.
Taking the human out of the loop is a great tutorial of Bayesian Optimization [Shahriari et al., 2016].
Since poli works mostly on discrete inputs, this baseline is implemented with the intention of optimizing in the latent spaces of deep generative models like Generative Adversarial Networks (GANs) or Variational Autoencoders (VAEs) [Gómez-Bombarelli et al., 2018].

References#

If you use this solver, we expect that you cite the following resources:

References as text

[1] Papenmeier, L., Nardi, L., & Poloczek, M. (2022, May 16). Increasing the Scope as You Learn: Adaptive Bayesian Optimization in Nested Subspaces. Advances in Neural Information Processing Systems. https://openreview.net/forum?id=e4Wf6112DI

[2] González-Duque, M., Bartels, S., & Michael, R. (2024). poli: a libary of discrete sequence objectives [Computer software]. MachineLearningLifeScience/poli

References as BibTeX

@inproceedings{Papenmeier:BAxUS:2022,
    title={Increasing the Scope as You Learn: Adaptive Bayesian Optimization in Nested Subspaces},
    url={https://openreview.net/forum?id=e4Wf6112DI},
    author={Papenmeier, Leonard and Nardi, Luigi and Poloczek, Matthias},
    year={2022},
    month=may,
    language={en}
}

@software{Gonzalez-Duque:poli:2024,
author = {González-Duque, Miguel and Bartels, Simon and Michael, Richard},
month = jan,
title = {{poli: a libary of discrete sequence objectives}},
url = {https://github.com/MachineLearningLifeScience/poli},
version = {0.0.1},
year = {2024}
}