hdbo-b 0.1.1

High-Dimentional Bayesian Benchmark (HDBO-B)

HDBO-B: High Dimensional Bayesian Optimization Benchmark

HDBO-B, a generalized and unified benchmark for High Dimensional Bayesian Optimization, is including testing functions, example algorithms and more examples.

:gear: Installation

We recommend using CONDA to isolate Python environments and their packages, this protects you from potential package version conflicts.

To install HDBO-B package, choose one below:

• git clone https://github.com/yiyuiii/HDBO-B && cd HDBO-B && pip install -e . (You may get example codes ONLY by this way.)
• pip install hdbo-b
• wget https://github.com/Yiyuiii/HDBO-B/releases/download/whl/hdbo_b-0.1.0-py3-none-any.whl && pip install hdbo_b-0.1.0-py3-none-any.whl

In case we use HEBO 0.3.5 for experiments and there is 0.3.2 on PyPI currently, user should manually install:

• git clone https://github.com/huawei-noah/HEBO && cd HEBO/HEBO && pip install -e .

:rocket: Quick Start

Uncomment lines in test.py, and run python test.py.

User may also check codes in folder ./example.

:wrench: Details

Test Functions

To import hand-designed 30 test functions:

from HDBOBenchmark import TestFuncs as func_list

We have designed a stricter framework for functions and algorithms, This brings a lot of convenience to scaling, while there are a bit more difficulties in getting started. Check them out in

• ./HDBOBenchmark/base/FunctionBase.py
• ./HDBOBenchmark/AdditiveFunction.py
• ./example/wrapper/hebo_wrapper.py

Bayesian Optimization Algorithms

Algorithms can be found in ./examples/algorithms (more original) and ./examples/wrapper (directly imported by ./examples/optimize.py), including

• GP(Gaussian Process) + UCB(Upper Confidence Bound) Implementation with BOTorch

M. Balandat, B. Karrer, D. R. Jiang, S. Daulton, B. Letham, A. G. Wilson, and E. Bakshy. BoTorch: A Framework for Efficient Monte-Carlo Bayesian Optimization. Advances in Neural Information Processing Systems 33, 2020.

• GP/GPy(warped GP)/gumbel/... + MACE(Multi-objective ACquisition function Ensemble) with HEBO

Cowen-Rivers, Alexander I., et al. HEBO: Pushing The Limits of Sample-Efficient Hyperparameter Optimisation. Journal of Artificial Intelligence Research, 2022.

• Python implementation of Add-GP-UCB

Kandasamy K, Schneider J, Póczos B. High dimensional Bayesian optimisation and bandits via additive models. International conference on machine learning, 2015.

• Python implementation of REMBO

Wang Z, Hutter F, Zoghi M, et al. Bayesian optimization in a billion dimensions via random embeddings. Journal of Artificial Intelligence Research, 2016.

• Wrapper of ALEBO

Letham B, Calandra R, Rai A, et al. Re-examining linear embeddings for high-dimensional Bayesian optimization. Advances in neural information processing systems, 2020.

• Wrapper of TuRBO

Eriksson D, Pearce M, Gardner J, et al. Scalable global optimization via local bayesian optimization. Advances in neural information processing systems, 2019.

• Wrapper of SAASBO

Eriksson D, Jankowiak M. High-dimensional Bayesian optimization with sparse axis-aligned subspaces. Uncertainty in Artificial Intelligence, 2021.

./examples/optimize.py saves optimizing histories in folder ./result, which are automatically read by ./example/plot_result.py.

:speech_balloon: Common Issues

:triangular_flag_on_post: TODO List

• Introducing more realistic test functions

• Introducing VAE-based algorithms and related datasets needed

:microscope: Cite Us

:clipboard: Changelog

2023.06.06 > v0.1.1 :fire:

• Fixed logging.

2023.06.03 > v0.1.0

• Initialization.