navidpanchi / morbo

Multi-Objective Bayesian Optimization over High-Dimensional Search Spaces

Multi-Objective Bayesian Optimization over High-Dimensional Search Spaces

This is the code associated with the paper "Multi-Objective Bayesian Optimization over High-Dimensional Search Spaces."

Please cite our work if you find it useful.

@InProceedings{pmlr-v162-daulton22a,
title = 	 {Multi-Objective Bayesian Optimization over High-Dimensional Search Spaces},
author =       {Daulton, Samuel and Eriksson, David and Balandat, Maximilian and Bakshy, Eytan},
booktitle = 	 {Proceedings of the Thirty-Eighth Conference on Uncertainty in Artificial Intelligence},
year = 	 {2022},
series = 	 {Proceedings of Machine Learning Research},
publisher =    {PMLR},
}

Getting started

From the base morbo directory run:

pip install -e .

Structure

The code is structured in three parts.

• The utilities for constructing the acquisition functions and other helper methods are defined in morbo/.
• The experiments are found in and ran from within experiments/. The main.py is used to run the experiments, and the experiment configurations are found in the config.json file of each sub-directory.

The individual experiment outputs were left out to avoid inflating the file size.

Running Experiments

To run a basic benchmark based on the config.json file in experiments/<experiment_name> using <algorithm>:

cd experiments
python main.py <experiment_name> <algorithm> <seed>

The code refers to the algorithms using the following labels:

algorithms = [
    ("morbo", "MORBO"),
]

Each folder under experiments/ corresponds to the experiments in the paper according to the following mapping:

experiments = {
    "dtlz2_10d": "DTLZ2 (d=10)",
    "dtlz2_30d": "DTLZ2 (d=30)",
    "dtlz2_100d": "DTLZ2 (d=100)",
    "dtlz3_m2": "DTLZ3 (M=2)",
    "dtlz5_m2": "DTLZ5 (M=2)",
    "dtlz7_m2": "DTLZ7 (M=2)",
    "dtlz3_m4": "DTLZ3 (M=4)",
    "dtlz5_m4": "DTLZ5 (M=4)",
    "dtlz7_m4": "DTLZ7 (M=4)",
    "rover": "Rover",
    "vehicle_safety": "Vehicle Safety",
    "welded_beam": "Welded Beam",
}

Note: this code can heavily exploit a GPU if available.

License

This repository is MIT licensed, as found in the LICENSE file.