Skip to content
/ pf2es Public

Information-Theoretic based Acquisition Function for Parallalble Multi-Objective Bayesian Optimization

License

Notifications You must be signed in to change notification settings

TsingQAQ/pf2es

Repository files navigation

{PF} $^2$ ES: A New Acquisition Function for Parallel Feasible Pareto Frontier Entropy Search

Installation

❗❗❗Caution: You are away from the main branch of Trieste, this branch contains certain other dependencies

install from sources, run

$ pip install -e.

in the repository root (tested with Python version 3.7.11).

Tutorial/Demonstration Notebook

There is a tutorial notebook multi_objective_entropy_search_via_PF2ES.py at (\docs\notebooks) demonstrating how to make use of the acquisition function for different multi-objective optimization tasks.

In order to run the notebook, install the following dependency:

$ pip install -r notebooks/requirements.txt

Then, run the notebooks with

$ jupyter-notebook notebooks

Reproduce the paper's result

If you'd like to reproduce the paper's result exactly, the following directories contain relevant experiments:

About

Information-Theoretic based Acquisition Function for Parallalble Multi-Objective Bayesian Optimization

Resources

License

Code of conduct

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages