Skip to content

Code to run the ASEBO algorithm from the paper: From Complexity to Simplicity: Adaptive ES-Active Subspaces for Blackbox Optimization... please get in touch if interested!!

License

Notifications You must be signed in to change notification settings

jparkerholder/ASEBO

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

9 Commits
 
 
 
 
 
 

Repository files navigation

ASEBO

This is code to accompany the paper: From Complexity to Simplicity: Adaptive ES-Active Subspaces for Blackbox Optimization implementing one version of the ASEBO algorithm. Here is the link to the paper:

It is not the implementation used for the experiments in the paper, which is proprietary, but is a re-implementation which has been tested. The benefit of this is that we use numpy and a single machine, so it should run locally on a laptop.

We hope you find this useful - if you do, please cite:

@incollection{asebo,
title = {From Complexity to Simplicity: Adaptive ES-Active Subspaces for Blackbox Optimization},
author = {Choromanski, Krzysztof M and Pacchiano, Aldo and Parker-Holder, Jack and Tang, Yunhao and Sindhwani, Vikas},
booktitle = {Advances in Neural Information Processing Systems 32},
year = {2019}
}

About

Code to run the ASEBO algorithm from the paper: From Complexity to Simplicity: Adaptive ES-Active Subspaces for Blackbox Optimization... please get in touch if interested!!

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages