https://github.com/wiseodd/bo-async-feedback
https://github.com/wiseodd/bo-async-feedback
Last synced: 9 months ago
JSON representation
- Host: GitHub
- URL: https://github.com/wiseodd/bo-async-feedback
- Owner: wiseodd
- License: mit
- Created: 2024-02-09T21:47:57.000Z (about 2 years ago)
- Default Branch: master
- Last Pushed: 2024-07-18T10:31:13.000Z (over 1 year ago)
- Last Synced: 2025-07-02T06:06:31.845Z (9 months ago)
- Language: Python
- Size: 28.5 MB
- Stars: 1
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Bayesian Optimization With _Asynchronous_ Expert Feedback
Code to reproduce the paper:
```
@inproceedings{kristiadi2024asyncBO,
title={How Useful is Intermittent, Asynchronous Expert Feedback for {B}ayesian Optimization?},
author={Kristiadi, Agustinus and Strieth-Kalthoff, Felix and Subramanian, Sriram Ganapathi and Fortuin, Vincent and Poupart, Pascal and Pleiss, Geoff},
booktitle={AABI Non Archival Track},
year={2024}
}
```
## Setup
Requires `python >= 3.10` and `pytorch >= 2.0`. Then install the dependencies:
```bash
pip install git+https://git@github.com/aleximmer/laplace
pip install git+https://git@github.com/wiseodd/laplace-bayesopt
pip install lightning rdkit tqdm gauche botorch
```
Next, download the submodules (MolSkill, for expert simulator in chemistry problems).
```bash
git submodule init
git submodule update
```
## Usage
The expert simulator for toy problems is trained using `toy_train_reward.py`. The experiment scripts are `toy_bo.py`, `chem_bo.py`. Run `.py --help` for available options.