Project 17: Comparative Analysis of Acquisition Functions in Bayesian Optimization for Drug Discovery
This project conducts a comparative analysis of acquisition functions in Bayesian Optimization (BO) for drug discovery, focusing on small, diverse, unbalanced, and noisy datasets. It evaluates the impact of different acquisition functions, molecular featurization methods, and applicability domain (AD) to uncover optimal strategies for employing AF effectively in drug discovery challenges.
Please refer the github page for our observations: https://suneelbvs.github.io/AC-BO-Hackathon.html#/
To install an editable version of the package, run the following command:
pip install PyTDC==0.3.6
pip install xgboost
pip install torch
pip install gpytorch
pip install requests!python main.py (assuming you updated correct data_loader) for model building
- Hugo Bellamy (2022), "Batched Bayesian Optimization for Drug Design in Noisy Environments," J Chem Inf Model. 2022 Sep 12; 62(17): 3970–3981.
- Jan Christopher Spies, University of Muenster
- Jakub Lála, Imperial College London
- Yunheng Zou, University of Waterloo
- Luis Walter, Heidelberg University
- Curtis Chong, University of Waterloo