Molecular Descriptors with Actively Identified Subsets
A code for efficient molecular property optimization. This repository is under development. A cleaner version with tutorials and extended functionality (e.g., constraints, multi-objective, and human-in-the-loop) will be made available upon journal publication.
- Build environment
- Run main.py
- Add your .csv file with a SMILES and columns
- Add src/config/log_p_test_exp/<new_exp_name>.json, e.g., change the "exp_name", "Data_loc", and "y_variable" fields in src/config/log_p_test_exp/log_P_test_exp.json
- Change line 56 in main to point to <new_exp_name>.json
- Run main.py
- under development
Sorourifar, Farshud, Thomas Banker, and Joel A. Paulson. "Accelerating Black-Box Molecular Property Optimization by Adaptively Learning Sparse Subspaces." arXiv preprint arXiv:2401.01398 (2024).
@misc{sorourifar2024accelerating, title={Accelerating Black-Box Molecular Property Optimization by Adaptively Learning Sparse Subspaces}, author={Farshud Sorourifar and Thomas Banker and Joel A. Paulson}, year={2024}, eprint={2401.01398}, archivePrefix={arXiv}, primaryClass={q-bio.BM} }