Enhancing generative molecular design <i>via</i> uncertainty-guided fine-tuning of variational autoencoders
This study explores low-dimensional active subspace of pre-trained VAE-based generative molecular design models to perform uncertainty-guided fine-tuning of the VAE parameters to enhance downstream design task performance of the pre-trained models.
Authors
- A N M Nafiz Abeer (ORCID: https://orcid.org/0009-0004-8134-1604)
- Nathan M. Urban (ORCID: https://orcid.org/0000-0002-2264-3512)
- Sanket Jantre (ORCID: https://orcid.org/0000-0003-3611-0255)
- Byung-Jun Yoon (ORCID: https://orcid.org/0000-0001-9328-1101)
Institutions
- Brookhaven National Laboratory (US)
- Texas A&M University (US)
Publication Details
- Journal
- Molecular Systems Design & Engineering
- Published
- 2025-10-17
- DOI
- https://doi.org/10.1039/d5me00081e
- Primary Topic
- Machine Learning in Materials Science
- Type
- article
- Field-Weighted Citation Impact
- 0.00
Funders
- U.S. Department of Energy
- Office of Science
- Advanced Scientific Computing Research