Evaluating the Diversity and Utility of Materials Proposed by Generative Models
Abstract
Generative machine learning models can use data generated by scientific modeling to create large quantities of novel material structures. Here, we assess how one state-of-the-art generative model, the physics-guided crystal generation model (PGCGM), can be used as part of the inverse design process. We show that the default PGCGM's input space is not smooth with respect to parameter variation, making material optimization difficult and limited. We also demonstrate that most generated structures are predicted to be thermodynamically unstable by a separate property-prediction model, partially due to out-of-domain data challenges. Our findings suggest how generative models might be improved to enable better inverse design.
Cite
Text
New et al. "Evaluating the Diversity and Utility of Materials Proposed by Generative Models." ICML 2023 Workshops: SynS_and_ML, 2023.Markdown
[New et al. "Evaluating the Diversity and Utility of Materials Proposed by Generative Models." ICML 2023 Workshops: SynS_and_ML, 2023.](https://mlanthology.org/icmlw/2023/new2023icmlw-evaluating/)BibTeX
@inproceedings{new2023icmlw-evaluating,
title = {{Evaluating the Diversity and Utility of Materials Proposed by Generative Models}},
author = {New, Alexander and Pekala, Michael and Pogue, Elizabeth A and Le, Nam Q and Domenico, Janna and Piatko, Christine D. and Stiles, Christopher D},
booktitle = {ICML 2023 Workshops: SynS_and_ML},
year = {2023},
url = {https://mlanthology.org/icmlw/2023/new2023icmlw-evaluating/}
}