Large Language Models (LLMs) have demonstrated great performance in few-shot In-Context Learning (ICL) for a variety of generative and discriminative chemical design tasks. The newly expanded context windows of LLMs can further improve ICL capabilities for molecular inverse design and lead optimization. To take full advantage of these capabilities we developed a new semi-supervised learning method that overcomes the lack of experimental data available for many-shot ICL. Our approach involves iterative inclusion of LLM generated molecules with high predicted performance, along with experimental data. We further integrated our method in a multi-modal LLM which allows for the interactive modification of generated molecular structures using text instructions. As we show, the new method greatly improves upon existing ICL methods for molecular design while being accessible and easy to use for scientists.
@article{arxiv.2407.19089,
title = {Many-Shot In-Context Learning for Molecular Inverse Design},
author = {Saeed Moayedpour and Alejandro Corrochano-Navarro and Faryad Sahneh and Shahriar Noroozizadeh and Alexander Koetter and Jiri Vymetal and Lorenzo Kogler-Anele and Pablo Mas and Yasser Jangjou and Sizhen Li and Michael Bailey and Marc Bianciotto and Hans Matter and Christoph Grebner and Gerhard Hessler and Ziv Bar-Joseph and Sven Jager},
journal= {arXiv preprint arXiv:2407.19089},
year = {2024}
}