Fine-tuning GPT-3 for machine learning electronic and functional properties of organic molecules.



Xie, Zikai, Evangelopoulos, Xenophon, Omar, Ömer H ORCID: 0000-0002-5073-4999, Troisi, Alessandro ORCID: 0000-0002-5447-5648, Cooper, Andrew I ORCID: 0000-0003-0201-1021 and Chen, Linjiang ORCID: 0000-0002-0382-5863
(2024) Fine-tuning GPT-3 for machine learning electronic and functional properties of organic molecules. Chemical science, 15 (2). pp. 500-510.

Access the full-text of this item by clicking on the Open Access link.

Abstract

We evaluate the effectiveness of fine-tuning GPT-3 for the prediction of electronic and functional properties of organic molecules. Our findings show that fine-tuned GPT-3 can successfully identify and distinguish between chemically meaningful patterns, and discern subtle differences among them, exhibiting robust predictive performance for the prediction of molecular properties. We focus on assessing the fine-tuned models' resilience to information loss, resulting from the absence of atoms or chemical groups, and to noise that we introduce <i>via</i> random alterations in atomic identities. We discuss the challenges and limitations inherent to the use of GPT-3 in molecular machine-learning tasks and suggest potential directions for future research and improvements to address these issues.

Item Type: Article
Uncontrolled Keywords: 34 Chemical Sciences, Machine Learning and Artificial Intelligence
Divisions: Faculty of Science and Engineering > School of Physical Sciences
Depositing User: Symplectic Admin
Date Deposited: 30 Jan 2024 10:06
Last Modified: 21 Jun 2024 08:27
DOI: 10.1039/d3sc04610a
Open Access URL: https://doi.org/10.1039/D3SC04610A
Related URLs:
URI: https://livrepository.liverpool.ac.uk/id/eprint/3178080