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

Aug 5, 2026

Fine-tuning GPT-3 is evaluated for predicting the electronic and functional properties of organic molecules. Results show that fine-tuned models can recognize chemically meaningful patterns, differentiate subtle distinctions, and deliver robust predictive performance for molecular property prediction. The study analyzes resilience to information loss caused by missing atoms or chemical groups and to noise introduced by random alterations in atomic identities. Challenges, limitations, and directions for future improvements of GPT-3 in molecular machine-learning tasks are discussed.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages12

Reading Time30 min

AccessFree to download

Introduction

Machine learning for organic molecules

Large language models in chemistry

Study focus and evaluation aspects

University of Birmingham Fine-tuning GPT-3 for machine learning electronic and functional properties of organic molecules Xie, Zikai; Evangelopoulos, Xenophon; Omar, Ömer H. ; Troisi, Alessandro; Cooper, Andrew I. ; Chen, Linjiang DOI: 10.1039/D3SC04610A License: Creative Commons: Attribution (CC BY) Document Version Publisher's PDF, also known as Version of record Citation for published version (Harvard): Xie, Z, Evangelopoulos, X, Omar, ÖH, Troisi, A, Cooper, AI & Chen, L 2024, 'Fine-tuning GPT-3 for machine learning electronic and functional properties of organic molecules', Chemical Science, vol. 15, no. 2, pp. 500- 510. [https://doi.org/10.1039/D3SC04610A](https://doi.org/10.1039/D3SC04610A) Link to publication on Research at Birmingham portal General rights Unless a licence is specified above, all rights (including copyright and moral rights) in this document are retained by the authors and/or the copyright holders. The express permission of the copyright holder must be obtained for any use of this material other than for purposes permitted by law. •Users may freely distribute the URL that is used to identify this publication. •Users may download and/or print one copy of the publication from the University of Birmingham research portal for the purpose of private study or non-commercial research. •User may use extracts from the document in line with the concept of ‘fair dealing’ under the Copyright, Designs and Patents Act 1988 (?) •Users may not further distribute the material nor use it for the purposes of commercial gain. Where a licence is displayed above, please note the terms and conditions of the licence govern your use of this document. When citing, please reference the published version. Take down policy While the University of Birmingham exercises care and attention in making items available there are rare occasions when an item has been uploaded in error or has been deemed to be commercially or otherwise sensitive. If you believe that this is the case for this document, [please contact UBIRA@lists.bham.ac.uk](please contact UBIRA@lists.bham.ac.uk) providing details and we will remove access to the work immediately and investigate. Download date: 05. Aug. 2026 Open Access Article . Pu on 05 Decemberblished 2023. Down on 5/14/2024loaded 10:51:19 AM . hicle is licensed under a Creative C mmons A 3 0 U d[ttr .](ttr .)ibution npor e nce. Chemical Science EDGE ARTICLE View Article Online View Journal | View Issue Cite this: Chem. Sci., 2024, 15, 500 All publication charges for this article have been paid for by the Royal Society of Chemistry Received 31st August 2023 Accepted 4th December 2023 DOI: 10.1039/d3sc04610a[rsc.li/chemical-science](rsc.li/chemical-science) Fine-tuning GPT-3 for machine learning electronic and functional properties of organic molecules Zikai Xie,a Xenophon Evangelopoulos,a ¨Omer H. Omar, b Alessandro Troisi, b Andrew I. Cooper *a and Linjiang Chen *c 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 via 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. Introduction There has been recent and growing interest in leveraging machine learning (ML) for diverse applications involving organic molecules, such as predicting molecular properties1–6 and usi

FAQs

What is the main goal of fine-tuning GPT-3 in this study?

To evaluate how effectively fine-tuned GPT-3 can predict the electronic and functional properties of organic molecules.

How do fine-tuned GPT-3 models perform in identifying molecular patterns?

They can successfully identify and distinguish chemically meaningful patterns and detect subtle differences, with robust predictive performance.

What robustness tests are used to stress the models?

The study tests resilience to information loss from missing atoms or chemical groups and to noise from random alterations in atomic identities.

Aug 5, 2026

Upload by Emma Wilson