Large Language Model Predicting the Performance of Small Organic Molecule Corrosion Inhibitors

Matthias Busch, Christian Feiler, Marius Tacke, Jonas F. Eichinger, Batoul Daou, Sviatlana V. Lamaka, Mikhail Zheludkevich, Kevin Linka, Christian J Cyron, Roland Aydin

Keywords: AI, LLM, GPT-4o, Machine Learning, Material Science, Corrosion, Magnesium, Regression, Prediction, Foundation Model

TL;DR: GPT-4o is used to predict the ability of different compounds to inhibit the corrosion of magnesium.

Abstract: Large language models (LLMs) like GPT-4o have shown promise in solving everyday tasks and addressing basic scientific challenges by utilizing extensive pre-trained knowledge. In this work, we explore their potential to predict the efficiency of various organic compounds for the inhibition of corrosion of the magnesium alloy ZE41, a material crucial for many industrial applications. Traditional approaches, such as basic neural networks, rely on non-contextual data, often requiring large datasets and significant effort per sample to achieve accurate predictions. They struggle particularly with small datasets, limiting their effectiveness in discovering new corrosion inhibitors. LLMs can contextualize and interpret limited data points by drawing on their vast knowledge, including chemical properties of molecules and their influence on corrosion processes in other materials like iron. By prompting the model with a small dataset, LLMs can provide meaningful predictions without the need for extensive training. Our study demonstrates that LLMs can predict corrosion inhibition outcomes, and reduce the amount of data needed.

Primary Area: foundation or frontier models, including LLMs

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.

Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2025/AuthorGuide.

Reciprocal Reviewing: I understand the reciprocal reviewing requirement as described on https://iclr.cc/Conferences/2025/CallForPapers. If none of the authors are registered as a reviewer, it may result in a desk rejection at the discretion of the program chairs. To request an exception, please complete this form at https://forms.gle/Huojr6VjkFxiQsUp6.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Submission Number: 7050