FEAT: A Fast, Effective, and Feasible Model for Molecular Property Prediction Based on Graph Neural Network

Mukesh Rohil, Ishan Sharma

2024

Abstract

Artificial Intelligence based methods and algorithms are being increasingly used by chemists to perform various tasks that would be rather difficult to perform using conventional methods. Whenever scientists design a new set of molecules for certain application, they need to experimentally validate if it possesses the desirable properties. Such (iterative) methods are often expensive and time-consuming. In the realm of Artificial Intelligence and Machine Learning, the molecules can themselves be viewed as graphs present in nature with bonds as edges and nodes as atoms. Therefore, it is worthwhile to exploit Graph Neural Networks for extracting the structural properties of these atoms and bonds, so as to further leverage these to predict the properties of these molecules (represented as graphs) as a whole. We propose a Graph Neural Network based model, FEAT, for this purpose. FEAT’s performance has been evaluated on multiple publicly available datasets and the results obtained are promising.

Paper Citation

in Harvard Style

Rohil M. and Sharma I. (2024). FEAT: A Fast, Effective, and Feasible Model for Molecular Property Prediction Based on Graph Neural Network. In Proceedings of the 16th International Conference on Agents and Artificial Intelligence - Volume 3: ICAART; ISBN 978-989-758-680-4, SciTePress, pages 878-881. DOI: 10.5220/0012410700003636

in Bibtex Style

@conference{icaart24,
author={Mukesh Rohil and Ishan Sharma},
title={FEAT: A Fast, Effective, and Feasible Model for Molecular Property Prediction Based on Graph Neural Network},
booktitle={Proceedings of the 16th International Conference on Agents and Artificial Intelligence - Volume 3: ICAART},
year={2024},
pages={878-881},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0012410700003636},
isbn={978-989-758-680-4},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 16th International Conference on Agents and Artificial Intelligence - Volume 3: ICAART
TI - FEAT: A Fast, Effective, and Feasible Model for Molecular Property Prediction Based on Graph Neural Network
SN - 978-989-758-680-4
AU - Rohil M.
AU - Sharma I.
PY - 2024
SP - 878
EP - 881
DO - 10.5220/0012410700003636
PB - SciTePress