Towards a Database of Bond Orders: combining quantum chemistry and Machine learning

Joyeta Saha, Hemant Lohumi

Keywords: Machine Learning, Chemical bond, Bond orders

Abstract:

Bond order is a key descriptor of chemical bonding, yet no general database of computed bond orders exists for machine learning applications. In this project, we curate Mayer Bond Orders for the GW100 benchmark molecules using quantum-chemical calculations. Our immediate goal is to complete dataset curation before the second presentation, after which we will train machine learning models to predict bond orders for unseen molecules. This work establishes one of the first systematic bond order datasets and explores ML as a fast, data-driven tool for bonding analysis.
https://youtu.be/ORNTGavbAsM?si=6x72RYpAjCXZoSa9

Submission Number: 2