Maseeh Summer Undergraduate Research Experience
Funding Source
Undergraduate Research and Mentoring Program (URMP)
Advisor
Yi Xia
Document Type
Poster
Publication Date
2023
Abstract
This study is training a machine learning model to identify each different carbon allotrope with the best thermal conductivity. This machine learning model uses the M3GNet model to predict the energies and forces of the carbon allotrope. Thermal conductivity is important to prevent chips from overheating. Carbon is known for its high thermal conductivity, however, there are many allotropes. M3Gnet is a massive pretrained neural network trained to predict interatomic potential energy given energies, forces, and stresses. Convergence of the model is achieved by varying the learning rate to the varying number of epoch.
Citation Details
Chung, Angel, "Modeling Carbon Allotropes using Machine Learning" (2023). Maseeh Summer Undergraduate Research Experience. 53.