FEIT Research Project Database

Quantum material design by machine learning

Project Leader: M Usman
Primary Contact: Muhammad Usman (muhammad.usman@unimelb.edu.au)
Keywords: machine learning; nanoelectronics; nanoengineered materials; nanophotonics; nanostructured materials
Disciplines: Computing and Information Systems,Electrical & Electronic Engineering

Quantum materials are fundamental building blocks of nearly all electronic technologies of this age including computing devices, photo-cells, integrated circuits, among many others. The investigation and understanding of material electronic properties is a highly tedious task and require extremely computationally intensive simulations utilising supercomputers. Furthermore, the large design space of quantum materials makes it nearly impossible to identify optimised structures which can lead to improved device performance. Artificial intelligence methods could provide help in learning material properties, developing a structure-property relationship, and identifying optimised materials with desired properties. This project will develop a machine learning algorithm trained on density functional theory, with the capabilities to propose new breakthrough material structures for next generation technologies.

Reference papers

Butler et al, Nature 559, 457, 2018

Luna et al, Nature 552, 23, 2017

Gibney et al, Nature 560, 151, 2018