Supervised learning and prediction of electronic properties: Discovery and Design of Materials and Interfaces for back-end-of-line interconnects

Supervised machine learning based techniques have found notable success in the recent past in the fields of atomic structure prediction and interatomic potential generation.
Such techniques, however, have not found similar success in the prediction of non-local electronic-structural quantities such as electronic transmission and total density of states.
In this work, we present approaches to accurately predict a range of electronic structure properties - from Hamiltonian Matrix elements, electronic transmission to density of states - using supervised learning.
Training and test data is obtained from open materials databases and from tailored first-principles Density Functional Theory simulations of materials and material interfaces.
The key role played by structural and electronic descriptors and their impact on the accuracy of predictions is discussed.
The use of simulation-based proxies for material properties accessible through experiment is also discussed.
Finally, recent progress on the discovery and design of conductor and barrier materials for back-end-of-line interconnect using supervised learning is presented.