Predicting the electronic structure of matter at scale with machine learning

Datum

17.10.2024

Zeit

13:00 - 15:00

Sprecher

Attila Cangi

Zugehörigkeit

Helmholtz-Zentrum Dresden-Rossendorf

Serie

TUD nanoSeminar

Sprache

en

Hauptthema

Physik

Andere Themen

Physik

Host

Arezoo Dianat

Beschreibung

In this presentation, I will discuss our recent advancements in utilizing machine learning to significantly enhance the efficiency of electronic structure calculations [1]. Specifically, I will focus on our efforts to accelerate Kohn-Sham density functional theory calculations by incorporating deep neural networks within the Materials Learning Algorithms framework [2,3]. Our results demonstrate substantial gains in calculation speed for metals across their melting point. Additionally, our implementation of automated machine learning has resulted in significant savings in computational resources when identifying optimal neural network architectures, laying the foundation for large-scale investigations [4]. Furthermore, I will present our most recent breakthrough, which enables neural-network-driven electronic structure calculations for systems containing over 100,000 atoms [5]. This achievement opens up new avenues for studying complex materials systems that were previously computationally intractable. 
 
 [1] L. Fiedler, K. Shah, M. Bussmann, A. Cangi, Phys. Rev. Materials, 6, 040301 (2022) 
 [2] A. Cangi, J. A. Ellis, L. Fiedler, D. Kotik, N. A. Modine, V. Oles, G. A. Popoola, S. Rajamanickam, S. Schmerler, J. A. Stephens, A. P. Thompson, Phys. Rev. B 104, 035120 (2021). 
 [3] J. Ellis, L. Fiedler, G. Popoola, N. Modine, J. Stephens, A. Thompson, A. Cangi, S. Rajamanickam, Phys. Rev. B, 104, 035120 (2021) 
 [4] L. Fiedler, N. Hoffmann, P. Mohammed, G. Popoola, T. Yovell, V. Oles, J. Austin Ellis, S. Rajamanickam, A. Cangi, Mach. Learn.: Sci. Technol., 3, 045008 (2022) 
 [5] L. Fiedler, N. Modine, S. Schmerler, D. Vogel, G. Popoola, A. Thompson, S. Rajamanickam, A. Cangi, npj. Comput. Mater., 9, 115 (2023)

Links

• More information

Letztmalig verändert: 02.10.2024, 20:26:18