45th CEMS Colloquium

Title

Sparse modeling in condensed matter physics

Abstract

The basic notions of sparse modeling (SpM) are as follows. First, high-dimensional data are assumed to have a sparse representation. Second, the number of explanatory variables should be reduced without loss of accuracy. Finally, explanatory variables are selected objectively, and effective models of target phenomena are constructed automatically [1]. In this talk, I explain why an SpM algorithm called LASSO works well, and give two applications of LASSO in condensed matter physics. The first example is an application of SpM to quasi-particle interference (QPI) experiment with scanning tunneling microscopy/spectroscopy (STM/STS) [2]. We showed that SpM improves efficiency and saves measurement time. The other is an application of Sparsity-promoting Dynamic Mode Decomposition (SpDMD) to coherent-phonon analysis [3]. We showed that SpDMD distinguishes signals of interest from the artifact noise of measurement system. Finally, I describe my future perspective of the data driven approach based on the SpM and the Bayesian inference in condensed matter physics [1].

[1] Y. Igarashi, K. Nagata, T. Kuwatani, T. Omori, Y. Nakanishi-Ohno, and M. Okada, J. Phys. Conf. Ser., 699(1), 012001 (2016).
[2] Y. Nakanishi-Ohno, M. Haze, Y. Yoshida, K. Hukushima, Y. Hasegawa and M. Okada, J. Phys. Soc. Jpn. 85, 093702 (2016).
[3] I. Akai, S. Murata, S. Aihara, S. Tokuda, K. Iwamitsu, and M. Okada, “Mode-decomposition analysis by SpDMD for coherent phonon signals (I)”, 2016 Autumn Meeting of The Physical Society of Japan, 14pAL-3 (2016).