Classifying surface probe images in strongly correlated electronic systems via machine learning
Classifying surface probe images in strongly correlated electronic systems via machine learning
复制标题
DOI：
10.1103/physrevmaterials.3.033805
复制
发表时间：
2019-03-29
影响因子：
3.4
通讯作者：
Carlson, E. W.
中科院分区：
文献类型：
--
作者：
Burzawa, L.;Liu, S.;Carlson, E. W.
扫描探针实验，如扫描隧道显微镜(STM)和原子力显微镜(AFM)在强关联电子系统上的实验，经常发现在多个长度尺度上形成复杂的图案。通过研究这些图像中的普遍标度，我们在几个不同的关联电子系统中表明，图案的形成是由接近无序驱动的临界点驱动的，揭示了这些材料中图案形成的统一性。作为解决这种新材料图像分类问题的另一种方法，我们报告了一种机器学习方法的研究，以确定哪个底层物理模型正在驱动系统中的图案形成。使用神经网络结构，我们能够达到97%的分类正确率从三个模型具有伊辛对称性。这项研究还表明，机器学习可以捕获物理系统的隐含通用行为。这拓宽了我们对机器学习能做什么的理解，我们期待未来机器学习和凝聚态物理之间有更多的协同作用。
Scanning probe experiments such as scanning tunneling microscopy (STM) and atomic force microscopy (AFM) on strongly correlated electronic systems often reveal complex pattern formation on multiple length scales. By studying the universal scaling in these images, we have shown in several distinct correlated electronic systems that the pattern formation is driven by proximity to a disorder-driven critical point, revealing a unification of the pattern formation in these materials. As an alternative approach to this image classification problem of novel materials, here we report an investigation of the machine learning method to determine which underlying physical model is driving pattern formation in a system. Using a neural network architecture, we are able to achieve 97% accuracy on classifying configuration images from three models with Ising symmetry. This investigation also demonstrates that machine learning can capture the implicit universal behavior of a physical system. This broadens our understanding of what machine learning can do, and we expect more synergy between machine learning and condensed matter physics in the future.