Learning the right channel in multimodal imaging: automated experiment in piezoresponse force microscopy

作     者：Yongtao Liu Rama K.Vasudevan Kyle P.Kelley Hiroshi Funakubo Maxim Ziatdinov Sergei V.Kalinin 

作者机构：Center for Nanophase Materials SciencesOak Ridge National LaboratoryOak RidgeTN 37923USA Department of Materials Science and EngineeringTokyo Institute of TechnologyYokohama 226-8502Japan Computational Sciences and Engineering DivisionOak Ridge National LaboratoryOak RidgeTN 37923USA Department of Materials Science and EngineeringUniversity of TennesseeKnoxvilleTN 37996USA 

出 版 物：《npj Computational Materials》 (计算材料学（英文）)

年 卷 期：2023年第9卷第1期

页      面：2008-2015页

核心收录：

学科分类：08[工学] 0803[工学-光学工程] 0812[工学-计算机科学与技术（可授工学、理学学位）] 

基　　金：This effort(implementation in SPM,measurement,data analysis)was primarily supported by the center for 3D Ferroelectric Microelectronics(3DFeM),an Energy Frontier Research Center funded by the U.S.Department of Energy(DOE),Office of Science,Basic Energy Sciences under Award Number DE-SC0021118 This research(ensemble-DKL)was supported by the Center for Nanophase Materials Sciences(CNMS),which is a US Department of Energy,Office of Science User Facility at Oak Ridge National Laboratory This work was also supported by MEXT Program:Data Creation and Utilization Type Material Research and Development Project Grant Number JPMXP1122683430 

主　　题：automated policy modal 

摘      要：We report the development and experimental implementation of the automated experiment workflows for the identification of thebest predictive channel for a phenomenon of interest in spectroscopic measurements. The approach is based on the combinationof ensembled deep kernel learning for probabilistic predictions and a basic reinforcement learning policy for channel selection. Itallows the identification of which of the available observational channels, sampled sequentially, are most predictive of selectedbehaviors, and hence have the strongest correlations. We implement this approach for multimodal imaging in piezoresponse forcemicroscopy (PFM), with the behaviors of interest manifesting in piezoresponse spectroscopy. We illustrate the best predictivechannel for polarization-voltage hysteresis loop and frequency-voltage hysteresis loop areas is amplitude in the model samples. Thesame workflow and code are applicable for any multimodal imaging and local characterization methods.