A5074142581- Machine Learning in Materials Science
- Electronic and Structural Properties of Oxides
- X-ray Diffraction in Crystallography
- Computational Drug Discovery Methods
- Semiconductor Lasers and Optical Devices
- Pharmacy and Medical Practices
- ZnO doping and properties
- Risk and Safety Analysis
- Optical Network Technologies
- Adhesion, Friction, and Surface Interactions
- Energy Load and Power Forecasting
- Arsenic contamination and mitigation
- Memory Processes and Influences
- Analytical Methods in Pharmaceuticals
- Pharmaceutical studies and practices
- Advanced Polymer Synthesis and Characterization
- Blood donation and transfusion practices
- Photosynthetic Processes and Mechanisms
- Magnetic and transport properties of perovskites and related materials
- Algal biology and biofuel production
- Gaze Tracking and Assistive Technology
- Usability and User Interface Design
- Solar Radiation and Photovoltaics
- Advanced Photonic Communication Systems
- Electrowetting and Microfluidic Technologies
Tokyo Institute of Technology
2017-2024
Nagoya Institute of Technology
2024
Shizuoka University
2008-2023
Kwansei Gakuin University
2020
Toyo University
2019
Kyoto University
2014-2018
University of Michigan
2018
National Institute of Technology, Ichinoseki College
2017
Yamagata University
2014-2017
Meiji University
2011-2012
The representations of a compound, called "descriptors" or "features", play an essential role in constructing machine-learning model its physical properties. In this study, we adopt procedure for generating systematic set descriptors from simple elemental and structural representations. First it is applied to large dataset composed the cohesive energy about 18000 compounds computed by density functional theory (DFT) calculation. As result, obtain kernel ridge prediction with error 0.041...
We propose a simple scheme to estimate the potential energy surface (PES) for which accuracy can be easily controlled and improved. It is based on model selection within framework of linear regression using least absolute shrinkage operator (LASSO) technique. Basis functions are selected from systematic large set candidate functions. The sparsity PES significantly reduces computational cost evaluating force in molecular dynamics simulations without losing accuracy. usefulness describing...
A simple method of the calculation mean amplitudes both directly bonded atom pairs and nonbonded ones is proposed by approximating function cothx 1/x+x/4. The enumeration L matrix, most time-consuming procedure, replaced inverse matrix F−1 from potential energy matrix. formulae obtained are applicable with reasonable accuracy (4 percent) to almost all cases except having higher frequencies than 1200 cm−1. For latter modified formulas presented using hyperbolic cotangent function. It also...
Interatomic potentials have been widely used in atomistic simulations such as molecular dynamics. Recently, frameworks to construct accurate interatomic that combine a set of density functional theory (DFT) calculations with machine learning techniques proposed. One these methods is use compressed sensing derive sparse representation for the potential. This facilitates control accuracy potentials. In this study, we demonstrate applicability deriving potential ten elemental metals, namely,...
Machine learning interatomic potentials (MLIPs) based on a large data set obtained by density functional theory calculation have been developed recently. This study gives both conceptual and practical bases for the high accuracy of MLIPs, although MLIPs considered to be simply an accurate black-box description atomic energy. We also construct most MLIP elemental Ti ever reported using linearized framework many angular-dependent descriptors, which corresponds generalization modified embedded...
Prediction models of both the electronic and ionic contributions to static dielectric constants have been constructed using data from density functional perturbation theory calculations approximately 1200 metal oxides via supervised machine learning. We developed two types random forest regression for with ground-state crystal structures: one model requires only compositional information other also uses structural information. Although training included various atomic frameworks, prediction...
Oxygen vacancies play significant roles in various properties of oxide materials. Therefore, insights into the oxygen can facilitate discovery better To achieve this, we developed codes for high-throughput point-defect calculations and applied them to characterize 937 oxides. From resulting large dataset, analyzed vacancy structures formation energies constructed machine-learning regression models predict energies. We have found that are predicted using random forest with accuracies...
A transparent p-type semiconductor, CuAlO2, shows selective and reversible response to ozone gas at room temperature. All existing commercial semiconductor sensors are of n type. This study demonstrates the feasibility developing an inexpensive p type sensor. Types p–n junction may be fabricated using CuAlO2 n-type materials such as In2O3.
It is considered that a decrease of the ratio polychromatic erythrocytes (PCE) to normochromatic (NCE) (P/N) in micronucleus test an indicator bone marrow toxicity induced by mutagens. However, exact meaning fluctuation P/N not yet known. We have studied this point counting total number and nucleated cells following various treatments. The decreased gradually with time after administration mitomycin C. Our data suggest was attributable increase numbers denominator, i.e. NCE, caused rapid...
A marine unicellular aerobic nitrogen‐fixing cyanobacterium Synechococcus sp. strain Miarni BG 043511 was pretreated with different light and dark regimes in order to induce higher growth synchrony. pretreatment of two cycles 16 h each yielded good synchrony for 3 cell division cycles. Longer treatments decreased the degree shorter caused irregular division. Once synchronous culture established, distinct phases cellular carbohydrate accumulation degradation were observed even under...
Nonlinear phenomena on the soft material surface are one of most exciting topics chemical physics. However, only a few reports exist friction under accelerated movement, because between two solid surfaces is considered linear phenomenon in many cases. We aim to investigate how nonlinear motion affects surfaces. In present study, we evaluate frictional forces polytetrafluoroethylene (PTFE) resins using an advanced evaluation system. On PTFE surfaces, normalized delay time δ, which lag...
Zinc tin nitride (ZnSnN2) is attracting growing interest as a non-toxic and earth-abundant photoabsorber for thin-film photovoltaics. Carrier transport in ZnSnN2 consequently cell performance are strongly affected by point defects with deep levels acting carrier recombination centers. In this study, the revisited careful first-principles modeling based on recent experimental theoretical findings. It shown that does not have low-energy levels, contrast to previously reported results....
Antiperovskites have recently been attracting considerable attention because of their intriguing physical properties. We theoretically investigate polymorphism mixed-anion antiperovskites ${M}_{3}^{2+}{X}^{3\ensuremath{-}}{\mathrm{N}}^{3\ensuremath{-}}\phantom{\rule{4pt}{0ex}}(M$ = Mg, Ca, Sr, Ba; $X$ P, As, Sb, Bi) using the seven representative crystal-structure prototypes $AB{Z}_{3}$ compounds. Stable exploration for four unreported compounds...
Abstract Mixed-stack complexes which comprise columns of alternating donors and acceptors are organic conductors with typically poor electrical conductivity because they either in a neutral or highly ionic state. This indicates that conductive carriers insufficient mainly localized. In this study, mixed-stack uniquely exist at the neutral–ionic boundary were synthesized by combining (bis(3,4-ethylenedichalcogenothiophene)) (fluorinated tetracyanoquinodimethanes) similar energy levels orbital...
We report a direct topological reorganization between intermolecularly cross-linked polymers and single-chain nanoparticles (SCNPs) via thermal exchange reactions based on bis(2,2,6,6-tetramethylpiperidin-1-yl)disulfide (BiTEMPS) linkages. The network architecture of poly(hexyl methacrylate) incorporating BiTEMPS at the cross-linking points can be directly transformed to SCNPs merely by heating in dilute solution. Diffusion-ordered NMR spectroscopy GPC analyses confirmed smaller hydrodynamic...
The cluster analysis of materials categorizes them according to similarities based on the features materials, providing insight into relationship between materials. Conventional analyses typically use basic derived from chemical composition and crystal structure without considering target material properties such as bandgap dielectric constant. However, approaches do not meet demands for grading interest simultaneously with structural similarities. Herein, a clustering method grouping...
This paper proposes a method for selecting explanatory variables global solar radiation forecasting. The weather conditions affect generation output of renewable energy significantly. As result, smart grids increase the uncertainties caused by power injections photovoltaic (PV) and/or wind generation. To smooth grid operation, it is necessary to select predicted input forecasting model. In this paper, how in model PV discussed because Japan gives much higher priority framework future policy....
Machine-learning interatomic potential (MLIP) has been of growing interest as a useful method to describe the energetics systems interest. In present study, we examine accuracy linearized pairwise MLIPs and angular-dependent for 31 elemental metals. Using all optimal metals, show robustness frameworks, general trend predictive power MLIPs, limitation MLIPs. As result, obtain accurate elements using same framework. This indicates that use numerous descriptors is most important practical...
We investigate the native point defects in delafossite $\mathrm{Cu}M{\mathrm{O}}_{2}$ ($M=\mathrm{Al}$, Ga, In) using first-principles calculations based on Heyd-Scuseria-Ernzerhof (HSE06) hybrid functional approach. The Cu vacancies all systems show low formation energies and form relatively shallow acceptor levels, which would contribute mainly to $p$-type conductivity. hole compensation by donor-type does not essentially limit doping of under controlled growth conditions. In contrast,...
This paper proposes a new hybrid intelligent method for probabilistic short-term load forecasting (STLF) in power systems. It consists of Relevance Vector Machine (RVM) the statistical learning called Kernel and regression tree (RT) data mining. As preconditioned technique data, RT is used to classify into some clusters with similarity. After classifying clusters, RVM constructed predict one-step ahead loads at each cluster. one efficient Machines that extend Support (SVM) deal continuous...
The areal density of holographic memory can be increased by decreasing the aperture size optical system; however, this causes deterioration a reproduced image due to elimination high-frequency components image. We have developed two-dimensional partial response maximum likelihood (2D PRML), which is based on previously proposed "decision feedback Viterbi algorithm" and added new idea "constant-weight constraint (CW constraint)" "amplitude PR method". confirmed improvement in error rate using...
Materials screening by high-throughput first-principles calculations is a powerful tool for exploring novel materials with preferable properties. Machine learning techniques are expected to accelerate constructing surrogate models and making fast predictions. Especially, black-box optimization methods such as Bayesian optimization, repeating the construction of prediction model selection data points, have attracted much attention. In this study, we constructed an autonomous system using...
A combined computational and experimental study of La 2 SnO S 3 reveals carrier generation compensation mechanisms associated with its moderate n-type conductivity, where hydrogen impurities electron self-trapping play crucial roles.