A5043798091- Matrix Theory and Algorithms
- Electromagnetic Scattering and Analysis
- Advanced Optimization Algorithms Research
- Quantum Computing Algorithms and Architecture
- Quantum and electron transport phenomena
- Neural Networks and Applications
- Polynomial and algebraic computation
- Quantum Chromodynamics and Particle Interactions
- Parallel Computing and Optimization Techniques
- Advanced NMR Techniques and Applications
- Complex Network Analysis Techniques
- Blind Source Separation Techniques
- Advanced Numerical Methods in Computational Mathematics
- Face and Expression Recognition
- Numerical methods for differential equations
- Machine Learning in Materials Science
- Theoretical and Computational Physics
- Quantum Information and Cryptography
- Bioinformatics and Genomic Networks
- Numerical Methods and Algorithms
- Advanced Chemical Physics Studies
- Single-cell and spatial transcriptomics
- Electromagnetic Simulation and Numerical Methods
- Advanced Clustering Algorithms Research
- Gene expression and cancer classification
University of Tsukuba
2016-2025
Applied Mathematics (United States)
2018
High Energy Accelerator Research Organization
2015
The location of microRNAs (miRNAs) in cells determines their function regulation activity. Studies have shown that miRNAs are stable the extracellular environment mediates cell-to-cell communication and located intracellular region responds to cellular stress environmental stimuli. Though situ detection techniques made great contributions study localization distribution miRNAs, miRNA subcellular role still progress. Recently, some machine learning-based algorithms been designed for...
We consider an eigensolver for computing eigenvalues in a given domain and the corresponding eigenvectors of large-scale matrix pencils. The Sakurai-Sugiura (SS) method is based on complex moments by contour integrals inverses with several shift points. This has good parallel scalability, suitable massively environments. SS parameters, choice these parameters crucial achieving high accuracy performance. discuss some numerical properties method, present efficient parameter estimation...
This study aims to create a deep learning-based classification model for cervical cancer biopsy before and during radiotherapy, visualize the results on whole slide images (WSIs), explore clinical significance of obtained features. included 95 patients with who received radiotherapy between April 2013 December 2020. Hematoxylin-eosin stained biopsies were digitized WSIs divided into small tiles. Our adopted feature extractor DenseNet121 classifier support vector machine. About 12 400 tiles...
Some kinds of eigensolver for large sparse matrices require specification parameters that are based on rough estimates the desired eigenvalues. In this paper, we propose a stochastic estimation method eigenvalue distribution using combination estimator matrix trace and contour integrations. The proposed can be easily parallelized applied to which factorization is infeasible. Numerical experiments executed show perform at low computational cost.
We introduce a novel method to obtain level densities in large-scale shell-model calculations. Our is stochastic estimation of eigenvalue count based on shifted Krylov-subspace method, which enables us huge Hamiltonian matrices. This framework leads successful description both low-lying spectroscopy and the experimentally observed equilibration Jπ=2+ 2− states 58Ni unified manner.
Triple negative breast cancer (TNBC) is associated with worse outcomes and results in high mortality; therefore, great efforts are required to find effective treatment. In the present study, we suggested a novel strategy treat TNBC using mesenchymal stem cell (MSC)-derived extracellular vesicles (EV) transform behaviors cellular communication of cells (BCC) other non-cancer related tumorigenesis metastasis. Our data showed that, BCC after being internalized EV derived from Wharton's Jelly...
The performance of some nonlinear eigenvalue problem solvers can be increased by setting parameters that are based on rough estimates the desired eigenvalues. In present paper, we propose a stochastic method for estimating density problems analytic matrix functions. proposed uses unbiased estimation traces and contour integrations. Its is evaluated through numerical experiments.
We propose an efficient way to calculate the electronic structure of large systems by combining a large-scale first-principles density functional theory code, Conquest, and interior eigenproblem solver, Sakurai-Sugiura method. The Hamiltonian charge are obtained eigenstates Hamiltonians then Applications hydrated DNA system adsorbed P2 molecules Ge hut clusters on Si substrates demonstrate applicability this combination with 10,000+ atoms high accuracy efficiency.
High-throughput sequencing technologies have enabled the generation of single-cell RNA-seq (scRNA-seq) data, which explore both genetic heterogeneity and phenotypic variation between cells. Some methods been proposed to detect related genes causing cell-to-cell variability for understanding tumor heterogeneity. However, most existing separately, without considering gene interactions. In this paper, we a novel learning framework interactive groups scRNA-seq data based on co-expression network...
Abstract Accurate prediction of drug-target interactions (DTIs) can reduce the cost and time drug repositioning discovery. Many current methods integrate information from multiple data sources target to improve DTIs accuracy. However, these do not consider complex relationship between different sources. In this study, we propose a novel computational framework, called MccDTI, predict potential by multiview network embedding, which heterogenous target. MccDTI learns high-quality...
The atomic descriptors used in machine learning to predict forces are often high dimensional. In general, by retrieving a significant amount of structural information from these descriptors, accurate force predictions can be achieved. On the other hand, acquire higher robustness for transferability without overfitting, sufficient reduction should necessary. this study, we propose method automatically determine hyperparameters aiming obtain while using small number descriptors. Our focuses on...
We propose a fast and efficient approach for solving the Bogoliubov-de Gennes (BdG) equations in superconductivity, with numerical matrix-size reduction procedure proposed by Sakurai Sugiura [J. Comput. Appl. Math. 159, 119 (2003)]. The resultant small-size Hamiltonian contains information of original BdG given energy domain. In other words, present leads to construction low-energy effective theory superconductivity. combination polynomial expansion method allows self-consistent calculation...
We propose a methodology based on unsupervised learning with the two-step locality preserving projections (TS-LPP) method to detect differences in local structures silica. Subtle changes can be detected.
We propose the reduced-shifted Conjugate-Gradient (RSCG) method, which is numerically efficient to calculate a matrix element of Green's function defined as resolvent Hamiltonian operator, by solving linear equations with desired accuracy. This method does not solution vectors but directly resolvent. The elements different frequencies are simultaneously obtained. Thus, it easy exception value expressed Matsubara summation these elements. To illustrate power our we choose nano-structured...
Owing to the advances in computational techniques and increase power, atomistic simulations of materials can simulate large systems with higher accuracy. Complex phenomena be observed such state-of-the-art simulations. However, it has become increasingly difficult understand what is actually happening mechanisms, for example, molecular dynamics (MD) We propose an unsupervised machine learning method analyze local structure around a target atom. The proposed method, which uses two-step...
Complex band structures (CBSs) are useful to characterize the static and dynamical electronic properties of materials. Despite intensive developments, first-principles calculation CBS for over several hundred atoms still computationally demanding. We here propose an efficient scalable computational method calculate CBSs. The basic idea is express Kohn-Sham equation real-space grid scheme as a quadratic eigenvalue problem compute only solutions which necessary construct by Sakurai-Sugiura...
To estimate the number of eigenvalues a Hermitian matrix that are located in given interval, existing methods include polynomial filtering and rational filtering. Both approaches based on stochastic approximations for trace. In this paper, we analyze method is which solutions to linear systems approximated by Krylov subspace method. Our analysis numerical experiments indicate effective when sparsely distributed target interval.