A5047435295- Matrix Theory and Algorithms
- Electromagnetic Scattering and Analysis
- Privacy-Preserving Technologies in Data
- Advanced Optimization Algorithms Research
- Face and Expression Recognition
- Numerical methods for differential equations
- Advanced Numerical Methods in Computational Mathematics
- Neural Networks and Applications
- Tensor decomposition and applications
- Remote-Sensing Image Classification
- Mobile Crowdsensing and Crowdsourcing
- Sparse and Compressive Sensing Techniques
- Advanced Clustering Algorithms Research
- Complex Network Analysis Techniques
- Cryptography and Data Security
- Polynomial and algebraic computation
- Advanced Causal Inference Techniques
- Explainable Artificial Intelligence (XAI)
- Model Reduction and Neural Networks
- Gamma-ray bursts and supernovae
- Stochastic Gradient Optimization Techniques
- Adversarial Robustness in Machine Learning
- Privacy, Security, and Data Protection
- Scientific Computing and Data Management
- Machine Learning and Data Classification
University of Tsukuba
2016-2025
Applied Mathematics (United States)
2018
RIKEN Center for Computational Science
2018
High Energy Accelerator Research Organization
2015
Nagoya University
2009-2012
Midorigaoka Hospital
2002
We present the first results of our spatially axisymmetric core-collapse supernova simulations with full Boltzmann neutrino transport, which amount to a time-dependent 5-dimensional (2 in space and 3 momentum space) problem fact. Special relativistic effects are fully taken into account two-energy-grid technique. performed two for progenitor 11.2M, employing different nuclear equations-of-state (EOS's): Lattimer Swesty's EOS incompressibility K = 220MeV (LS EOS) Furusawa's based on mean...
The growing amount of data and advances in science have created a need for new kind cloud platform that provides users with flexibility, strong security, the ability to couple supercomputers edge devices through high-performance networks. We built such nation-wide platform, called "mdx" meet this need. mdx platform's virtualization service, jointly operated by 9 national universities 2 research institutes Japan, launched 2021, more features are development. Currently is used researchers wide...
Recently, data collaboration (DC) analysis has been developed for privacy-preserving integrated across multiple institutions. DC centralizes individually constructed dimensionality-reduced intermediate representations and realizes via without sharing the original data. To construct representations, each institution generates shares a shareable anchor dataset its representation. Although, random functions well in general, using an whose distribution is close to that of raw expected improve...
Abstract Ensuring the transparency of machine learning models is vital for their ethical application in various industries. There has been a concurrent trend distributed designed to limit access training data privacy concerns. Such models, trained over horizontally or vertically partitioned data, present challenge explainable AI because explaining party may have biased view background partial feature space. As result, explanations obtained from different participants might not be consistent...
This paper proposes a data collaboration analysis framework for distributed sets. The proposed involves centralized machine learning while the original sets and models remain over number of institutions. Recently, has become larger more with decreasing costs collection. Centralizing analyzing them as one set can allow novel insights attainment higher prediction performance than that individually. However, it is generally difficult to centralize because large size or privacy concerns. does...
Observational studies enable causal inferences when randomized controlled trials (RCTs) are not feasible. However, integrating sensitive medical data across multiple institutions introduces significant privacy challenges. The collaboration quasi-experiment (DC-QE) framework addresses these concerns by sharing "intermediate representations"—dimensionality-reduced derived from raw data—instead of the data. Although DC-QE can estimate treatment effects, its application to remains unexplored....
We present a newly developed moving-mesh technique for the multi-dimensional Boltzmann-Hydro code simulation of core-collapse supernovae (CCSNe). What makes this different from others is fact that it treats not only hydrodynamics but also neutrino transfer in language 3+1 formalism general relativity (GR), making use shift vector to specify time evolution coordinate system. This means transport part our essentially relativistic although paper applied moving curvilinear coordinates flat...
Feature selection is an efficient dimensionality reduction technique for artificial intelligence and machine learning. Many feature methods learn the data structure to select most discriminative features distinguishing different classes. However, sometimes distributed in multiple parties sharing original difficult due privacy requirement. As a result, one party may be lack of useful information features. In this paper, we propose novel method which allows collaborative without revealing...
A fundamental problem in machine learning is ensemble clustering, that is, combining multiple base clusterings to obtain improved clustering result. However, most of the existing methods are unsuitable for large-scale tasks owing efficiency bottlenecks. In this paper, we propose a spectral (LSEC) method balance and effectiveness. LSEC, clustering-based efficient generation framework designed generate various with low computational complexity. Thereafter, all combined using bipartite graph...
Ensemble clustering has attracted much attention in machine learning and data mining for the high performance task of clustering. Spectral is one most popular methods superior compared with traditional methods. Existing ensemble usually directly use results base algorithms learning, which cannot make good intrinsic structures explored by graph Laplacians spectral clustering, thus obtain desired result. In this paper, we propose a new method clustering-based algorithms. Instead using obtained...
This paper proposes an interpretable non-model sharing collaborative data analysis method as a federated learning system, which is emerging technology for analyzing distributed data.Analyzing essential in many applications, such medicine, finance, and manufacturing, due to privacy confidentiality concerns.In addition, interpretability of the obtained model plays important role practical applications systems.By centralizing intermediate representations, are individually constructed by each...
Tensor completion using multiway delay-embedding transform (MDT) (or Hankelization) suffers from the large memory requirement and high computational cost in spite of its potentiality for image modeling. Recent studies have shown performance with a relatively small window size, but experiments sizes require huge amount cannot be easily calculated. In this study, we address serious issue, propose fast efficient algorithm. Key techniques proposed method are based on two properties: (1) signal...
Multi-source data fusion, in which multiple sources are jointly analyzed to obtain improved information, has attracted considerable research attention. Data confidentiality and cross-institutional communication critical for the construction of a prediction model using datasets medical institutions. In such cases, collaboration (DC) analysis by sharing dimensionality-reduced intermediate representations without iterative communications may be appropriate. Identifiability shared is essential...
Dimensionality reduction methods that project highdimensional data to a low-dimensional space by matrix trace optimization are widely used for clustering and classification. The problem leads an eigenvalue subspace construction, preserving certain properties of the original data. However, most existing use only few eigenvectors construct space, which may lead loss useful information achieving successful Herein, overcome deficiency loss, we propose novel complex moment-based supervised...
The demand for the privacy-preserving survival analysis of medical data integrated from multiple institutions or countries has been increased. However, sharing original is difficult because privacy concerns, and even if it could be achieved, we have to pay huge costs cross-institutional cross-border communications. To tackle these difficulties on parties, this study proposes a novel collaboration Cox proportional hazards (DC-COX) model based framework horizontally vertically partitioned...