- Please see

Basic theories of signal processing such as Fourier and wavelet transforms, principal component analysis and independent component analysis are explained. Practical examples will be given, using sound processing, image processing, biomedical measurement signals and astronomical data analysis as examples.

The course will discuss the basics of the inference for un-normalized models and sparse modeling and its applications. Also, methodologies of formulating problems into "solvable" forms for various real-world problems are discussed.

Research work is directed on the methodology of statistical machine learning such as deep learning and kernel methods for analyzing large high-dimensional data. The course aims to provide students with knowledge of machine learning theory and methods, and the ability to apply machine learning methods to data analysis. The course will be conducted in the form of seminars, and students will be evaluated based on their presentations.

This course deals with various data analysis methods for practical applications, including statistical machine learning, deep learning, Bayesian inference, Python/R programming, etc., through several case studies including materials data analysis.

Students will learn about the concept and implementation of uncertainty quantification for Bayesianinference of physical models. Examples of model comes from structural engineering, geotechnicalengineering, etc.

Lectures or seminars on basic topics of combinatorial optimization and algorithm design will be given. The topics will include linear programming, matching, network flows, matroids, and submodular functions, but other topics may be covered depending on the interests of the students.

Lecture and research guidance on nonlinear optimization algorithm and analysis, with a focus on those suitable for real problems in large-scale machine learning

This course is intended to serve an introduction to systems design and focuses on the theoretical aspects of mathematical optimization based on convex analysis, duality theory, numerical linear algebra, and functional analysis.

This course deals with advanced computational statistics with their applications. Examples are specialized topics in Markov Chain Monte Carlo, sequential Monte Carlo, bootstrap, and their applications in Bayesian statistics and model selection.

Survival analysis deals with patients' and/or equipments' survival time data. When analyzing survival data, one typically entounters incompletely observed survival times that are censored or truncated. This type of survival data arises in a variety of fields, including medicine, reliability, ecology, insurance, economics, finance, and marketing. In this course, we study statistical models and inference methods based on survival data.

The course covers some theoretical and methodological topics of Bayesian computation. Markov chain Monte Carlo methods are also covered.

One of the topics below will be chosen: (1) Distribution theory, asymptotic theory, statistical inference, (2) Gaussian random processes/fields, (3) Contingency table and graphical model, (4) Differential and integral geometric approach to statistics, (5) Algebraic statistics, (6) Random matrices, (7) Mathematics in statistics including convex analysis, combinatorics, and measure theory.

This course is on the foundamental mathematical theory of point processes. It introduces basic concepts and theories related point processes,
including random measures, Janossy measures, Janossy density, Campbell measure, moment measure, conditional intensity, Papangelou intensity, and Palm intensity. Its objectives include:

(1) To learn the point process theory from the viewpoint of modern probability theory. (2) To understand the theoretical relationship between each type of point processes.

After outlining asymptotic theory for regular statistical models, asymptotic theory for locally conic models will be introduced and the geometric methods required in their evaluation will be explained. Alternatively, after outlining the basic methods for causal inference, semiparametric approaches using propensity scores will be introduced and information criteria for causal inference models will be explained. The goal of the lecture is to provide fundamentals and developments of statistical asymptotics or causal inference.

This course introduces the estimation and testing of regression models with nonlinearities and related statistical theory as basic topics, and treats sparse modeling, robust statistics, missing data analysis, deivergence-based inference as advanced topics.

This course discusses algebraic and combinatorial methods arising in stochastic computing for statistical inference around count data.

This course deals with either of the following two topics based on interests of students: (i) the theory of well-known probability distributions and related statistical models, or (ii) the theory of generalized linear models and its applications to real problems.

Students will learn the mean field theory for random systems: (1) how to analyze magnetic models (2) basic analytical methods

Probability theory is basic mathematics for statistical science. This course deals with mathematical topics that link to statistical applications.

This course deals with numerical algorithms in mathematical optimization and related areas. Specifically, we choose some topics in numerical algorithms for continuous optimization problems, matrix and eigenvalue problems, and so on, and discuss their mathematical foundations.

This course focus on the study of statistical models, such as linear mixed-effects models and their extensions, used in the longitudinal data analysis in which a response variable is measured repeatedly over time for multiple subjects. The course will also focus on the study of research designs, such as randomization, and on the study of statistical methods used in actual problems.

In this course we will discuss the fundamentals of convex analysis such as separation theorems, subdifferential theory and several duality results. At the end, we will discuss some extensions and related topics. Alternatively, we might also discuss modelling and theoretical aspects of conic linear programs.

We study the theory of semiparametric inference, its application and (or) some related topics. The standard knowledge of mathematical statistics and the basic mathematical skill on calculus, abstract linear algebra, metric space (or general topology) and probability theory are required.

This lecture introduces statistical methods for spatial data and their applications. Specifically, methods in geostatistics, spatial econometrics, and relevant areas are introducesd together with their implementation with R.

This course deals with the basic probabilistic framework for statistical treatment of natural language and related discrete data. It will cover the basic statistical and computational methods necessary to understand the characteristics of high-dimensional discrete data.

This course discusses topics in probability and statistics in higher and infinite dimensions.

This is a course of seminar and practice on sequential data assimilation methods or variational data assimilation methods. On the basis of the state-space model, students derive the sequential methods or the variational methods, and implement the procedure.

The goal of this course is to provide students with the ability to understand and implement the various statistical methods used in environmental data analysis. Students will read the textbooks and attend lectures on specific topics.

After covering the basics of stationary time series models that were not fully covered in Introduction to Time Series and Spatial Modeling, causality analysis based on multivariate autoregressive (VAR) models, impulse responses, etc. will be covered. After introducing unit root tests, we will expand our understanding to cointegration models. With financial time series (rate of return data) in mind, the conditional heteroscedasticity models will be outlined. In addition, time-varying variance modeling will be discussed in relation to local stationary AR models and non-Gaussian filters. Computer exercises using R or R Shiny applications will be conducted once after every two or three lectures.

Lectures on statistical computing using a parallel computer will be given in this course. In particular，the following subjects will be discussed: problems which requires huge matrices，the particle filter using a parallel computer，and implementation of the ensemble Kalman filter on a parallel computer.

This course deals with recent relevant topics on biostatistics, especially, (i) Biostatistical methodology on clinical and epidemiologic studies, (ii) Designs and analyses of clinical trials, (iii) Evidence synthesis methods, and (iv) statistical analyses of large-scale genomic data.

The digital age has fostered the broadcasting of an ever increasing quantity of complex multimedia documents， be it through the internet or more versatile electronic channels. These evolutions have called for new tools and technologies to classify and analyze multimedia contents. We study in this course algorithms which are useful for these tasks.

This course covers major privacy-preserving techniques such as anonymization, differential privacy, and statistical disclosure control, which realize safe analysis of big data. We also cover topics on machine learning security with emphasis on empirical evaluation methods.

Concrete cases will be presented and research lectures will be given on theoretical methodologies for quantification of credit risk, risk assessment of financial markets, and investment strategies. In particular, students will acquire practical knowledge of statistical models that are consistent with relevant laws and regulations such as the Basel Accord and corporate accounting, forecasting using stochastic processes and time-series models, and risk assessment through the use of actual data.

In this lecture, we will outline the basics of stochastic processes, including Poisson processes, random walks, and Brownian motion, while showing examples of applications of stochastic processes. In addition, we will explain the mathematical economic analysis of resource management risks by applying macro-renewable resource supply forecasting and option theory, targeting renewable resources such as forests.

This course deals with data analysis of genetic information by applying the methods of statistical science. To infer the phylogenetic relationships of organisms, we compare homologous genes between organisms. Explanations and exercises will be given on how to construct a model for that purpose and maximum likelihood estimation of phylogenetic trees.

This course provides an introduction to stochastic processes, with their applications to real-world situations. This course covers elementary stochastic processes such as Gaussian, Poisson, Markov and renewal processes.

In this course, I will lecture spatial modeling for spatial data such as geostatistics, lattice model, directional statistics, and spatial point process, as well as random partitioning and random packing. Unlike the generalized linear model and its related statistics, a likelihood of a spatial model is not a simple product over independent data, which is one of essences for spatial modeling.

Shannon's information theory is lectured as a basic theory for analysis of information sources. The amount of information, entropy, and communication sysytems are discussed. And actual radio communication systems are lectured.

This course covers systematic explanations of practical methodologies of survey design for organizations or regions.

The course will cover the design of data collection, with a focus on sample survey methods, and statistical inference under that design. The goal is to deepen understanding of the importance of the design stage in the research process.

The lecture will discuss deterministic and probabilistic approaches to time series data analysis and modeling. In addition, students take turns reading books and papers about detecting significant signals, spatial correlations, and causality in time series data.

This course deals with differential privacy (DP), which is known as a de facto standard privacy notion for privacy protection. We first study some basic topics, such as the privacy properties of DP, the Laplace and exponential mechanisms, the composition theorems, and the SVT (Sparse Vector Technique). Then, we study more advanced topics, such as various extensions/variants of DP, the local model, and the shuffle model.

This is a general research course of statistical science. Students are requested to present progress of their research by giving seminars and talks.

This is a general course of statistical science. Students are requested to attend the statistical mathematics seminar held at the institute of statistical mathematics to learn various recent developments in statistical science.