Welcome

I am Kengo Kamatani Professor at ISM, Japan.

My background is Mathematical Sciences, and my main interest is in Bayesian Computation. Bayesian Statistics is a branch of Statistics (or Data Science), and Bayesian Computation helps the evaluation of the posterior inference, which is the most important object for Bayesian Statistics. See monographs such as Doing Bayesian Data Analysis by Kruschke, and Bayesian Data Analysis by Gelman et al. See also The International Society for Bayesian Analysis webpage. You might be interested in a nice blog by Prof. A. Gelman and another excellent blog by C. P. Robert.

I am interested in the theoretical aspects of Bayesian Computation. More specifically, I am working on the analysis of Markov chain Monte Carlo, Particle Filter/Sequential Monte Carlo, and other related methods. In addition, I am interested in the inference of stochastic processes. Here are some recent papers: 

• Boomerang sampler (joint work with J. Bierkens, S. Grazzi and G. O. Roberts)

•  Non-reversible guided Metropolis-Hastings kernel. (joint work with X. Song)

• An efficient strategy for the Markov chain Monte Carlo in high-dimension with a heavy-tailed target probability distribution.

I am currently working on a research project  JST CREST  Modeling Methods allied with Modern Mathematics  (Research Director: Nakahiro Yoshida, U. Tokyo). This project is aiming for developing methodologies together with user-friendly tools (YUIMA Project) and for a better understanding of statistical inference for complicated dependent systems. 

I am also working on a five-year project JSPS Grant for Scientific Research (B) 20H04149 in April 2020.  In this project, I will develop robust Monte Carlo methodologies based on Hamiltonian dynamics. 

Another three-year project  JSPS Grant for Challenging Exploratory Research 21K18589 has also been started in April 2021. In this project, I will develop practical methods for scalable Bayesian inference. 

A new JST CREST project has started in October 2021. The theme of the project is "New developments in statistics for stochastic systems toward data science for large-scale spatiotemporal dependence" and the research director is Prof. Nakahiro Yoshida. 

Collaboration with the Railway Technical Research Institute is also continuing.