Takuo is a first year PhD student in the School of Mathematics, Statistics and Physics, Newcastle University under the supervision of Professor Chris Oates. He started his doctoral studies at The Alan Turing Institute in October 2019. He previously obtained M.Eng. and B.Eng. degrees in the Department of Electrical Engineering and Bioscience at Waseda University in Tokyo, Japan. Before starting his PhD, he worked as a research assistant at both RIKEN Center for Advanced Intelligence Project and The Alan Turing Institute in a short period. His research interests involve in probabilistic numerics, statistical machine learning, Bayesian statistics, and, theory of reproducing kernel Hilbert space.
The nascent research field of probabilistic numerics aims to exploit the mathematics of probability and statistics to develop a unified framework in which error accumulation due to the repeated use of numerical methods (such as quadrature, numerical optimization and the numerical solution of differential equations) in a computational procedure can be analysed. In this approach, numerical methods are considered to be statistical estimation procedures and concepts from statistical theory can be deployed to achieve diagnosis and control of numerical error. The probabilistic numerics agenda is producing cutting-edge research at the interface of statistics, machine learning, engineering, and computer science. Takuo is particularly interested in probabilistic numerical methods to construct deeply hierarchised models such deep neural networks, which has a potential to exploit statistical advantages of complex learning models for numerical methods and vice versa. His research interests cover statistical machine learning, Bayesian statistics, and, reproducing kernel Hilbert space.