Koh Takeuchi

Koh takeuchi

Koh Takeuchi is an assistant professor at Department of Intelligence Science and Technology, Kyoto University. Before joining the faculty, he was a research scientist of Learning and Intelligent Systems Research Group and Ueda Research Laboratory in Communication Science Laboratory of NTT during April, 2011-January, 2020. He obtained a Ph.D. degree in informatics in 2019 fromKyoto University in Japan.

MISSION

We seek to solve social problems through the development of artificial intelligence, especially machine learning, and the design of spatio-temporal data analysis systems for understanding urban and natural environments.

PROFILE＆VISION

After working on machine learning and data mining research at NTT Communication Science Laboratories for 10 years from 2011, I started working at Kyoto University from 2020.

With the boom of big data and deep learning in the 2010s, technologies related to artificial intelligence and data science have become familiar in our daily lives. However, looking back at the boom, I consider that our society is still in the process of change, with limited applications of advanced AI-based systems to solve social and industrial problems.

I am engaged in research on developing data analysis methods based on machine learning that focus on understanding the current state and predicting the future of urban and natural environments, and realizing information technologies that can be used for decision making. I am also interested in technologies for analyzing social data that address potential social biases in human behavior by maximizing social welfare while maintaining fairness.

I am working to design data analysis systems that incorporate the advanced technologies of our field in collaboration with the business community.

APPROACH

Developing Statistical Machine Learning Methods

Machine learning can be described as a technology that predicts unknown data relationships by automatically learning the hidden relationships from data given by humans. We study high-performance analysis methods by efficiently handling auxiliary information such as the location and time at which data was measured. In recent years, based on causal inference techniques in statistics, we have been researching techniques to effectively handle biases derived from individuals and attributes latent in data, which have attracted attention in economics and sociology.

Advancing Technology for Data Analysis Infrastructure Systems

Our goal is to apply data analysis techniques, such as data mining and data science, to real-world applications. By effectively handling information such as when, where, who, and what was measured, we can solve problems such as demand forecasting for commercial facilities, predicting traffic congestion in urban transportation networks, and extracting purchasing patterns using latent attribute information. We design data analysis systems in cooperation with companies and organizations, and conduct joint research and advising with the aim of realizing technologies that are useful to society.

News

2021/12/12

A Paper Accepted fo EAAI2022.

Naoki Miyaguchi, Koh Takeuchi, Hisashi Kashima, Mizuki Morita, Hiroshi Morimatsu.Predicting Anesthetic Infusion Events Using Machine Learning.Scientific Reports, 2022.

2021/11/19

A Paper Accepted fo Scientific Reports.

Naoki Miyaguchi, Koh Takeuchi, Hisashi Kashima, Mizuki Morita, Hiroshi Morimatsu.Predicting Anesthetic Infusion Events Using Machine Learning.Scientific Reports, 2022.

2021/11/03

We received JSAI Annual Conference Award 2021.

Koh Takeuchi, Ryo Nishida, Hisashi Kashima, Masaki Onishi, JSAI 2021

see more info

Get In Touch

Address

204, Research Bldg. No.7, Kyoto University, Yoshida-honmachi, Sakyo-ku, Kyoto, 606-8501.

E-mail

takeuchi<a.t.>i.kyoto-u.ac.jp

More Info

Profile / CV

News

Get In Touch