Research Topic Summary
The development of measurement and computer technologies has made it possible to acquire big data in various fields of science and technology. The approach that aims to make new discoveries by analyzing big data is called data-driven science and technology, and is expected to become the fourth approach following theory, experiment, and simulation-based approaches. The technology to create computer programs to analyze big data is called machine learning. We contribute to society through research and education on machine learning and the practice of data-driven approaches in science and technology.
AI and Machine Learning Method Development
Hypothesis Discovery by Machine Learning
Data-driven approaches have a potential to discover promising hypotheses that cannot be easily recalled by human experts. We are working on theoretical analysis and method development of pattern mining and generative modeling techniques to enhance data-driven hypothesis discovery.
Reliability Quantification of Machine Learning-Driven Knowledge
Complex machine learning models, such as deep neural networks, are capable of making highly accurate predictions, but their complexity raise the problems in terms of explainability and reliability. We are developing methods for statistical inference for complex machine learning models.
Experimental Design by Machine Learning
In scientific research and technological development, new knowledge can be obtained by analyzing data related to the subject of research and development. We are developing methods for systematic design of experiments in science and technology based on the knowledge obtained from data analysis.
Data-Driven Science and Technology Practices
AI for Manufacturing
The use of digital technology (DX) has become indispensable in the field of manufacturing. We are working with industrial partners to create data-driven manufacturing technologies by introducing AI and machine learning to the field of manufacturing.
AI for Biomedical Science and Technology
The current mainstream in life science is to understand biological phenomena quantitatively by analyzing comprehensive large-scale biological data, such as genetic information. We contribute to the life science field through collaborative research with medical and biological researchers.
AI for Material Science and Engineering
Discovering new materials with desirable properties is challenging. We are working on a data-driven approach to effectively discovering novel materials (called material informatics) through collaborative research with materials scientists.