- Day 1 (Nov.17), 13:30-16:00
- Room 1101 (11th floor)
- Speakers: Ryo Yoshida (Institute of Statistical Mathematics), Koji Tsuda (University of Tokyo)
The ability of machine learning (ML) models trained on massive amounts of data has reached or even outperformed humans in intellectually demanding tasks across various fields. As such, ML has received considerable attention in manufacturing to reap substantial time and cost savings in many potential applications in industry and science. In this tutorial, we introduce a set of ML technologies that would be a key driver to the next frontier of creative design and manufacturing. The primary objective is to identify a set of parameters, such as the structure of materials and process parameters for the devise manufacturing, such that resulting response variables meet arbitrary given requirements. In general, a ML workflow consists of two steps; the first step is to build a prediction model that describes forwardly the response variables as a function of the input parameters, and this forward model is inverted to the backward one. With the given backward model conditioned by the design target, a set of parameters that exhibits the desired response is computationally explored. In this tutorial, some outstanding successes of ML are demonstrated with examples from materials science. The topics cover novel applications of deep learning technologies for designing materials structures or synthetic routes, transfer learning for overcoming a limited supply of materials data, Bayesian optimization frameworks that integrate ML and data from computer experiments such as the first principles calculation, and so on. The first speaker, Prof. Tsuda at the University of Tokyo, shows some outstanding progresses made by state-of-the art ML technologies for inorganic solid-state materials. The second speaker, Prof. Yoshida at the Institute of Statistical Mathematics, presents emerging new technologies for creative design and discovery of new functional molecules.
Ryo Yoshida, a Professor for Department of Data Science at the Institute of Statistical Mathematics (ISM), has served as the director of Data Science Center for Creative Design and Manufacturing in ISM since the center’s opening in July 2017. After receiving his Ph.D. in Statistical Mathematics from the Graduate University for Advanced Studies in 2004, he worked as a Project Assistant Professor for the Human Genome Center at Institute of Medical Science, the University of Tokyo – a position he has maintained after joining the ISM in 2007. In addition, he serves as an invited researcher for National Institute for Materials Science (NIMS). He received the IIBMP Research Encouragement Prize (2016 and 2017). He has the experience of using his expertise in data science for research work in both biology and materials science. He plays leadership roles in several cutting-edge research projects, including the Japan Science and Technology Agency (JST)’s the “Material Research by Information Integration” initiative (2015-present). He is also conducting the development of XenonPy–a machine learning platform for materials science. He is devoted to foster and practice machine learning technologies for design and manufacturing through industry-academia collaboration.
Koji Tsuda received B.E., M.E., and Ph.D degrees from Kyoto University, Japan, in 1994, 1995, and 1998, respectively. Subsequently, he joined former Electrotechnical Laboratory (ETL), Tsukuba, Japan, as Research Scientist. When ETL was reorganized as AIST in 2001, he joined newly established Computational Biology Research Center, Tokyo, Japan. In 2000–2001, he worked at GMD FIRST (currently Fraunhofer FIRST) in Berlin, Germany, as Visiting Scientist. In 2003–2004 and 2006–2008, he worked at Max Planck Institute for Biological Cybernetics, Tübingen, Germany, first as Research Scientist and later as Project Leader. Currently, he is Professor at Department of Computational Biology and Medical Sciences, Graduate School of Frontier Sciences, the University of Tokyo. He is also affiliated with National Institute of Material Science (NIMS) and RIKEN Center for Advanced Intelligence Project.