AI for Science: The Science of New Scientific Methods
Dr. Ichigaku Takigawa
Department of Computational Biology and Medical Sciences, Graduate School of Frontier Sciences, The University of Tokyo.
[Abstract]
Machine learning has been put to practical use as a core technology supporting audio, image, video, and even language processing, and has deeply penetrated into everyday tools and services. The next frontier is natural science, which is being actively researched worldwide as AI for Science. In this seminar, I will introduce several research cases from the perspective of having been engaged in machine learning research for more than 20 years and at the same time being involved in practical research in natural science fields such as life science and chemistry. The seminar will also provide an overview of current developments at the intersection of life sciences and machine learning, as well as new opportunities and challenges posed by the rapid development of machine learning. In addition, while the rapid evolution of generative language models in recent years, starting with ChatGPT, has exceeded even the expectations of experts, there is criticism and debate that it has not led to any understanding of “why humans can use language. In considering what is scientific understanding and what is a scientific model, this fact raises skepticism about the very nature of constitutive understanding, and I would like to discuss this meta-scientific perspective if time permits.
Date： May 29, 2025 (Thu.) 15:00 – 16:30
Place： Zoom and Room 412, Faculty of Science Building 3, University of Tokyo
Host： Shinya Kuroda
If you would like to attend, please email us at info.kuroda-lab [at] bs.s.u-tokyo.ac.jp and we will send you the Zoom URL. Please use your institution’s e-mail address and let us know your name and affiliation.