This note is intended to avoid mismatches and misunderstandings regarding the graduate school application.
Please carefully read this note and follow the instructions.
The applicants for master or doctor courses under supervision of our lab are requested to
Take a survey and understand what we are doing.
See our webpage and publication list.
Contact us before the application to discuss the research topics in the graduate course
Have the sufficient knowledge and training of (if you want to work on theory and modeling)
Mathematics: linear algebra, calculus, differential equation, stochastic processes, etc
Physics: thermodynamics and statistical physics
Have the sufficient knowledge and training of (if you want to work on data analysis)
Mathematics: statistics, Bayesian inference, machine learning theory, etc.
Informatics: bioinformatics, programming, and data structure.
Some applicants misunderstand that we are working on deep learning or machine learning. We sometimes use it in our project to understand biological phenomena, but we are not interested in developing the deep learning method itself. We are working on understanding the design principle of biological systems from the viewpoint of physics and mathematics. Please carefully see our website for individual projects and topics. Some papers on DL (e.g., [1], [2]) were mainly conducted by our collaborators not by us. So, we cannot supervise such topics. Also, we no longer work on bioimage analysis such as [3], [4].
More importantly, we do not have the GPU servers necessary to run the state-of-the-art DL algorithms for image and language. The computational power required to run these algorithms has been increasing, which is affordable only at labs where they work on DL as the primary target with sufficient grant support. If you really want to work on and practice DL for image and language, it is crucial to find a lab with adequate equipment (it is not our lab).
Some applicants also misunderstood that we are working on medical applications. But we are mainly interested in science rather than applications. While we work on immunology, it is purely for unveiling the principle of why the immune cells can recognize, learn, and remember invading pathogens. For medical applications, the accessibility to data is crucial. Without that, we cannot do any research. If you really want to do medical applications, there are many other good labs that have established tight connections with medical doctors and hospitals. Please survey such places.
Theory, and informatics can be learned by yourself because they do not require any expensive or special equipment. This means you can start studying what you want to do by yourself anytime if you really want to do it. We now have many good and free textbooks and movies on basic knowledge of math, physics, and informatics published by universities worldwide.
We expect that the applicants have already taught themselves before application.