April 1, 2019

From astrophysics to biometric authentication

Biometrics Laboratories
Makoto Takamoto

For about 10 years, Makoto Takamoto had conducted research in high-energy astrophysics at university in Japan and at the Max Planck Institute in Germany. In 2018, after receiving an award for outstanding achievement in research using the K supercomputer, Takamoto joined NEC. At present, he is enthusiastically pursuing research in biometric authentication.

From astrophysics researcher to biometric authentication researcher

Before entering NEC in 2018, I had carried out research in astrophysics for about 10 years, including my time at graduate school and at a research institute in Germany. When my research entered a period of stability, I decided to take the opportunity to enter NEC. At the time, I thought that I would like to pursue machine learning and artificial intelligence as new research themes.
I became interested in these fields through my research in astrophysics. In cutting-edge astrophysics research that uses numerical simulations, data is handled in enormous volumes on the order of petabytes, which are larger than terabytes. As might be expected, such vast amounts of data pose limitations on the data analysis that can be performed by humans. Consequently, my calculations were performed using the K supercomputer, which is the fastest in Japan. As part of the analysis, I used elementary machine learning technology. While working with the technology, I had a growing interest in this field, considering that the technology would make giant leaps forward in the future. I started thinking that I might like to take it up as a new research theme.
The transition from astrophysics to biometric authentication is a pretty significant change. The biggest difference is the sense of time involved in the research. In astrophysics, the theories can in some cases take 100 or even 1,000 years to verify. The main research subjects involve phenomena that occur at distances of over 1,000 light-years from our planet. However, in biometric authentication, if all goes well, the technology can spread all around the world a year after it is developed (laughter). I think that this sense of speed is a big difference between them.
Another big difference is the knowledge, so I had to learn a lot of new things. However, the basic approach to research and the mathematical fundamentals are the same. With support from my superiors and peers, I have been able to advance my career smoothly.

Striving for new possibilities by drawing on expertise from other fields

I am currently working on biometric authentication that uses face recognition, as a member of the team led by Hitoshi Imaoka. In 2018, I was also involved in the development of an engine for determining face orientation in a face recognition system.
I am still getting accustomed to conducting research in a new field, but eventually I would like to make use of my research expertise to create new technology. I think that knowledge and concepts from different fields are an extremely powerful tool, regardless of the area. Knowledge and concepts that did not previously exist in that area can cultivate new applications and ideas. As Nobel Prize winning scholars often say, thinking unconventionally and coming up with astonishing ideas can lead to new innovations. In the fields of machine learning and artificial intelligence, there is a large community of researchers and the pace of development is extremely rapid. For this reason, as long as doing the same things as other companies with large-scale operations, we might go down to defeat. There is remarkable technological evolution in deep learning, which is at the core of the present machine learning field. The question now is what to do with this technology, or how to come up with an approach that nobody has considered, from all the countless possibilities. In machine learning and artificial intelligence research, this is the key.
I believe that, as somebody with experience in astrophysics research, my way of thinking and knowledge can lead to the pioneering of new possibilities in machine learning and biometric authentication technology. Now, as I prepare to start my second year at NEC, I hope to pursue this possibility more ambitiously.

Work environment that is conducive to raising children

Another thing that makes me feel glad about entering NEC is the extremely comfortable work environment. My superiors recommend that I take my vacation time, and they are very mindful of my working hours. In the time that I have been here, I have never been forced to work overtime.
I am presently employed under the discretionary work system, so I have a flexible work style that allows me to adjust my schedule around the end of business hours of 5:35 PM. This gives me plenty of time to spend with my family after I return home, so that I can take a bath with my first child who was born in February 2018, and cook to feed the baby. NEC has a childcare leave system that allows employees to take time off work for childcare, regardless of their gender. It is also a large company, so a variety of benefits and systems are available including social security. I still do not have a good grasp of what is offered (laughter).
Before I entered NEC, I was working as a post-doctorate at a university. Compared to that situation, the environment at NEC is much more comfortable to work in. It is very helpful to be able to work in a good working environment that even supports raising children.