March 26, 2021

Conducting Research That Anticipates Social Implementation and Commercialization

Data Science Research Laboratories
Kunihiro Takeoka

After receiving a master's degree in engineering, Takeoka joined NEC in April 2018. His work spans the fields of machine learning and natural language processing as he continues to produce results in the research of technologies which automate the integration of human knowledge with data and between different types of data.

Consistent experience from research to commercialization

I am now conducting research in machine learning and natural language processing at NEC's Data Science Research Laboratories. Our team's work produced a technology that understands the meaning of data.
Last year, I was also involved in the launch of the "NEC Data Enrichment" service which was based on this technology. It was my first experience with commercialization, so it was extremely stimulating and interesting. Becoming able to think about technology from different perspectives was an extremely valuable experience. For example, what customers require is not technology "novelty." If they can solve issues, even existing technologies are fine. Taken to an extreme, if issues can be efficiently solved even with manpower without relying on technology, customers are fine with that. Research is only required when issues cannot be solved. I realized that, on the contrary, new research themes are created to solve social issues precisely from this perspective.
Moreover, delivering a prototype to the customer and receiving feedback enables you to conduct further research and also advance the development in a speedy manner. Because the global speed of development in the field of machine learning is accelerating, research and development which integrates the customers with the development team is important. If a technology is developed, how can you quickly turn it into a service? In my current team as well, we are promoting development over an extremely short span of time.

Ability to balance both academic research and social application

To tell the truth, I joined my current research team as an intern during my student days. The opportunity arose when I participated in a machine learning conference and Dr. Oyamada, my current boss, invited me to join the team. Because I was already involved in researching the technology to understand the meaning of data during the latter half of my internship, he allowed me to fully participate in the research from that time.
I later participated in internships with several other companies, but I felt that NEC was number one in terms of their connections with academia. It is probably unusual for a company to recommend conference participation and the submission of research papers to such a degree. Of course, many companies are focused on development rather than research. After all, development itself has many interesting aspects. However, in my case, I sensed the appeal of thinking about real-world applications while also being able to take on the challenge of academic research and decided to join NEC.
In addition, NEC is a company which has various business areas. In conducting machine learning research, the ability to examine every possible application and not just specialized areas such as security and document processing was also appealing. Being able to conduct a wide range of research without shutting out possibilities is a positive aspect.
In addition, NEC has many excellent researchers in a variety of fields. I think that I am very blessed to be in an environment with so many experts around where you can think up some idea and know exactly the right person to talk to about a specific field. For example, I can go ask researchers who are working on the core areas of machine learning about cutting-edge theories or debate with researchers in different fields about security or biometric authentication, etc. An environment such as this is extremely stimulating and efficient for conducting research.

Researching while anticipating a path to social implementation

I would like to quickly become the kind of researcher who can imagine how a solution to an issue can be achieved. As I just mentioned, the purpose is to ultimately solve issues through the social implementation of technology. As long as you can achieve the purpose, it does not matter what the required technologies are. Within that process, you discover the necessary research and carry it out. My goal is to become the kind of researcher who can clearly foresee such a path.
My boss, Dr. Oyamada, is such a person. He is an excellent researcher who continues to reliably produce results at international conferences, but he is also an extremely capable developer who effortlessly rewrites prototype code increase the processing speed many times over. In addition, when it comes to commercialization, I read marketing related books to reach a level where I can discuss on equal terms with the people from the business units. Thanks to such comprehensive capabilities, mutual feedback between research, development, and business will go smoothly. And I will be able to firmly establish a path to solve issues.
Particularly in our research field, it is not the case that research alone can realize a solution to issues. There are some issues and ideas which become visible in the development process while others become visible through the broad penetration of technology and the collection of data. I would like to pursue research in anticipation of social implementation while thoroughly investigating my own research themes.

A day at work

Message to my past self in my school days