Ministry of Economy, Trade and Industry website that delivers “now” of economy and industry


Living and learning in the Reiwa Era

AI startup co-founder, expert discuss 20 years of MITOU Program

Graduates boost profile

Takeuchi: I had the good fortune to mentor Okanohara when he was in his second year in college. How do you and he divide roles in your company?

Nishikawa: We used to think of his role as doing research and mine as working out business strategies, but actually, we ended up not having such a clear division between our roles.

In the first place, I don’t particularly enjoy designing business plans (laughs). To me, it’s not much fun to see things just go along a planned path. I rely on Okanohara for many business-related aspects as he is good with figures.

Takeuchi: I understand you were developing a search engine before launching PFN, but you reached a turning point while investigating deep learning (which is a core technology in the field of artificial intelligence). Whose idea was it to try deep learning, yours or his?

Nishikawa: Okanohara first focussed on neural networks, but I think maybe I initially suggested shifting our focus from traditional machine learning technology, which we were working on at the time.

Takeuchi: I understand you have worked to develop frameworks for the real-world application of deep-learning technology and build a large computer cluster to support research and development. Was it your suggestion to develop a dedicated processor to accelerate deep learning to speed up these efforts?

Nishikawa: Correct. Developing a computer has always been one of my ambitions.

Takeuchi: MITOU started in 2000 as a “millennium” program with an aim to identify and develop genius programmers.

Organized by Information-technology Promotion Agency, Japan, the MITOU Demo Day event presents various projects led by young, talented personnel.

Similar programs existed when it first began, but MITOU is a rare example in that it has lasted as long as two decades. They say education needs a long-term plan; I insisted that MITOU should not be a short-term project.

Nishikawa: I qualified for the program in fiscal 2005. That was its fifth year.

Takeuchi: Even in the business community, MITOU remained fairly unknown for the first five years or so of its existence. People who were groomed in the program, you included, started to make their presence felt after around its 10th year and raised the program’s profile. People came to appreciate MITOU’s work.


Introducing noteworthy topics and relevant efforts concerning economic and industrial circles