This section introduces the contents of press conference and keynote lectures held at Upgrade 2024, the NTT Group's R&D event held by NTT Research from April 10 to 11.

This is a report by Waichi Sekiguchi, President of MM Research Institute, Ltd., who covered the event on-site.
< Part 2 of 2 / Go to Part 1>

Before the Upgrade 2024 keynote lectures, a press conference was held on April 10 by Kazuhiro Gomi, CEO of NTT Research, and Yosuke Aragane, head of the NTT IOWN Promotion Office. First, CEO Gomi gave an overview of NTT's R&D structure, NTT Research, and its efforts over the past five years since its establishment. He also announced a new donation to Harvard University.

"NTT has about 2300 researchers worldwide and spends about $3 billion annually on R&D," said Gomi. He also noted that NTT Research has three research centers, with 86 patents pending over the past five years, and that 450 academic research papers have been published in prestigious research journals and conferences, with seven of the research papers winning awards at prestigious conferences. He also stated that the company collaborates with 15 world-renowned academic institutions and organizations, including Harvard University and Stanford University to promote basic research.

NTT announced that they made a gift to Harvard University to establish the Harvard University Center for Brain Science-NTT Fellowship Program in the field of Physics of Intelligence. This will serve as a bridge for computer science research, brain science, psychology, and physics. The donation of up to $1.7 million will be used to support postdoctoral research.

Aragane, head of the NTT IOWN Promotion Office, explained the IOWN APN and presented the newly started demonstration tests of APNs in the U.K. and the U.S.

While demand for data centers is increasing worldwide due to the emergence of generative AI and other technologies, Aragane, head of the IOWN Promotion Office, pointed out that it is becoming increasingly difficult to establish data centers in urban areas due to environmental problems and land shortages. Geographically, suburban data centers experience latency, and he explained the benefits of leveraging APNs with higher capacity, higher quality, lower power consumption, and lower latency to solve this problem.

He introduced APN commercially available in Japan since 2023, as a use case for remote processing of real-time AI analysis, which would be useful for services that require real-time response, such as automated driving.

He announced that they succeeded in the demonstration experiment connecting long-distance data centers in the U.K. and the U.S. with APN. By connecting data centers near 100 km apart using APN, low-latency communication of 1 millisecond or less was achieved, confirming the potential use for distributed real-time AI analysis. Conventional networks experience a delay of over 2 milliseconds.

In the future, he said that IOWN APN will be developed globally with NTT DATA, and will promote a joint demonstration experiment to create use cases for the distributed data center constructions including the financial field.

https://group.ntt/en/newsrelease/2024/04/12/240412a.html

Following CEO Gomi's opening remarks, the first panelist to speak at the keynote lecture session was Joe Alexander (on the left), Director of the Medical & Health Informatics Laboratories—NTT Research's third research field. He emphasized that if the ecological information of the cardiovascular system could be collected by sensors and reproduced on a computer as a bio-digital twin of the human health condition, it would contribute greatly to the digitization of the medical field, such as medication and surgical preparation.

NTT Senior Executive Vice President Katsuhiko Kawazoe (in the center), who oversees R&D at NTT, joined Gomi on the panel. He stated that NTT is focused on developing the next-generation optical information and communications platform IOWN and the generative AI technology Tsuzumi. NTT is spending an annual R&D budget of 3 billion dollars to achieve these goals.

On the same day that the Upgrade event opened, Japanese Prime Minister Fumio Kishida and U.S. President Joe Biden held a Japan-US summit meeting on the East Coast of the U.S., and Senior Executive Vice President Kawazoe introduced that Japan-US cooperation on IOWN was included in the agreement between the two countries. Kawazoe looked ahead to the future of the digital society, saying that photonic chips (optical semiconductors) will replace conventional silicon chips (semiconductors) in the computer age.

The final speaker of the lecture session was Tim McKenna, Principal Scientist of the Physics & Informatics Laboratories who is researching coherent ising machine (CIM) to develop optical semiconductors and other technologies. He pointed out that the convergence of optical technology with quantum physics, will enable the development of semiconductors with far greater processing power than current silicon chips, and that they have discovered that thin-film lithium niobate (LN) is effective in creating circuit substrates for such devices. McKenna predicted the use of thin-film LNs is still in the beginning, but will grow to a $2 billion market by 2028.

I spoke with Tim McKenna, Principal Scientist, Physics & Informatics Laboratories at NTT Research, about his research on optical semiconductor technology.

McKenna: At NTT Research, we are working on the development of coherent optical computers. This is an analog computer useful for computing difficult optimization problems. As part of this research, we are developing thin film substrate technology for lithium niobate (LN). This is a substrate for fabricating optical semiconductor circuits and is a technology that will enable next-generation optical computers.

McKenna: That's right. It can be used as an optical accelerator in the high-speed network in the IOWN, and it is a very effective communication technology for connecting computers. In the future, computation and communication will be possible using light, creating a seamless all-optical network.

McKenna: That's a good question. Generative AI has increased the demand for all technologies. All ecosystems need to evolve to meet the needs of generative AI. The current hardware is not enough so we will need better hardware in the future. Electronic technology needs to get more efficient, and the optical network needs improvement. I believe the IOWN will play an important role.