
You need Adobe Reader 7.0 or later in order to read PDF files on this site.
If Adobe Reader is not installed on your computer, click the button below and go to the download site.

|
August 2026 Vol. 24 No. 8 |
|
Front-line Researchers
-
Seishi Takamura, Visiting Senior Distinguished Researcher, Computer and Data Science Laboratories, NTT, Inc.

Abstract
Encoding technology, which effectively compresses the capacity of still images and videos, has developed into a cutting-edge research field, which is necessary in today’s world to transmit vast amounts of information efficiently in a multimodal manner. Dr. Seishi Takamura, a visiting senior distinguished researcher at NTT Computer and Data Science Laboratories, is a world-renowned researcher in this field and has proposed a series of new encoding methods based on various ideas. Recently, he has been challenging himself to explore the ultimate video imaging system that captures, transmits, stores, and displays images at the level of photons, the smallest units of light that make up images. In this interview, we asked him about his latest research results, vision for the future, and perspective as an IEEE (Institute of Electrical and Electronic Engineers) Fellow on raising Japan’s research capabilities and nurturing young researchers, including students.
Rising Researchers
-
Akihiro Suda, Distinguished Researcher, Software Innovation Center, NTT, Inc.

Abstract
The discovery of a backdoor in the open-source software (OSS) “liblzma” (a file-compression library) in 2024 demonstrated that even core operating-system libraries could be targeted in a manner that fundamentally damage the credibility of the OSS supply chain. Fortunately, this incident was discovered and addressed before it could be deployed to countless systems worldwide. In light of this serious situation, further strengthening of cybersecurity has become an urgent necessity. In this interview, we spoke with Distinguished Researcher Akihiro Suda at NTT Software Innovation Center [affiliation as of the time of the interview], a leading expert in strengthening the security of the OSS supply chain.
Feature Articles: AI for Quality Growth¡½NTT Group’s AI Research and Business
-
NTT’s LLM tsuzumi 2: A High-performance, Secure, and Low-cost Japanese LLM

Abstract
The NTT Group is promoting artificial intelligence (AI) business with “AI for Quality Growth,” a unique strategy to support sustainable and high-quality growth of enterprises with AI. The Japanese large language model (LLM) “tsuzumi 2,” lightweight yet high performing, was announced through a press release on October 20, 2025, and has started to be commercially available. This article outlines the NTT Group’s AI strategy and introduces the current status of NTT’s LLM “tsuzumi,” focusing on the evolution points from tsuzumi to tsuzumi 2.
-
tsuzumi: Challenges in Developing a Sovereign Large Language Model

Abstract
While large language models (LLMs) have attracted increasing attention, they also face challenges such as the need for enormous computational resources and security risks. To address these challenges, NTT has been conducting research and development on “tsuzumi,” a lightweight, high-performance Japanese LLM. By constructing a tokenizer optimized for the Japanese language, conducting high-quality pretraining, and applying alignment learning tailored to business needs, tsuzumi achieves high performance with excellent cost-effectiveness. NTT is also promoting practical functional enhancements, including an evaluation process for rapid business decision-making, improved software development capabilities, and extension to a large vision-language model capable of image understanding. This article introduces the technical initiatives of the tsuzumi project and its outlook.
-
tsuzumi Orchestra Search: Connecting People, AI, and Systems through Search

Abstract
Enterprise search has been moving toward the integrated use of multiple information sources, such as internal documents, business systems, and the web. In practice, however, a considerable amount of information still needs to be confirmed by consulting people. This article introduces the concept of tsuzumi Orchestra Search, which is based on NTT’s large language model “tsuzumi” and uses an orchestrator to coordinate multiple specialized agents. By combining internal and external information sources with inquiries to employees, this system aims to support the collection and organization of information necessary for business decisions and proposals while supplementing missing information.
Global Standardization Activities
Practical Field Information about Telecommunication Technologies
External Awards
|