国立大学法人東北大学電気通信研究所公式ウェブサイト

Industry-Academia-Government-Collaboration Research and Development Division AI Hardware Security Project

>> Research Center for 21st Century Information Technology(IT-21 Center)

Industry-Academia-Government-Collaboration Research and Development Division AI Hardware Security Project

Industry-Academia-Government-Collaboration Research and Development Division AI Hardware Security Project

Staff

[ Project Leader, Professor* ] Naofumi Homma

Research Activities

In the coming super-smart society, various physical and biological data will be collected and integrated in applications such as automobiles, industrial/household robots, and customized medical care. Accordingly, individualized and comprehensive services that utilize the knowledge gained from them are expected.
On the other hand, security is considered to be an inseparable issue in utilizing and expanding such services in the future. In particular, cryptography is the foundational technology of modern security, and strong security technology based on cryptography will be required as its core in the future society where artificial intelligence (AI) will be widespread.

IT-21 Center has been promoting research and development of next-generation information communication technology (ICT) through industry-academia collaboration projects. From FY2025, we have launched a new project to promote research and development of AI-related security (AI security) with the aim of establishing highly secure ICT in a society where AI is widespread.

The main purpose of this project is the research and development of cryptographic hardware that is highly resistant and robust against various attacks. When assuming implementation on IoT terminals and edges, in addition to cyber attacks, physical attacks (attacks that physically access the system and steal or tamper with information) will be a major risk. Among these, the most realistic threat is the side channel attack, which exploits unintentional physical quantities (power consumption, electromagnetic radiation, execution time, etc.) during system operations to steal secret data and control in a non-destructive and non-invasive manner. Even for next-generation cryptography, it is essential to implement it with consideration for resistance to physical attacks including side channel attacks. In light of the above, this project promotes the research and development of advanced AI-based side channel attacks and their countermeasures.

Specifically, we are conducting the research mainly in the industry-academia collaboration project “Development of AI hardware security technology” in the framework of the JST K Program “Establishment of AI security technology necessary for a data-driven economic society where artificial intelligence (AI) is widespread,” which we were commissioned to do from FY2024. Through this project, we also aim to contribute to the social implementation of AI security technology and the promotion of industry-academia collaboration.

Fig.1 Overview of research topics.
Fig.2 Side-channel attacks exploiting machine learning.