June 30, 2022

Creating a Social System in Which People Can Live in Peace

Becoming a "provider of platforms that enable every system to be safely protected and continue to run normally." That is the very research vision put forth by NEC's security area. In the current era, social systems have become inseparable from ICT. NEC's various technologies are being broadly utilized in social infrastructure areas which support people's daily lives. We are advancing research and development on a daily basis with the goal of thoroughly managing cyber attacks, data leaks, and other risks to ensure the reliability of social infrastructure.
NEC's strengths include know-how accumulated through many years of involvement in social infrastructure and specialized knowledge in each area. The ability to create practical security systems that can be utilized in the field can perhaps be described as a unique characteristic of NEC. In addition, the company has had many research papers accepted by international conferences as it continues to send advanced technologies out into the world.

Utilizing Digital Twins and World-leading Encryption Technologies

NEC's research in the security area is currently pursuing the three themes of "system security," "data security," and "AI security." In system security, our goal is to build and operate secure systems utilizing digital twins. In data security, we are researching technologies that can utilize data that stays encrypted end to end while in AI security, we are advancing technology research to prevent attacks against AI.

Primary research technologies

Technology for automatic cyber attack risk assessment

Automated secure system design technology

This technology automatically designs security-conscious systems and networks based on customer requirements. Converting the secure system design knowledge possessed by NEC into a database enables design without any dependence on individuals or examination omissions.

Lightweight authenticated encryption

Secure computation (Multi-party computation)

Secure biometric authentication (homomorphic encryption)

This technology recognizes biological features used in face recognition in an encrypted state to reduce the risk of leaking such personal information. In addition to the secret sharing method mentioned above, it also realizes a method using homomorphic encryption which does not require multiple servers.

Highly-confidential federated learning

This technology is compatible with both high security and analysis accuracy in federated learning which utilizes data between multiple organizations. By applying secure computation and attack simulation, it realizes attack-resistant learning with little accuracy degradation to eliminate the risk that the original data will be inferred from the AI model that is the learning result.