Study on Japan’s Cybersecurity Strategy and Regional Cybersecurity Construction in the Face of Global Competition

Abstract

Cybersecurity has emerged as a crucial security concern for countries worldwide, and Japan has made significant strides in this area, providing other countries with valuable lessons from its experience. In the context of global competition, this paper presents Japan’s cybersecurity situation, examines the top-level design of the Japanese government’s cybersecurity strategy, and outlines a three-pronged approach to building regional cybersecurity. Aiming to address the shortcomings of the existing DAEDALUS cyber attack warning system, this paper proposes an unknown attack detection algorithm based on an open-set classification network, which improves the algorithm’s recognition accuracy in the face of unknown attacks through loss function optimization and classifier replacement on the basis of a convolutional neural network’s open-set classification architecture. The experimental results demonstrate that, under various openness conditions, this algorithm optimizes the average F1 value of the three recognition models to 78.47 and enhances the detection accuracy of OCN in unknown attacks across two datasets to 84.94% and 76.53%, respectively. This paper demonstrates the accuracy of the OCN algorithm in recognizing unknown attacks.

Keywords

• japanese network security
• artificial intelligence
• unknown attack detection
• open set classification
• 97p20