An Efficient Quantum-inspired Computing Approach for Intrusion Detection System
Jyun–Yi Shen, Ching-Hsuan Wu, Cheng-Yen Hua, Ming-Ho Chang, Shu–Yu Kuo, Yao–Hsin Chou, Sy‐Yen Kuo · 2024
Quantum-inspired computing (QIC) algorithms sim-ulate the principles of quantum mechanics to enable an efficient search for optimal solutions and substantial reductions in computational costs. This study marks the first attempt to propose a QIC algorithm designed for constructing an intrusion detection system (IDS). With the rapid advancement of technology comes a surge in malicious activities. The IDS is an important defense against internet attacks, which is crucial for efficiently identifying anomalous network traffic and preventing detrimental consequences such as data breaches. The proposed system introduces a new encoding method for IDS on QIC, complemented by a novel technique: a voting mechanism. This mechanism enhances accuracy and reliability in the decision-making process by aggregating multiple inputs or opinions, thereby reducing the rate of false negatives and false positives. The results demonstrate that the QIC-based IDS effectively distinguishes between normal network traffic and malicious ac-tivities, achieving high accuracy on two well-known benchmarks with lower floating-point operations (FLOPs). This lightweight design and high efficiency offer significant advantages for real-time systems and various industries.