Leveraging Kolmogorov-Arnold Networks for Enhanced IoT Security Threat Detection
R Yamuna, C N Pushpa, J Thriveni, K. R. Venugopal · 2025
The rapid growth of Internet of Things (IoT) systems has enhanced connectivity while introducing significant security challenges, necessitating precise threat detection mechanisms. This paper presents a novel approach utilizing Kolmogorov-Arnold Networks (KAN) for high-accuracy security threat detection in IoT environments. By leveraging KAN’s exceptional ability to approximate multivariate functions with precision, the proposed framework efficiently identifies anomalous patterns indicative of security threats. Evaluated on the CICIDS 2017 dataset, the KAN-based model achieves remarkable performance, with an accuracy of 98.7 % and are call of $98.9 \%$ significantly surpassing traditional deep learning models. Additionally, the approach demonstrates scalability and adaptability across diverse IoT applications, offering a robust and intelligent solution to bolster IoT security.