Cross Domain Attack Detection Model Based on Transfer Learning
Shangwen Ouyang, Yong Wang, Lingmei Gao · 2025
With the advancement of network technology, security threats have intensified, and cross domain attacks have brought significant challenges. Traditional detection methods based on fixed rules are difficult to cope with constantly evolving attack strategies, resulting in poor protection effectiveness. To this end, this article introduces the Cross Domain Attack Detection Model Based On Transfer Learning(TLCAD), as shown in Fig. 1, which utilizes source domain knowledge to quickly adapt to the target domain, reducing training time and data requirements. TLCAD simplifies model selection and hyperparameter tuning by extracting domain invariant features to optimize classifiers. Experimental results show that it performs well in simplicity, accuracy, and efficiency, significantly improving cross domain attack detection capabilities and outperforming traditional and deep learning methods.