Open-Set Recognition in Unknown DDOS Attacks Detection with Reciprocal Points Learning

Abstract

The internet plays a crucial role in modern society, yet its extensive use introduces serious challenges in privacy and cybersecurity. Among these, Distributed Denial-ofService (DDoS) attacks are particularly disruptive, often crippling the network operations of targeted organizations. Although security mechanisms such as firewalls and intrusion detection systems (IDS) are employed, detecting DDoS attacks especially previously unseen or unknown types remains a significant challenge. Recent advancements in machine learning (ML) and deep learning (DL) have improved IDS capabilities, but limitations persist for unknown attack detection. This study proposes NN-RPL, a hybrid model combining Neural Networks (NN) with Reciprocal Points Learning (RPL), an advanced Open-Set Recognition (OSR) technique. The model effectively distinguishes both known and unknown attack patterns while maintaining a lightweight architecture with reduced training complexity. Finally, NN-RPL offers an efficient, flexible, and practical security solution suited to organizations facing. In particular, the NN-RPL model simplifies the architecture of the deep neural network by significantly reducing the number of training parameters without compromising defense capabilities. Therefore, our proposed method is genuinely efficient, particularly flexible, and lightweight compared to traditional methods.