Title: SD-CGAN: Conditional Sinkhorn Divergence GAN for DDoS Anomaly Detection in IoT Networks Show Chinese title

Title: SD-CGAN：用于物联网网络中DDoS异常检测的条件Sinkhorn散度GAN

Abstract: The increasing complexity of IoT edge networks presents significant challenges for anomaly detection, particularly in identifying sophisticated Denial-of-Service (DoS) attacks and zero-day exploits under highly dynamic and imbalanced traffic conditions. This paper proposes SD-CGAN, a Conditional Generative Adversarial Network framework enhanced with Sinkhorn Divergence, tailored for robust anomaly detection in IoT edge environments. The framework incorporates CTGAN-based synthetic data augmentation to address class imbalance and leverages Sinkhorn Divergence as a geometry-aware loss function to improve training stability and reduce mode collapse. The model is evaluated on exploitative attack subsets from the CICDDoS2019 dataset and compared against baseline deep learning and GAN-based approaches. Results show that SD-CGAN achieves superior detection accuracy, precision, recall, and F1-score while maintaining computational efficiency suitable for deployment in edge-enabled IoT environments. Show Chinese abstract

Abstract: 物联网边缘网络复杂性的增加给异常检测带来了重大挑战，特别是在高度动态和不平衡的流量条件下识别复杂的拒绝服务（DoS）攻击和零日漏洞时。 本文提出了SD-CGAN，这是一种增强的条件生成对抗网络框架，结合了Sinkhorn散度，专门用于物联网边缘环境中的鲁棒异常检测。 该框架结合了基于CTGAN的合成数据增强以解决类别不平衡问题，并利用Sinkhorn散度作为几何感知损失函数，以提高训练稳定性和减少模式崩溃。 该模型在CICDDoS2019数据集的剥削性攻击子集上进行评估，并与基线深度学习和GAN方法进行比较。 结果表明，SD-CGAN在保持适合边缘启用物联网环境部署的计算效率的同时，实现了更高的检测准确率、精确率、召回率和F1分数。