Explainable AI in Distributed Denial of Service Detection
Raj Kumar Batchu, Seshu Bhavani Mallampati, Hari Seetha · 2024
The rapid expansion of networks and associated innovations results in the production of a huge volume of data from a variety of services, which in turn creates potential vulnerabilities for networks. The exponential growth in cyber-attacks that hinder the usage of online resources has led to an imminent need to detect them. Although several intrusion detection systems have been introduced, still several IT organizations face substantial financial loss due to these cyber threats. Obtaining optimal dimensionality is a crucial task in modeling an effective detection mechanism. The emergence of AI has brought enormous models for detecting a variety of cyber-attacks. But some of these models are black box models with complex underlying algorithms having less transparency and hence trusting them was questionable. In this regard, the Explainable AI-based methods would be more powerful in not only providing trustworthiness but also in determining the contribution of the features toward prediction. This chapter presents various techniques and explainable AI methods used in obtaining specific features that could distinguish cyber-attacks.