KnowML: Improving Generalization of ML-NIDS with Attack Knowledge Graphs

TL;DR

KnowML enhances ML-based Network Intrusion Detection Systems by integrating attack knowledge graphs derived from large language models, significantly improving detection accuracy across diverse attack variants.

Contribution

The paper introduces KnowML, a novel framework that incorporates attack knowledge graphs into ML-NIDS, enabling better generalization and detection of diverse attack variants.

Findings

Baseline ML-NIDS fail to detect some attack variants (F1=0%).

KnowML achieves up to 99% F1 score on attack variants.

False positive rate remains below 0.1% with KnowML.

Abstract

Despite extensive research on Machine Learning-based Network Intrusion Detection Systems (ML-NIDS), their capability to detect diverse attack variants remains uncertain. Prior studies have largely relied on homogeneous datasets, which artificially inflate performance scores and offer a false sense of security. Designing systems that can effectively detect a wide range of attack variants remains a significant challenge. The progress of ML-NIDS continues to depend heavily on human expertise, which can embed subjective judgments of system designers into the model, potentially hindering its ability to generalize across diverse attack types. To address this gap, we propose KnowML, a framework for knowledge-guided machine learning that integrates attack knowledge into ML-NIDS. KnowML systematically explores the threat landscape by leveraging Large Language Models (LLMs) to perform automated analysis of attack implementations. It constructs a unified Knowledge Graph (KG) of attack strategies, on which it applies symbolic reasoning to generate KG-Augmented Input, embedding domain knowledge directly into the design process of ML-NIDS. We evaluate KnowML on 28 realistic attack variants, of which 10 are newly collected for this study. Our findings reveal that baseline ML-NIDS models fail to detect several variants entirely, achieving F1 scores as low as 0 %. In contrast, our knowledge-guided approach achieves up to 99 % F1 score while maintaining a False Positive Rate below 0.1 %.