Advancing Cybersecurity with AI: A Multimodal Fusion Approach for Intrusion Detection Systems

This paper presents a novel AI-enabled Intrusion Detection System (IDS) that enhances cybersecurity by integrating multimodal data analysis and AI fusion techniques. Through analysis of network traffic data from four test environments, it illustrates the benefits of a more robust detection strategy utilizing multiple modalities of the network traffic. The modalities extracted are designed to be protocol-agnostic, enabling their application across various network protocols, thereby broadening the system's applicability and effectiveness. Fusion of these models' outputs results in a robust, adaptable solution capable of real-time threat detection with improved accuracy across different network protocols.