Guarding Organizations Against Malware Risk: A Novel Graph-Based Malware Detection Method

2026-07-29

AI Summary

This work addresses the limitations of existing bytecode-based malware detection methods, which are vulnerable to evasion attacks, and conventional graph-based approaches that struggle to effectively capture semantically meaningful groups of basic blocks, thereby constraining detection accuracy. To overcome these challenges, the paper proposes MalGuard, a novel framework that introduces an operational role identification mechanism to cluster basic blocks into semantically coherent roles. It then models their dependencies through a role-interaction-oriented hierarchical graph representation learning approach. By integrating program dependence graphs with graph neural networks, MalGuard preserves structural hierarchy while precisely capturing sparse malicious signals and semantic program behaviors. Experimental results demonstrate that MalGuard significantly improves malware detection accuracy in enterprise environments and effectively reduces the expected cost associated with undetected threats.

Abstract

Organizational digitalization expands cybersecurity risks, making cybersecurity an increasingly important research area in Information Systems (IS). Among these risks, malware has become a pervasive and destructive threat. Byte-based machine learning (ML) methods are widely used for malware detection but remain vulnerable to evasive behaviors that manipulate raw bytes to evade detection. Graph-based methods are less affected by such manipulations because they represent software as program graphs that capture execution behavior. However, they do not explicitly identify cohesive groups of basic blocks that jointly realize meaningful program behaviors, nor do they learn sufficiently expressive program graph representations for accurate detection. To this end, we propose MalGuard, a graph-based malware detection method for organizational malware risk management. MalGuard introduces two methodological innovations: an operational role identification approach and a program graph representation learning method. The former identifies these cohesive groups of basic blocks as operational roles, enabling the detector to capture program behaviors that may not be visible from isolated basic blocks. The latter learns expressive program graph representations by modeling interactions among operational roles, preserving sparse malicious signals, and capturing hierarchical graph structure. Extensive experiments show that MalGuard improves detection performance and reduces the expected cost of undetected malware.

Problem

• malware detection
• program graph
• basic blocks
• graph representation learning
• operational roles

Innovation

• graph-based malware detection
• operational role identification
• program graph representation learning
• basic block cohesion
• malware risk management