Adaptive Malware Defense: A Multi-Agent System for Malware Detection and Adaptive Response
Summary
No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.
Summarise with VIRAT Free to try with your Science Catalogs account.Abstract
Malware remains a persistent threat to governments, corporations, and individuals, with conventional signature-based detection methods proving inadequate against polymorphic, fileless, and adversarially designed threats. This paper presents Adaptive Malware Defense (AMD), a multi-agent system integrating static analysis, dynamic sandboxing, threat intelligence enrichment, explainable machine learning, and Large Language Model (LLM)-based reasoning for comprehensive malware detection. The system achieves 94.5% accuracy using Logistic Regression with SHAP(SHapley Additive exPlanations)based interpretability, with advanced models reaching 97.2% accuracy. External threat intelligence from VirusTotal, AbuseIPDB, and the National Vulnerability Database enriches detection context. The complete analysis pipeline executes in an average of 9.07 seconds, demonstrating practical feasibility for Security Operations Center workflows. Adversarial robustness testing reveals resilience under obfuscation, with transformer-based models maintaining 92.5% accuracy. A Streamlit-based prototype validates operational utility through interactive dashboards, automated reporting, and analyst-friendly interfaces. By unifying detection mechanisms with explainable AI and LLM-based contextual reasoning, AMD bridges the gap between detection and actionable defense.
Fields of Study & Subjects
Advanced Malware Detection Techniques
Spam and Phishing Detection
Network Security and Intrusion Detection
Publication details
- Journal / source
- International Conference on Computing, Communication, Control and Cyber-Physical Systems (I5CPS)
- Pages
- 1–6
- Publication date
- 26 February 2026
- Language
- EN