AI forensic pipeline ensures model integrity for intrusion detection

Researchers have developed MODEXRE, a forensic pipeline designed to ensure the integrity and verifiability of explainable AI models used in intrusion detection. The system extends a previously validated XGBoost classifier with SHAP explainability, operating in two modes: Laboratory for training and certification, and Formal for inference on case evidence. MODEXRE incorporates a SHA-256 hash-chained chain of custody to guarantee that the certified model is the exact artifact applied to a case, addressing issues like timestamp leaks and feature-coverage gaps that previously affected model performance. The pipeline has been refined through four key design decisions, leading to significant improvements in classification accuracy and the ability to identify specific threat behaviors. AI

IMPACT Enhances trust and reliability in AI systems used for critical security and forensic applications.

RANK_REASON Academic paper detailing a new system and methodology for AI model verification in a specific domain. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• CICIDS2017 dataset: performance improvements and validation as a robust intrusion detection system testbed
• Common European Framework of Reference for Languages
• Kitsune
• MODEXRE
• Open Cybersecurity Schema Framework
• SHA-256
• SHAP
• Suricata
• XGBoost

• paper
• safety
• product

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →