Machine learning models detect LDAP reconnaissance with high accuracy

Researchers have developed two machine learning frameworks to detect malicious Lightweight Directory Access Protocol (LDAP) reconnaissance activities. The first framework uses weak supervision to label a large dataset and classify LDAP queries as malicious or benign, achieving up to a 65% True Positive Rate. The second framework employs statistical hypothesis testing to extract novel malicious LDAP signatures, demonstrating 81.48% field precision. This approach leverages automated corpus construction to reduce costs and time compared to manual labeling. AI

IMPACT Enhances security protocols by enabling early detection of malicious reconnaissance activities.

RANK_REASON Academic paper detailing novel ML methods for security. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• Active Directory
• CrowdStrike
• Lightweight Directory Access Protocol
• machine learning
• Managed Detection and Response team
• ML-based data-mining method
• ML classifier
• statistical hypothesis-testing framework

• paper
• safety

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