New ML framework PHASE evaluates human behavior realism in cybersecurity simulations

Researchers have developed PHASE, a machine learning framework designed to evaluate the realism of human behavior in cybersecurity simulation environments. This passive system analyzes Zeek connection logs to distinguish human activity from non-human activity with over 90% accuracy. By utilizing local DNS records for labeling and SHAP analysis to identify behavioral signatures, PHASE can uncover patterns that undermine the realism of synthetic user personas, enabling improvements to create more effective simulations. AI

IMPACT Enhances the realism of cybersecurity training and testing environments by improving synthetic user behavior.

RANK_REASON The cluster contains a research paper detailing a new machine learning framework. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• SHAP
• Steven Lampier
• Zeek

• paper
• product
• other

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