AutoLabel: Automated Fine-Grained Log Labeling for Cyber Attack Dataset
summary
AutoLabel turns log labeling into extracting an attack subgraph from a provenance graph, using instrumentation to emit auxiliary alignment info, inject attack flags as unambiguous anchors, and apply unit partitioning to avoid dependency explosion. The system fuses audit, application, and traffic logs to produce fine-grained labels, achieves 100% labeling accuracy across 29 scenarios (25 real CVEs plus a Sandworm CTID), and generates datasets in ≤96 minutes per scenario, producing 580 benchmarks.
AI-written summary · about
related talks
13mAPTGen: An Approach towards Generating Practical Dataset Labelled with Targeted...
30mData Driven Security
22mMachine Learning Based Attack Vector Modeling for CyberSecurity
26mNLP for security log analysis : Learning to crawl before you run
48mHunting Stealth Adversaries with Graphs & AI
47m Natural Language Processing & Anomaly detection in Sys call logs
40mMe, My Adversary & AI: Investigating & Hunting with Machine Learning
31mEnhancing Secret Detection in Cybersecurity...