Jeremiah O'Connor
summary
Research team builds a pipeline that crawls clear and dark web sources, applies unsupervised anomaly detection to find cybercrime patterns at internet scale, and reports that machines must replace human labeling as data grows astronomically. The stack uses Apache Spark for scale and gensim and NLTK for text processing, with practical lessons on linear algebra and algorithm choices. The system runs in production and surfaces meaningful anomalies for further investigation.
AI-written summary · about
related talks
48mHunting Stealth Adversaries with Graphs & AI
52mDevil in the Haystack
27m"Cybertelligence": How AI is changing the face of cybersecurity
34mBeyond Manual Hunting: OSINT Copilot with LLMs & Evidence-Based Collection, Enrichment ...
49mA Real-World Example of Using Network Behaviour to Detect Cyber-Compromise
44mApplied Machine Learning for Data Exfil and Other Fun Topics
47m Natural Language Processing & Anomaly detection in Sys call logs
22mDavid Talby, SVP Engineering, Atigeo