Dmitrijs Trizna

Speaker

Designation :

Security Data Science at Avast


Talk Title :

Shell Language Processing: parsing Unix commands for Machine Learning


Abstract :

In this workshop, we present a Shell Language Preprocessing (SLP) concept, which is focused on adjusting Unix commandline telemetry for Machine Learning pipeline. We describe the rationale behind the need for a new approach with specific examples when conventional Natural Language Processing (NLP) pipelines fail. Furthermore, we we will share a Notebook with dataset that allows on a hands-on manner to evaluate our methodology on a security classification task against widely accepted information and communications technology (ICT) tokenization techniques and achieve significant improvement of an F1-score from 0.392 to 0.874

Bio :

Dmitrijs is a Security Data Scientist at Avast. In past: Red Teaming, Threat Hunting, membership in NATO cybersecurity events. Publications and speaker at DefCon and CAMLIS. Two higher educations (MSc Data Science, MSc Network Security). Certified OSCP, CCNA, CCSA, deeplearning.ai.

Want to connect?