Tiberiu Boros is a Ph.D. in computer science, specialized in Machine Learning for Spoken Language Processing. Currently, his research activity focuses on applied Deep Learning and Statistical Models to mainstream tasks in both Natural Language Processing and Security. He published more than 50 papers in conferences that include A and A* publications (ACL, EACL, CONLL, LREC) and a book chapter in a Springer publication (“Where humans meet machines”, 2013). He maintains several security-related open-source projects (Stringlifier, OSAS, Tripod and libLOL) and one Natural Language Processing (NLP) pipeline (NLPCube). In recognition for his work he received a Romanian Academy Research Award and a Microsoft Research Award. His most recent work in the field of security includes several conference talks, a published research paper entitled “A Principled Approach to Enriching Security-related Data for Running Processes through Statistics and Natural Language Processing”, and a currently under-review paper “Machine Learning and Feature Engineering for detecting Living off the Land Attacks”.

Hunting for LoLs (a ML Living of the Land Classifier)

Living of the Land is not a brand-new concept. The knowledge and resources have been out there for several years now. And still, LoL is one of the preferred approaches when we are speaking about highly skilled attackers or security professionals. Two main reasons why:

1. Experts tend not to reinvent the wheel
2. Attackers like to keep a low profile/footprint (no random binaries/scripts on the disk)

The talk focuses on detecting attacker activity/Living of the Land commands using Machine Learning, for both Linux and Windows systems.

The presentation covers why it is hard to detect LoLs, the feature engineering used in our approach, comparison between different classifiers as well as hands-on experience using our library and integration into one of our previous open-source projects called One-Stop-Anomaly Shop (OSAS – https://github.com/adobe/OSAS). Additionally, we also discuss why OSAS and the LoL classifier are complementary solutions and how evading one will lead to being detected by the other.

This presentation is co-presented with Andrei Cotaie, Technical Lead, Security Intelligence & Engineering at Adobe.