Elemendar is Three! 

It feels a little odd celebrating during this very challenging time, but we also have to accept that life continues and that change is the only constant. We realised a week ago that we are three years old and it’s a good time to reflect on how far we’ve come and where we are going. Read what Giorgos thinks about this: https://elemendar.com/may-the-3rd-be-with-you/

Our new release v0.0.8 is out!

  • includes Identity, Tool and Vulnerability objects, meaning we now automatically extract 8 (up from 7) of the 11 relevant STIX 2.0 domain objects
  • integrates multiple new models to improve accuracy, including a voting system between those models to reduce false positives (for those interested in the tech, we’ll be writing more about this approach soon)
  • extracts relationships more than twice as thoroughly, from 7 types of relationships to 19, including linking attack patterns directly to their indicators.

Coming next in v0.0.9, a smaller release will integrate allow-lists to reduce false positives in indicator extraction.

Nadia has written about cybersecurity and the nuclear power industry

It turns out we do need to worry about it. Read more: https://elemendar.com/cyberproofing-nuclear-power-plants-in-the-era-of-a-pandemic/

Rita has written about the challenges of working with real world datasets

For the ML and data geeks among you, our NLP Engineer Rita is working with imbalanced data sets in the real (cyber) world. Read her strategies to handle them using 1. Data driven, 2. Algorithmic, and 3. Synthetic data driven methods: https://elemendar.com/working-with-real-world-datasets/

Things We Have Found Interesting

This fascinating and super in-depth resource by Privacy Affairs on Cyberwarfare statistics: A decade of geopolitical attacks. It breaks down:

  • the nature of cyber attacks
  • where the attacks originate from
  • who they target

Read it here: https://www.privacyaffairs.com/geopolitical-attacks/

That’s all from us this month. 

Please stay safe

Tristan Palmer

Head of Growth

Our AI for Cyber Threat Intelligence engine reads and translates human authored CTI feeds into machine readable and actionable data (STIX2 inc. MITRE ATT&CK™). This saves significant time and money and better protects your whole organisation from threats. Try our free trial here.