The Enterprise Immune System: Proven Mathematics and Machine Learning for Cyber Defence

Thursday, 16 February, 2017 In


By leveraging the power of advanced machine learning and mathematics, Darktrace takes a fundamentally new approach to the challenge of defending against sophisticated cyber-attacks.

This approach, known as the Enterprise Immune System, represents a new category of cyber defense technology. It deploys cutting-edge technology modeled after one of the most powerful biological systems — the human immune system. By understanding what ‘normal’ looks like, our immune systems can distinguish ‘us’ from ‘not us’ to quickly identify potential threats. This self-learning capability gives it the means to adapt and evolve in an intelligent manner, defending against new threats and changing environments.

The self-learning mechanisms of the human immune system inspired Darktrace’s flagship technology: the Enterprise Immune System. Powered by advances in unsupervised machine learning developed by specialists from the University of Cambridge, the Enterprise Immune System intelligently detects emerging threats that other security tools miss. The system can also automatically fight back with ‘digital antibodies’ — targeted responses that complete the end-to-end functionality of the Enterprise Immune System.

To devise this new approach, mathematicians from the University of Cambridge developed a breakthrough in Bayesian theory. This new branch of probabilistic mathematics, called Recursive Bayesian Estimation, helps to make sense of huge data sets, deal with uncertainty, and, ultimately, identify cyber-attacks amid the noise of a network.

The full white paper can be read below: