Machine Learning and Artificial Intelligence on Big Data for Cybersecurity

Author:

Fred Cohen, Principal Consulting Analyst

Abstract

The use of machine learning and artificial intelligence for cyber-security are nothing new. But the availability of larger data sets and the evolution of techniques applicable to big data have produced a new generation of systems that improve efficiency and utility. The need for far greater scale brought about by the dramatic increase in the number of users, uses, and systems involved, has driven the development of machine learning and artificial intelligence for cyber-security.

The basic technological changes involve mathematical algorithms that examine large data sets containing known bad and known good samples. These methods create equations that cluster known “good” and “bad” samples and differentiate them from each other, then apply those same equations to new samples to classify them as “good” or “bad”. This can then be applied to any of a wide range of problems, including many of those of cyber-security.

The enterprise benefits of these emerging technologies are economies of scale, efficiency of labor, and detection in areas not previously addressed.

This report discusses; (1) the basics of these techniques, (2) the words used to describe them and what those words mean, (3) the limitations, benefits, and costs of these techniques, and (4) their application to CySec today and into the future. It then discusses product types emerging in the markets and the current and likely future utility of applying these product types to enterprises.

In this report, we investigate this emerging trend, and what should be the next steps for TechVision Research clients.