What happens when cyber security and machine learning work together? The results are pretty positive. Many technologies are leveraging machine learning in cyber security functions nowadays in order to automate and augment their cyber workforce. How? Most recently in training and skill building.
Machine learning helps emulate human cognition (e.g. learning based on experiences and patterns rather than inference) so autonomous agents in a cyber security system for instance, can “teach themselves” how to build models for pattern recognition—while engaging with real human cyber professionals.
Machine learning as a training support system
Machine learning becomes particularly valuable in cyber security training for professionals when it can support human activities like malware detection, incident response, network analysis, and more. One way machine learning shows up is in our gamified cyber learning platform Project Ares , under our AI-advisor “Athena” who generates responses to player’s queries when they get stuck on an activity and/or need hints to progress through a problem.
Athena generates a response from its learning corpus, using machine learning to aggregate and correlate all player conversations it has, while integrating knowledge about each player in the platform to recommend the most efficient path to solving a problem. It’s like modeling the “two heads are better than one” saying, but with a lot more “heads” at play.
Machine learning as an autonomous adversary
Likewise, machine learning models provide a general mechanism for organization-tailored obscuring of malicious intent during professional training—enabling adversaries to disguise their network traffic or on-system behavior to look more typical to evade detection. Machine learning’s ability to continually model and adapt enables the technology to persist undetected for longer (if it is acting as an autonomous agent against a trainee in our platform). This act challenges the trainee in the platform in a good way, so they begin to think like an adversary and understand their response to defensive behavior.
Machine learning supports cyber skills building
Companies like Uber use machine learning to understand the various routes a driver takes to transport people from point A to point B. It uses data collected to recommend the most efficient route to its destination.
It increases the learning potential for professionals looking to hone their cyber skills and competencies using machine learning.
Now imagine that concept applied to cyber training in a way that can both help cyber pros through cyber activities while also activating a trainee’s cognitive functions in ways we previously could not with traditional, off-site courses.
Machine learning abilities can analyze user behavior for both fraud detection and malicious network activity. It can aggregate and enrich data from multiple sources, act as virtual assistants with specialized knowledge, and augment cyber operators’ daily tasks. It’s powerful stuff!