Why deep learning for cybersecurity?

Deep learning is a subset of machine learning that uses multi-layered artificial neural networks to learn data representation. Deep learning not only builds more effective models from threat data, it also learns from data itself — no features are defined by humans. By training our models with the massive threat data that exists today, our platform learns to identify threats in real-time with very high efficacy. 

Watch the lightboard video by Arri Ciptadi, our Principal Machine Learning Scientist, on the differences between AI, machine learning and deep learning.

 

How Our Platform Works

The Blue Hexagon platform includes two components — (1) Our cloud where our models are trained (2) Our appliance that sits on-premises.

In our cloud, we curate a diverse set of threat data to train our deep learning models. Optimized models are deployed on our hardware or virtual appliance.  Our on-premises appliances in network tap or span mode, inspect North-South traffic for threats. You’ll not only stop threats in real time, but also gain real-time threat categorization and complete kill chain visibility into threats found. Orchestrate prevention to endpoints and network devices (firewalls, web proxies, routers and more). 

 

Benefits

Speed of Detection

Speed matters. Our threat detection occurs in less than a second so you can stop patient zero and lateral movement in the network. You won't have to deal with human triage, remediation, or SOC analyst investigation.

Prevention is Possible

We know detection isn't good enough. That's why we have a myriad of prevention options for you. Enable millisecond prevention inline or automate prevention policies to your endpoint and firewall.

High-Efficacy Verdicts

Our deep learning inspection applies to the complete network flow. This complete picture of an attack delivers very high efficacy threat verdicts. It also makes it harder for an adversary to evade us.

Complete kill chain visibility

When we detect a threat, we also deliver real-time threat categorization, including threat family and IoC. You'll also gain comprehensive kill chain visibility into infected systems and hosts, and attack timeline.

Works Out-of-the-Box

Our platform works out of the box on day one. You won't have to baseline what's on your network or fine-tune any of our models to optimize performance. Install in network tap mode and it'll just work.

Ease of Deployment

We are priced and managed as a modern SaaS with a usage-based pricing model. You can deploy our hardware or VM on-premises. Any hardware deployment is right-sized for you but priced based on actual traffic inspected.

Whitepaper: Outsmarting Attackers with Deep Learning

Cybersecurity continues to be an asymmetric war. Attackers are taking advantage of AI to create and distribute threats at unprecedented rates. In this whitepaper, read about how deep learning has evolved, why advances are possible today, and how it can address the new threat landscape of automated attacks.