What is deep learning and how does it impact cyber security?

Deep learning is changing the landscape of cyber security with more than 90% accuracy in stopping malware and ransomware attacks.

All public sector organisations are vulnerable to being targeted with ransomware and malware attacks. These cyber threats are evolving at alarming rates and it’s time that ransomware prevention and cyber security adapted too.

The public sector today

The public sector is facing an increase of around 800% in ransomware attacks in the last 24 months and public sector security teams are often taking too long to respond to security incidents leaving organisations exposed.

According to a recent global Deep Instinct study from October 2021, the average time for a response was 24.4 hours – almost three full working days. The public sector is already a high value target for attackers, but response times are making them even more vulnerable.

The traditional solution

Even with the sophisticated cyber security solutions available on today’s market, high-profile breaches still happen, and public sector organisations of any size can become a target without the right protection in place.

Often organisations will assume that they are fully protected using traditional solutions, such as machine learning. However, machine learning can leave gaps in your cyber defences due to its process, which relies on human intervention and means that human training must also stay up to date.

Security gaps in machine learning:

  • Less than 2% of available data
  • Human reliant
  • Linear model
  • Limited file types covered (PE)
  • Leads to 1-2% false positives
  • Accuracy of unknown threats 50-70%

Introducing deep learning

Deep learning is not a new technology and it is already being applied by organisations, such as Netflix and YouTube. The platforms utilise it to make recommendations on films and TV based on your viewing history.

Organisations like Tesla are also employing the technology to aid with autonomous vehicles, and companies like Amazon and Google for image and speech recognition technology.

Why was deep learning not available for cyber security before now?

Deep learning technology requires knowledgeable data scientists and vast amounts of computing power. Thanks to the cloud and advances in processing power, deep learning is now a mainstream reality.

There are only six deep learning frameworks existing within the world. One of which, number six, is built by Deep Instinct and is a purpose built, end-to-end deep learning cyber security solution. The only one of its kind in the world.

Due to nature of this deep learning technology, Deep Instinct offers more than 99% threat detection accuracy with less than 0.1% false positives. Deep learning self learns, much like the human brain making it a reliable and robust solution to support your organisation’s cyber security strategy.

Watch the video to learn more


Dan Clipston
Dan Clipston

Dan is Phoenix's cyber security endpoint specialist. He has four years of experience within the IT industry and works with our cyber security and GRCE vendors to deliver a core security message.

See all posts by Dan Clipston