Market Insights: the benefits of deep learning with Sam Linford, Deep Instinct

We catch up with Sam Linford from Deep Instinct to find out how deep learning is changing the cyber protection landscape and why organisations need to invest in it to plug cyber security gaps.

What is today’s greatest cyber security challenge?

State sponsored cyber attacks designed to cause massive economical manipulation, disruption, and the theft of intellectual property and personal data are certainly the biggest challenges today. Typically, cyber attacks start with a malicious file landing on a hard disk, and organised crime groups are no different in their technical approach. By implementing protection – such as deep learning – the risk surface is reduced enormously.

Legacy solutions, such as reputation-based traditional AV and EDR technologies are no longer enough against modern cyber threats. Pre-execution prevention is key, completely changing the way a SOC team operates and investigating security gaps where cyber criminals can gain access to your systems.

What gaps do you often see in cyber security strategies?

When it comes to endpoint security, there is a risk surface gap that sits above all of the more traditional technologies we commonly use to protect ourselves, including traditional antivirus and endpoint detection and response (EDR) tools. At Deep Instinct, we rate your level of protection based on:

  • 0% protection: passwords written down
  • 30% protection: covered by traditional AV
  • 60% protection: more advanced EDR solutions implemented

To defend against modern threats and an ever-changing threat landscape, organisations with 60% coverage from EDR protection, must address that 40% gap with deep learning-based prevention technology to ensure full protection.

Why deep learning vs machine learning?

Machine learning has been around for a very long time – in fact, it’s legacy technology! Machine learning is totally reliant on human interaction; feature engineers extracting interesting attributes of a file and then feeding this into a detection framework. But in reality, only 2% of the raw data makes its way into this framework, which means efficacy is low and limited file types are covered.

On the other hand, a specifically designed cyber security deep learning framework has the autonomous ability to analyse 100% of the raw data available, both malicious and benign. This gives our solution the unique ability to stop over 99% of ‘unknown’ threats (threats that have not yet been seen in the wild).

Why is deep learning so important and should it become the new standard?

Yes, it definitely should be the new standard. Deep learning is fast. It’s as quick as real time when it comes to detection and prevention and has the unique ability to stop zero-day attacks in a way that no other form of AI can. Efficiency is much higher compared to any previously designed machine learning AI.

Deep learning has a high entry barrier that requires elite data scientists and vast computing power. There are only a handful of organisations out there that truly provide a real deep learning framework and I’m very pleased to say that Deep Instinct is one of these!

What advice do you have for organisations with limited cyber security in place? 

Attacks from organised crime groups and state sponsored initiatives are rapidly becoming more complex and organisations with limited cyber security defences must reduce their attack risk surface dramatically. Pre-execution prevention should be the main focus before implementing post-execution solutions like EDR. As I always say, “What would you prefer, CCTV on the inside of your house so that you can see what is being stolen or a really strong front door to stop them getting in in the first place?”.

What type of organisations are often targeted and need to be prepared? 

Organised crime groups are often focused on large enterprise, government, and utilities. However, we regularly hear about small and mid-size organisations coming under attack from ransomware. You should never assume that your organisation won’t be hit by an attack and every organisation should be prepared. Prevention is far more cost-effective than remediation.

Why is it important to secure budget and buy-in to support cyber security? 

Hoping that a breach doesn’t occur isn’t a cyber security strategy – however, we still see organisations taking this approach. The associated costs with paying ransoms and remediation are astronomical and if you don’t secure budget and buy-in to invest in a proper cyber security strategy, you’re leaving your organisation vulnerable to losing a lot of money.

Plug the gap now – invest in deep learning for your organisation 

Chat to our cyber security specialists now to check for gaps in your organisation’s security strategy and how to fill them for complete protection against modern threats.

Book your free consultation now


Read the other blogs in our Market Insights series:

Market Insights: cyber security with Matt Knell, Sophos – part one

Market Insights: cyber security with Matt Knell, Sophos – part two


Ben Murden
Ben Murden

Ben has over two decades in the IT industry, delivering both online and offline campaigns across all platforms to meet business goals and objectives. Joining Phoenix in 1999 as a graphic designer, Ben has evolved over the years into a fully-rounded marketing professional, before being promoted to Phoenix Marketing Manager early in 2018, reporting directly to the MD. With his background in creative design, Ben takes projects from inception to execution and can identify the correct strategy based on the subject, audience, and goals – while increasing the brand profile and revenue. His passion for digital marketing is evident in everything he does, and both vendors and strategic partners often comment on his incredibly positive attitude to ‘make things happen’.

See all posts by Ben Murden