End-to-End Deep Learning Prevents and Detects Malware

The ability to predict and prevent attacks that have never been seen before.

We had the opportunity to meet with Jonathan Kaftzan, V.P. of Marketing at Deep Instinct as part of IT Press Tour #35. According to Jonathan, Deep Instinct is the only company today using end-to-end deep learning to predict and prevent malware.


Jonathan began the discussion with an interesting query, “What would have happened if we could have prevented the coronavirus from causing any damage? What would you say if we would have a solution that could have predicted something like it based on a past a situation that was similar and could have predicted and therefore could have prevented this thing from happening? When in cybersecurity, the reality is that we have a Coronavirus situation almost every day. Deep Instinct has invented a solution that is based on deep neural networks that can predict attacks that are completely new that we have never seen before. And therefore can prevent these attacks from causing any harm.”


“When I'm talking about these things, think about the situation that we are facing right now. We are all in a situation where we're trying to detect the infection. Identify good people and to contain them. If I take the analogy to cybersecurity where many of the solutions that exist in the market are trying to detect infected computers, infected machines, infected networks, and are trying to contain that damage. And if you think about the situation we face today, where we are trying to detect to contain the damage, and the impact that it has on our life in cybersecurity, that's exactly what we're trying to avoid.”


Deep Instinct was founded in 2015 by three co-founders that came from very interesting backgrounds in cybersecurity, AI, machine learning, big data, and also deep learning. They realized that by combining the two, we can bring all the advantages and all the sophistication deep learning can provide to cybersecurity focusing on prediction and prevention with zero-time threat prevention rather than detection and response.


The cybersecurity challenge is not abating. There are more than 350,000 new malicious malware programs every day. 67% of CIOs think a data breach or cybersecurity attack is a given. Enterprise breaches cost anywhere from $40 million to $800 million. 69% of CISOs say their cybersecurity teams are understaffed. And, there are projected to be 3.5 million cybersecurity job openings in 2021.


Deep Instinct’s multi-layered protection uses more than 3 billion files to train its deep learning algorithm to detect and prevent threats in 20 milliseconds, analyze and investigate threats in 50 milliseconds, and remediate and contain threats in less than one minute.


HP Sure Sense, powered by Deep Instinct, is providing zero-time detection and prevention of zero-day threats and advanced persistent threat (APT) attacks for all Windows endpoints.

Deep learning identifies more than 99% of unknown malware with less than 0.0001% false positives while machine learning solutions are 50 to 70% accurate with 1 - 2% false positives.


Key Takeaways


The benefits of deep learning for cybersecurity are five-fold:


  • Prediction of unknown (future) threats

  • Zero-time detection and prevention

  • Zero-time classification

  • Protect any device, operating system, and file

  • Connectionless (edge deployment)

Drop Me a Line, Let Me Know What You Think

© 2020 by Tom Smith | ctsmithiii@gmail.com | @ctsmithiii