Del Machine Learning al Deep Learning, neutralizing the Ransomware associated with Industry 4.0

Automatic learning is the future of cybersecurity and so Deep learning will help you neutralize the Ransomware linked to the industry 4.0

As technology evolves for efficiency and innovation, so does malware by generating infinitely more complex malicious files to be detected, just as Ransomware is. It is not a surprise that, in the study carried out by Sophos to 2700 medium sized companies, the central conclusion is that one of every two companies analyzed in Latin America suffered attacks even when 77% of these had protection with updated security Systems.

It can be very discouraging to cite these statistics, when attacks also reach Medium-high-level corporates in cybersecurity culture, however, not everything is bad news: as the attacks evolve the same way the technology should do of Protection. What is the answer to this problem? Migrating from Machine Learning to a Deep learning capable of making decisions and learning through its artificial neurons previously loaded with information on the Ransomware.

Deep Learning’s scopes on machine learning are exponential:

  • While machine learning identifies two-dimensional malware, regularly size and compression of a file, deep learning makes decisions from the updated DNA of the Ransomware as a typical vertical attacker of industry, multiplying in hundreds of thousands the possibilities of Detection.
  • Deep Learning significantly reduces the weight and performance of a protection system, since it chooses a precise decision path, oriented by the information of the preloaded Ransomware dna, with hyper sophisticated characteristics and capable of recognizing Even the best disguised Ransomware.
  • A tool based on Deep learning is capable of evolving and learning while managing real-time detection, even without requiring updates via the internet, since it feeds evolves and learns from the operation of each Workstation.
  • The artificial neural networks (ANN) used by Deep Learning mimic biological neuronal behavior by taking an entry, manipulating and sending information to other neurons from the right decision making previously, which makes it smarter as you progress in analyzing a company.
  • Through Deep Learning, The protection system reduces false positives in its entirety, because it recognizes from the behavior and characteristics in hundreds of thousands of layers of Ransomware as a tomography in a human brain.

The real power of Deep Learning is that if you are properly trained, you can reliably detect the threats you have not seen before. Additionally, through its growing intelligence it becomes especially effective in blocking new threats proactively and predictably.

When analyzing options to protect the prestige and continuity of your business it is advisable to make specific questions to your security provider:

  1. Does your Machine Learning algorithm make real-time decisions?

The dataset must be compact enough to keep you in memory, thus avoiding the need to keep the reading of detection data off the disk as each file is scanned.

  1. Is it real time?
  2. How much time do you need to make a decision?
  3. What happens to performance and accuracy if the computer is offline?
  4. What is your training set?
  5. Where do the training data for the security system come from?
  6. What makes it realistic? How much data is there? How do you keep training sets up to date?

These factors are usually the main vulnerabilities that make it insufficient to protect with only Machine Learning, for this reason Sophos and OnLine licenses recommend exploring tools that involve technologies with Deep Learning and an expert in Cybersecurity capable of maximizing the potential of these disruptive technologies.

For more information see