Most programs aimed at killing malicious software need someone to tell them something is malicious. Deep Instinct, a company specialized in operational cyber-security, says its security solution, launched last week, can identify bad code on their own.
The offering works its magic with a technology called ‘deep learning’. With traditional programming, as you code, you have to anticipate all the situations that arise that you have to deal with. What deep learning does is take the data and build a model from what it finds in the data that's statistically relevant.
That kind of ‘learning’ can be very important in a security app. What makes malware difficult to detect through traditional programming methods is that the slightest change in malicious code can fool a program.
It I like looking at the picture of a cat and then not being able to recognize the image because someone modifies a few pixels of that picture. Deep leaning is extremely resilient to variance and modification; it will be able to identify an image as that of a cat by just seeing the tail of the cat.
Deep Instinct's security solution agent takes up about 10 MB of memory installed at each endpoint — laptop, mobile device or server — with deep learning technology in it. Most of the time this agent does nothing, but when it detects any type of new file it passes it through the deep learning module on the device. If the file is malware, it will immediately remove it or quarantines it.
When compared to 61 other solutions and in all the benchmarks it was tested against, Deep Instinct’s solution had 98 to 99 percent detection. The best solution out of the other 61 solutions scored 79 percent detection.