Intel and Microsoft are collaborating to create a perfect anti-malware product.
STAMINA, which stands for STAtic Malware-as-Image Network Analysis grabs samples of different malware, turning them into 2D grayscale images which structural textures can be singled out, helping to identify similar malicious software.
STAMINA operates on deep learning — a subfield of machine learning that allows computers to imitate the way the human brain. In STAMINA’s case, this knowledge is based on a binary of malware that is transformed into raw pixel data. It is then scanned by a deep neural network, taught to recognize 60% of malware known to man.
While STAMINA does resize the converted images to speed up the scanning, the Intel-Microsoft developers say it doesn’t affect the results in a negative way. In fact, the project has shown 99.07% accuracy, with 2.58% in false positives, ZDNet reported on May 11th.