The Microsoft Threat Protection Intelligence Team and Intel Labs are working together to facilitate the identification of malicious software through Deep Learning. For this, researchers have made use of STAMINA, from the English “Static Malware Analysis Network as an Image”.
The program transforms the binary code of threats into grayscale images, converting, reshaping and resizing each pixel. Each byte receives a value from 0 to 255, indicating the intensity of gray. The file size is then used to determine the scale of the image, which is transformed into a two-dimensional photo.
At the current stage, the results are very promising, with an overall accuracy of 99.07% and a false positive index of 2.58%. However, the technique is effective with static analysis, making it necessary to use standard software behavior procedures for more complex threats. Still, the methods can be combined so that antivirus software can achieve an even higher level of protection.
It is worth remembering that Microsoft started testing for the implementation of Edge Chromium as the standard Windows browser, with the novelty initially coming to members of the Insider program. The company also made available to all users of the Office suite on iPadOS today official support for the system’s Split View mode. Meanwhile, Intel has updated the driver for its integrated GPUs in preparation for the major May update of Windows 10.