“Hide and seek” malware creates “botnets” by quietly infecting massive numbers of devices using advanced communications. CPC’s team created real novel breakthrough by turning evasive malware into graphic images. They “mixed and matched” computer vision with AI algorithms, resulting in breakthrough malware classification.
CPC’s AI algorithm for malware classification uses Machine Learning and Visual Computing to identify malware, without reading file contents, without taking time to observe behavior, or even require very much computation. CPC’s technique cleverly processes suspicious files with a 3D RGB color image algorithm. Then a specialized “autoencoder” and a “weakly supervised learning network” discover hidden features of suspicious files. Small visual images can represent the entire original data, and are easily managed. The amazing breakthrough is when the CPC method uses AI-enhanced computer vision to “view” transformed images, detecting malware. In other words, this is “facial recognition” for malware, successful even when malware “hides” using disguises. CPC’s new method can use GPUs to handle image tasks, freeing up the main CPUs. CPC’s new method can more efficiently and effectively tackle increasing, more frequent and complex threats, with thorough analysis of past threats and behaviors so as to prevent them in the future. AI Visual Security can also identify upcoming mutated threats because even mutated threats under the same family could still be easily identified before they pose any risk.
Thank you for your enquiry.