Malware Classification Based on Deep Learning

This paper proposes a novel malware classification method for personal computers, which I have named MalCNN. The motivation behind selecting this topic is that, in the era of the pandemic, as online work continues to grow, computer security has become an increasing concern.Currently, most malware classification methods rely on either a hard decision strategy or analysis conducted by cybersecurity professionals. However, traditional classification models and approaches are no longer suitable for the big data era. In the digital world, millions of malware programs spread every moment, making traditional classification methods inefficient due to their slow speed and inaccurate results. As a result, these methods struggle to meet modern security demands.To address these challenges, MalCNN attempts to establish a soft decision classification method based on computer vision techniques, allowing for the annotation of samples that do not meet predefined classification conditions.