Research on Malware Detection and Classification Based on Artificial Intelligence.

Abstract

Malware remains one of the major threats to network security. As the types of network devices increase, in addition to attacking computers, the amount of malware that affects mobile phones and the Internet of Things devices has also significantly increased. Malicious software can alter the regular operation of the victim's machine, damage user files, steal private information from the user,steal user permissions, and perform unauthorized activities on the device. For users, in addition to the inconvenience caused by using the device, it also poses a threat to property and information. Therefore, in the face of malware threats, if it can accurately and quickly detect its presence and deal with it, it can help reduce the impact of malware. To improve the accuracy and efficiency of malware detection, this article will use deep learning technology in the field of artificial intelligence to study and implement high-precision classification models to improve the effectiveness of malware detection. We will use convolutional neural networks and long and short-term memory as the primary training model. When using convolutional neural networks for training, we use malware visualization techniques. By converting malware features into images for input, and adjusting the input features and input methods, models with higher classification accuracy will be found; in long-term and short-term memory models, appropriate features and preprocessing methods are used to find Model with high classification accuracy. Finally, the accuracy of small sample training is optimized by generating features for network output samples. In the above training, all of us want to use malware as a sample that affects different devices. In this article, we propose three research topics: 1). When importing images, high-precision models are used to study malware. 2). When importing non-images, a high-precision model will be used to study the malware. 3). By using this model, the generated adversarial network is optimized for small sample malware detection.