EvilModel: Hiding Malware Inside of Neural Network Models
Published in 26th IEEE Symposium on Computers and Communications (ISCC 2021), 2021
Recommended citation: Wang, Z., Liu, C., & Cui, X. (2021). Evilmodel: hiding malware inside of neural network models. In 2021 IEEE Symposium on Computers and Communications (ISCC) (pp. 1-7). IEEE. https://doi.org/10.1109/ISCC53001.2021.9631425
BIB Paper Implementation 1 Implementation 2 (by @Gábor Vecsei) Implementation 3 Implementation 4
Abstract Delivering malware covertly and evasively is critical to advanced malware campaigns. In this paper, we present a new method to covertly and evasively deliver malware through a neural network model. Neural network models are poorly explainable and have a good generalization ability. By embedding malware in neurons, the malware can be delivered covertly, with minor or no impact on the performance of neural network. Meanwhile, because the structure of the neural network model remains unchanged, it can pass the security scan of anti-virus engines. Experiments show that 36.9MB of malware can be embedded in a 178MB-AlexNet model within 1% accuracy loss, and no suspicion is raised by anti-virus engines in VirusTotal, which verifies the feasibility of this method. With the widespread application of artificial intelligence, utilizing neural networks for attacks becomes a forwarding trend. We hope this work can provide a reference scenario for the defense on neural network-assisted attacks.
Keywords Neural Networks, Malware, Steganography, Artificial Intelligence