Deep Learning-Based Mitigation of Fabricated Attacks in the Industrial Internet of Things
Ramu Kumar Beldar · 2024
The rapid growth of data-driven interconnected devices in the Industrial Internet of Things (IIoT) and the emergence of advanced 5G/6G networks highlight the need for secure data extraction. However, the increasing complexity of Android malware complicates conventional detection methods, as adversaries adeptly evade classification. To address counterfeit attacks in the IIoT, a novel strategy is proposed: utilizing deep learning models to extract application-aware features from Android applications for training a robust model. The proposed deep learning approach achieved remarkable accuracy in detecting malware using just 56 synthesized input features. This method effectively counters adversarial evasion attacks and enhances IIoT security. By leveraging deep learning and application-aware feature extraction, this approach offers a promising solution for combating counterfeit attacks in the IIoT ecosystem, contributing significantly to its security and resilience amid the evolving Android malware landscape.