Android malware detection using deep learning

Android malware detection is a critical cybersecurity practice that aims to identify and mitigate malware threats that target the Android operating system. With the widespread use of Android devices, the importance of effective malware detection has increased significantly. This abstract provides an overview of the key aspects of Android malware detection, including its goals, techniques, challenges, and importance.Android malware detection covers a wide range of goals, including identifying known and emerging malware types, early detection of threats, minimizing false positives, and protecting user privacy. In this process, various techniques such as signature-based identification, behavioral analysis and deep learning methods and machine learning and network monitoring are used to identify and effectively respond to malware threats.Challenges in detecting Android malware stem from the diversity and complexity of Android malware, as attackers continually develop new evasion techniques and attack vectors. Balancing detection accuracy with minimal impact on device performance and user experience is also a constant challenge.The importance of Android malware detection is characterized by its role in protecting user data, financial security, and personal privacy. It reduces the risks of identity theft, data loss, and fraudulent activity while strengthening user trust in the Android ecosystem. In addition, it helps with network security and compliance with legal and regulatory requirements.In this research, we propose a new approach for Android malware analysis and classification that uses the power of one of the BERT (Bidirectional Encoder Representations of Transformers) (DistilBERT) models to classify API call sequences generated from the Android API call graph. Using an API call graph, our approach captures the complex relationships and dependencies between API calls, enabling a deeper understanding of the behavior exhibited by Android malware. Our results show that our approach achieves high accuracy in classifying API call sequences as malicious or benign, and this method also provides a promising solution for Android malware classification and can help reduce the risks caused by Android malicious apps.

کلیدواژه ها:

Deep Learning ، Android Malware Detection ، Malware Analysis ، Machine Learning