Key points are not available for this paper at this time.
Efficient malware identification is essential to safe the system resources and privacy of data for cybersecurity system. The use of android smartphones has increased tremendously that attracting various types of malware attacks. Nowadays, malware writers use Artificial Intelligence (AI)-enabled malware attack techniques to bypass the detection of malicious activities. Hence, designing an efficient, effective and robust malware detection system to identify malware variants remains a critical problem and challenge. However, number of deep learning (DL) models applied for effective android malware detection in existing methods at large scale, but these models actually lacks interpretability of the models to explain the contribution of each features to the detection system. Therefore, this paper propose an Explainable Artificial Intelligence (XAI) based hybrid Convolutional Neural network (CNN) and Bi-Gated Recurrent Unit (Bi-GRU) Android Malware Detection (AMD) System using DL models named as XAI-AMD-DL. The proposed model is evaluated the using the CICAndMal2019 android malware dataset. The results obtained by the proposed XAI-AMD-DL model is 97.98% accuracy, and 97.75 %, 97.76%, 97.75% precision, recall and f1score, respectively outperforms the existing DL models.
Smmarwar et al. (Sat,) studied this question.