.png){kind=icon}
.png){kind=icon}
Return
Efficient malware detection through inter-component communication analysis
.png){kind=icon}
PRE
AI
DOI：10.1007/s10586-024-04530-3.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
With the development of science and technology, the number of smartphones has increased dramatically. This also exposes Android-based smartphones to an increasing number of malware attacks. Currently, feature extraction schemes based on sensitive API calls have become mainstream in malware detection. However, such an approach can only capture the behavior of malware when the API is called, but it cannot detect malicious behavior implemented through other means. In the real world, attackers often use the Inter-Component Communication (ICC) mechanism to hide and conceal their malicious intent. In this paper, we propose a novel malware detection framework (named ADACapsNet). This framework first employs an entropy-based approach to extract sensitive API features to reflect the behavior patterns of malware and then captures the information flow across components by monitoring the ICC interactions in the Android systems. We transform sensitive API calls and ICC features into vector representations that are used as inputs to the learning model. Moreover, we propose an adaptive capsule network to mine deep program semantics, which uses an adaptive factor to dynamically assign weights for features, enhancing the model's ability to focus on relevant features and capture complex spatial relationships. We conducted a number of experiments to demonstrate the effectiveness of the proposed ADACapsNet in detecting malware. Experimental results show that the proposed method is robust against malware attacks.
Keywords:
Internet of things
Malware detection
Inter-component communication
Capsule network
Journal
C
IF:
4.1
Papers:
5.0K
Citations:
7.5K