Malware Detection in Internet of Things (IoT) Devices Using Deep Learning.

Area of Science:

• Computer Science
• Cybersecurity
• Artificial Intelligence

Background:

• Internet of Things (IoT) device proliferation increases data capacity, making them vulnerable to sophisticated malware attacks.
• Accurate and efficient malware detection mechanisms are crucial for securing IoT ecosystems.
• Existing malware detection methods face challenges in achieving high accuracy and reliability.

Purpose of the Study:

• To propose a novel deep learning-based ensemble classification method for effective malware detection in IoT devices.
• To address the need for reliable and time-efficient identification of sophisticated malware targeting IoT systems.
• To enhance the security posture of Internet of Things devices against evolving cyber threats.

Main Methods:

• A three-step approach involving data preprocessing (scaling, normalization, de-noising).
• Feature selection and one-hot encoding for data transformation.
• An ensemble classifier combining Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM) outputs.

Main Results:

• The proposed deep learning ensemble method achieved an average accuracy of 99.5% on standard datasets.
• Demonstrated superior performance compared to state-of-the-art malware detection techniques.
• Validated the effectiveness and efficiency of the proposed approach for IoT malware identification.

Conclusions:

• The deep learning-based ensemble classification method offers a highly accurate and reliable solution for IoT malware detection.
• The proposed approach significantly advances the state-of-the-art in securing Internet of Things devices.
• This method provides a robust framework for combating sophisticated malware threats in the expanding IoT landscape.