Explainable Artificial Intelligence-Based IoT Device Malware Detection Mechanism Using Image Visualization and Fine-Tuned CNN-Based Transfer Learning Model

Abstract

Automated malware detection is a prominent issue in the world of network security because of the rising number and complexity of malware threats. It is time-consuming and resource intensive to manually analyze all malware files in an application using traditional malware detection methods. Polymorphism and code obfuscation were created by malware authors to bypass the standard signature-based detection methods used by antivirus vendors. Malware detection using deep learning (DL) approaches has recently been implemented in an effort to address this problem. This study compares the detection of IoT device malware using three current state-of-the-art CNN models that have been pretrained. Large-scale learning performance using GNB, SVM, DT, LR, K-NN, and ensemble classifiers with CNN models is also included in the results. In light of the findings, a pretrained Inception-v3 CNN-based transfer learned model with fine-tuned strategy is proposed to identify IoT device malware by utilizing color image malware display of android Dalvik Executable File (DEX). Inception-v3 retrieves the malware’s most important features. After that, a global max-pooling layer is applied, and a SoftMax classifier is used to classify the features. Finally, gradient-weighted class activation mapping (Grad-CAM) along the t-distributed stochastic neighbor embedding (t-SNE) is used to understand the overall performance of the proposed method. The proposed method achieved an accuracy of 98.5% and 91%, respectively, in the binary and multiclass prediction of malware images from IoT devices, exceeding the comparison methods in different evaluation parameters.

Deep learning techniques for detection and prediction of pandemic diseases: a systematic literature review

“… The study used an hybrid method comprising DL and ML models, yet the detection and classification performance of the study was low, this might be liken to computational complexity on the dataset. [ 69 ] In this study, three pretrained cutting-edge CNN models were used to compare the detection of malware on IoT devices. CT chest images in the COVID-19 Radiography Database Inception-v3 CNN-based NA accuracy of 98.5% and 91%, respectively The results of this study can be further improved with the introduction of better optimization and feature selection techniques [ 48 ] The study developed and test a new computer-aided diagnosis (CAD) for detection of coronavirus chest X-ray images CNN-based CAD scheme NA Accuracy 94.5%; Sensitivity 96.3% Specificity 97.2% The study demonstrated that additional two image preprocessing steps and generating a pseudo color image in a DL CAD scheme.…”

A novel method of detecting malware on Android mobile devices with explainable artificial intelligence

“…By leveraging both textual and visual attributes, the method aims to improve detection accuracy. Naeem et al [19] assesses pertained CNN models for detecting malware from IoT device. It also explores the effectiveness of combining these CNN models with different classifiers in large-scale learning.…”

Pengenalan Pola Serangan pada Internet of Thing (IoT) Menggunakan Support Vector Mechine (SVM) dengan Tiga Kernel

“…Penelitian [21] mendeteksi malware menggunakan pendekatan low access memory dan juga melakukan proses klasifikasi delapan kelas dengan metode SVM dan KNN. Dengan semakin banyak nya jenis malware yang mampu merusakan sistem IoT, maka sejalan dengan para peneliti dibidang keamaan IoT melakukan perbaikan dalam deteksi malware pada IoT berbasis Deep Learning, penelitian pada tahun 2022 yang dilakukan oleh [22] mendeteksi malware pada IoT CCN Inception-v3 untuk mengidentifikasi malware perangkat IoT dengan memanfaatkan tampilan gambar malware berwarna dari Android Dalvik Executable File (DEX), pada sebuah lapoan ilmiah (scientific reports) yang dibuat oleh [23] mereka pengembangan arsitektur baru untuk mendeteksi malware pada perangkat Internet of Things (IoT) yang menggunakan pendekatan berbasis Convolutional Neural Network (CNN), penelitian ini menggunakan arsitektur iMDA mempelajari kumpulan fitur yang beragam. Penelitian yang dibuat oleh [24] mendeteksi malware pada protokol IP Flow Information Export (IPFIX) dengan membuat kombinasi machine learning dan Manufacturer Usage Description (MUD) dengan tujuan mengurangi atau memperkecil nilai false positive sehingga nilai true positive rate menjadi tinggi.…”