New AI Model Achieves 99.44% Accuracy in Android Malware Classification

Global: New AI Model Achieves 99.44% Accuracy in Android Malware Classification

A team of researchers has introduced a novel machine‑learning approach that combines the Fast Gradient Sign Method with a Diluted Convolutional Neural Network to classify Android malware, reporting an accuracy of 99.44 % in initial tests.

Method Overview

The proposed FGSM DICNN framework leverages diluted convolutions, which expand the receptive field without increasing the number of parameters. This design enables the network to capture dispersed malicious patterns across longer sequences while relying on fewer input features.

Diluted Convolutional Architecture

By inserting gaps between kernel elements, the diluted convolutional layers reduce computational overhead and maintain model compactness. Consequently, the architecture can process extensive code fragments typical of Android applications without the memory and processing demands of traditional deep networks.

Integration of Fast Gradient Sign Method

The Fast Gradient Sign Method (FGSM) is applied during training to generate one‑step adversarial perturbations. This strategy enhances the model’s robustness and improves classification accuracy while keeping training costs low.

Performance Evaluation

Experimental results, based on a benchmark dataset of Android applications, show that FGSM DICNN attains 99.44 % accuracy, surpassing comparable models such as a custom Deep Convolutional Neural Network (DCNN). The study attributes the improvement to the combined effect of a broader receptive field and adversarial training.

Implications and Future Work

According to the authors, the approach reduces dependence on large feature sets, potentially streamlining malware detection pipelines for enterprises and security vendors. Future research will explore real‑time deployment scenarios and assess resilience against evolving malware families.

This report is based on information from arXiv, licensed under Academic Preprint / Open Access. Based on the abstract of the research paper. Full text available via ArXiv.