A RobustEmbedded Long-Short Term Memory (eLSTM) Learning Based Framework for Classification of Malware

Abstract

Malware poses a significant threat to digital security, as it is designed to access or manipulate information from systems without user consent or authorization. As the digital landscape evolves, malware attacks by cyber-criminals are becoming increasingly sophisticated and frequent. To effectively counter these threats, this paper introduces a novel framework that combines Long-Short Term Memory (LSTM) and Convolutional Neural Networks (CNN) for malware detection and classification. The proposed framework leverages the strengths of LSTM and CNN neurons to process and learn from locally available data that exhibits malware-like characteristics. By abstracting and correlating relevant data, the model is capable of identifying patterns and anomalies associated with malware activity. This combination of LSTM's ability to capture temporal dependencies and CNN's feature extraction capabilities enhances the accuracy and robustness of malware detection. Through rigorous experimentation and evaluation, the framework demonstrates superior performance in classifying malware compared to traditional methods. The results suggest that this approach provides a powerful tool for identifying and mitigating potential malware threats, contributing to a safer and more secure digital environment.