MALWARE DETECTION BASED ON DEEP LEARNING AND BEHAVIOR GRAPHS

MALWARE DETECTION BASED ON DEEP LEARNING AND BEHAVIOR GRAPHS

1Ch.Modini, 1C.Sravya Reddy, 1D.Pranoy Chowdary, 1D.Shashank Reddy,

2B. R Sivasubramanian and 3Dr.Thayyaba Khatoon

1UG Students , 2Assistant Professor , 3Professor & HoD
Department of Artificial Intelligence and Machine Learning, School of Engineering , Malla Reddy University, Maisammaguda, Dulapally,

Hyderabad, Telangana 500100

Abstract
Malware is one of the most frequent cyberattacks, with its prevalence growing daily across the network. Malware traffic is always asymmetrical compared to benign traffic, which is always symmetrical. Fortunately, a variety of artificial intelligence approaches are available for identifying malware and differentiating it from everyday operations. The goal of this paper is to develop a reliable and effective malware detection system for computer networks. Traditional signature-based malware detection techniques are not sufficient to detect the increasingly sophisticated and polymorphic malware threats. The proposed deep learning-based behavior graph approach aims to address this problem by analyzing malware's behavioral patterns rather than relying on specific signatures. The paper uses graph neural networks to model the behavior of malware. The model learns the characteristics of malware by analyzing the behavior graphs of legitimate and malicious software. The behavior graphs represent the system calls and API functions used by the software, which the model uses to detect malicious activities.

Keywords: Deep Learning, Malware detection, Behavioral graphs, Malware, Benign.