DEVELOPMENT OF A FEDERATED LEARNING-BASED MALWARE DETECTION MODEL FOR INTERCONNECTED CLOUD INFRASTRUCTURES
No Thumbnail Available
Files
Date
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Description
Due to the large number of heterogeneous applications using the same infrastructure, enforcing
security and reliability in the cloud is a difficult but crucial task. A security analysis system that
detects threats for example malicious software (malware) should exist within the cloud
infrastructure. Different malware techniques that bypass network-based and host-based security
protections have led to the development of new methods for analysing and detecting malware,
which have evolved over the past decades. Due to the complexity of learning the ever-changing
configurations of malware, it is challenging for forensics investigators to keep up with the
exponential rise in the number and variety of malware species. In this research work, a malware
detection model was developed for interconnected cloud infrastructures based on federated
learning. This technique enables collaboration between multiple devices in the training of machine
learning models without exchanging data, thereby preserving the privacy of individual users. Three
different deep-learning algorithms were selected and used in the training, validation, and testing of
the models. By the model training with eight clients and twenty-five federation rounds, the
FeedForward Neural Networks(FFNN) model provided the best performance with a precision of
84%, an F1-score of 84%, and an accuracy of 84% whereas the Multi-Layer Perceptron(MLP)
model provided 83% of precision, 83% of F1-score, and 83% of accuracy and the Long Short-Term
Memory(LSTM) model provided a performance with 80% of precision, 80% of F1-score, and 80%
of accuracy as well.
Keywords
T Technology (General), TK Electrical engineering. Electronics Nuclear engineering