Deep Learning Framework for Cybersecurity: Framework, Applications, and Future Research Trends

Abstract

With the development and modernization in technology and Internet, cybercrimes are also increasing rapidly and cybercriminals are using different and new approaches or methods almost every time. Previously, Petri Nets formalization were used to model and tackle cyberattacks but due to advancement in technology and Internet, cybercriminals are using various machine learning (ML) and deep learning (DL) approaches to commence cyberattacks. As DL approaches were used in processing of image and recognition of speech, cybersecurity experts were busy using DL techniques to tackle those cyberattacks. This survey report focuses on DL methods put forward recently for cybersecurity. First, some basic need of cybersecurity including issues is described. Then favored models and algorithms of DL are explained. Subsequently, frameworks of DL required in the area of cybersecurity applications are proposed. Afterward, work on cybersecurity based on DL is discussed which include detection of malwares, intrusions, phishing, spam, and Web site defacement. Finally, improvement of various cybersecurity applications is suggested as discussed in the future research scopes followed by some concluding remarks.