Rooted Learning Model at Fog Computing Analysis for Crime Incident Surveillance
Romil Rawat, Vinod Mahor, Josefa Díaz‐Álvarez, Francisco Chávez
Abstract
Cyber Loopholes in smart devices' applications invited intruders to conduct malicious activities. The growing quantity and diversity of smart devices has posed significant cyber security concerns. Cloud computing centralization makes it difficult to provide dispersed vulnerability tracing services for IoT designs at network's edge. Despite the existing encryption methods on the conventional Internet, the inevitability of detection mechanisms has been demonstrated by growing attack surfaces and hacking talents. Machine Intelligence focused threat detection mechanisms have demonstrated lower accuracy and scalability for threat detection in massively dispersed nodes -Internet of things. Modern Deep Learning may pave the way for identifying the present level of complexity of vulnerabilities towards edge networks and has proven effectiveness in large data applications. Due to huge quantity of data created by Internet of things Nodes, Deep Learning can learn better than Shallow Learning algorithms. Fog computing or fog networking (fogging) might be the ultimate benefactor of the method for threat detection. Here we present a novel distributed Deep Learning technique for cyber attack and vulnerability injection detection. The Deep Learning model outperforms Shallow Learning models in terms of assessment parameters evaluated in the experiments.