Integrating Web Server Log Forensics through Deep Learning

Abstract

The world of Cyber Forensics is often filled with gigantic amounts of information, often more than what you would get from engagements in other branches of forensics. This not only makes the engagement much more thrilling for forensic experts, it also makes it much more tedious and a huge time-consuming factor when it comes to analysis. There are several tools available both from the open-source community and private devs, but not much from the fields of Artificial Intelligence (AI). Deep Learning, being at the core of Artificial Intelligence, will provide us with much better and more refined processing and predictions based on the available data. The setbacks and the breakthrough of using Deep Learning in Cyber Forensics are more or less the same as in every other branch, where AI is used to solve tasks critical to a person, or most of the time, crucial to an organization. To start with, for Deep Learning to be integrated with the fields of Cyber Forensics, i.e., after an incident, it must also be trained in the areas of Cyber Security, or to be exact, in Cyber Defense, i.e., before an incident. This idea is pretty intuitive. This paper looks at Deep Learning models as much similar to the most complex structure in the known universe, the human brain. After all, these models have been inspired and based on the human brain. This paper attempts to find existing solutions on how to best implement a Deep Learning model in the fields of Cyber Forensics and proposed how Deep Learning models could help the world of Cyber Security, especially for the IR teams.