Applications of Machine Learning in Cybersecurity: Network Packet Classifer

Machine learning and cybersecurity are two of the technology fields which companies invest most of their money into. This project aims to explore potential next-generation cybersecurity solutions by utilizing artificial intelligence and machine learning in the hopes of reducing costs spent on fixing software vulnerabilities for enterprises. The ultimate goal is to provide a blueprint for a cyber resilient system that allows for real-time dynamic endurance. One of the applications of machine learning in the cybersecurity field is intrusion detection. Intrusion detection is a security practice in which companies monitor their networks and applications and attempt to preemptively discover malicious attacks on their systems. The earlier companies are able to detect a threat, the more time they have to appropriately respond and maintain a secure environment. However, this is a costly practice, especially if threats remain unidentified for an extended period of time. My proposed solution to this problem is to combine machine learning techniques with the intrusion detection field in order to create a project that is able to accurately identify and respond to these attacks without human intervention. This would drastically decrease the cost and time spent for the business on identifying potential security breaches. The intent of the solution is to build the concept of an intrusion detection capability which collects information on monitored events for companies, detects signs of intrusion, and enables intrusion response based on the system requirements. This capability will be based on machine learning and analytics from previous cyber attacks and it will eventually be integrated into company workflows and development pipelines to run in an automated fashion. My project will specifically contribute towards this goal by providing a proof of concept of the ability to apply machine learning to intrusion detection. The project will focus on using machine learning techniques to analyze network packets from a simulated environment and determine whether the packets are malicious or benign. This is the first step in intrusion detection, and it helps with warning a system on whether incoming network traffic contains a malicious attack. This project can be expanded on in the future to include attempts to block or otherwise resolve these attacks and malicious network packets.