Agile Defense: Leveraging Continual Learning for Adaptive Cybersecurity

yber-attack techniques are becoming increasingly sophisticated, breaching even the toughest defenses and requiring smarter, more agile solutions. Current machine learning (ML) algorithms identify and predict threats but rely heavily on past datasets, requiring significant updates. Continual learning offers a solution by enabling automatic adaptation to new threats. In this PhD project, you will:

- Explore continual learning techniques to improve cyber-attack identification;

- Investigate the application of continual learning and Natural Language Processing (NLP) to automate the cyber-attack attribution process. - Establish solutions for automatic identification of attacker-oriented countermeasures.

You will explore how attackers adapt and evolve their techniques to overcome existing defenses and how continual learning can help in detecting these evolving threats. Additionally, you will analyze the provenance of used malware/attack techniques and the context of attacks to enhance the attribution process. This includes leveraging NLP techniques to parse and analyze threat intelligence reports, which will help in identifying the attackers and understanding their motivations and methods.

Your work will involve a blend of technical and analytical skills, including malware analysis, intrusion detection system (IDS) features, and understanding the broader ecosystem of cyber threats. The goal is to create solutions that are not only technically robust but also contextually aware, leading to more effective and tailored cybersecurity measures.