Cybersecurity refers to the technologies, processes, and practices designed to protect networks, computers, programs, and data from attack, damage, or unauthorized access. In today’s digital world, cybersecurity is an essential element of protecting personal and business information, and it is increasingly important as more and more of people’s lives and operations move online. Some of the major challenges in the field of cybersecurity include the rise of sophisticated cyber attacks, the increasing use of artificial intelligence and machine learning, the growing complexity of the technology landscape, and the shortage of skilled professionals.
Cybersecurity is the act of protecting valuable assets from attacks that are initiated mainly through digital means. The main targets of cyberattacks are not limited to Information Technology (IT) assets, but also include Operational Technology (OT) as well as human assets. A successful attack can result in considerable costs in terms of resources and data as well as reputation damage for the organisation concerned.
In order to enhance the preparedness and cyber-resilience of targeted assets to the greatest extent possible, there is currently a variety of tools that may be utilized. Those include Intrusion Detection Systems (IDS), Intrusion Prevention System (IPS) and Cyber Threat Intelligence (CTI) tools.
M4D is focused on the development of AI-supported CTI tools for the identification, gathering, analysis, correlation, classification, enrichment and sharing. To this end, relevant data from multiple system-based (e.g., logs) and external sources (e.g., online content from the Surface & Dark Web) are being collected through the utilization of honeypots and custom made crawlers. Actionable CTI is extracted from the gathered data (e.g., through Named Entity Recognition and relation extraction using rule-based and machine learning-based techniques), correlated, and shared (e.g., through MISP), while it can also be further enriched through its association to relevant taxonomies.
M4D is also developing dynamic taxonomies through the aforementioned CTIs and based on existing taxonomies (e.g., from MISP). To this end, information extraction techniques such as Name Entity Recognition, BERTopic modeling, and pattern matching are employed to create the cyber threat related taxonomies and update the existing taxonomies. Finally, the new taxonomies are used to develop dynamic ontologies focused on cyber threats raising the awareness and enhancing the preparedness against cyber attacks.
Research activities also expand to the analysis of logs produced by Host Based Intrusion Detection systems to identify attack patterns using Association Rule Learning while using Federation Learning techniques M4D is researching how Collaborative Intrusion Detection Systems (CIDS) can enhance the detection accuracy of cyber attacks and their overall effectiveness.
OPEN ACCESS