Using a Machine Learning approach to develop a multilingual capable system for collecting and evaluating cyber threat intelligence from online communities.

In recent years many cyber threat intelligence (CTI) solutions and services have been developed, primarily for use by companies and institutions. These organisations collect cyber threat related to computers, networks and information technology and transform the information to threat intelligence. However, traditionally cyber threat intelligence services collect and analyse data after attacks have happened. Conversely, data from online communities are not widely analysed by these solutions due to cost and lack of perceived necessity.

In today’s society, online communities and forums have become increasingly interwoven with our day-to-day life. People use these platforms to communicate and connect with like-minded people and, as result, a considerable amount of data and information is transferred via these online communities.

This project will explore the threat information that presents in online forums and communities, and novelly will include a multi language capability system, and utilising machine learning algorithms that could be employed to collect and analyse either real-time or historical data from these communities. The core of this project will be to identify any correlation that might exist between the likely information available and its usefulness to adversaries in the exploitation of the vulnerabilities. Hence, the project proposes to develop an evidence based multi-lingual knowledge system that can be used to identify cyber security vulnerabilities and threats.