Real-time Early Detection and Alert System for Online Terrorist Content based on Natural Language Processing, Social Network Analysis, Artificial Intelligence and Complex Event Processing

The RED-Alert project will bring data mining and predictive analytics tools to the next level, developing novel natural language processing (NLP), semantic media analysis (SMA), social network analysis (SNA), Complex Event Processing (CEP) and artificial intelligence (AI) technologies. These technologies will be combined for the first time and validated by 6 law enforcement agencies (LEAs) to collect, process, visualize and store online data related to terrorist groups, allowing them to take coordinated action in real-time while preserving the privacy of citizens. The RED-Alert solution will outperform state-of-the-art solutions in terms of number of languages supported, privacy-preserving capabilities, usability, detection performance, real-time capabilities and integration capabilities. The RED-Alert approach combines for the first time the CEP methodology with NLP/SMA and SNA applications in the context of social media data analytics, transforming (unstructured) social media data into (structured) events enhanced by semantic attributes. For example, a tweet will be an event consisting of content (expressed as NLP features e.g. concepts, sentiment, entities, etc.) and context (time and the author including SNA features e.g. number of followers, number of links, etc.). Turning unstructured social media data into structured events is key, as it allows the system to use (event) rules (event temporal logic, event logic patterns, even counting, absence of events) to infer insights or create alerts in real-time.The project impact is supported by the participation of Europol and specific dissemination activities around the World Counter-Terrorism Summit, organized by one of the partners. The total requested EC funding is 5M Euros and the project duration 36 months.