Detecting Events and Emerging Threats in Cyberspace

Recorded Future builds a threat-oriented digital twin of the world, by combining information from human language text and technical data sources — at internet scale and in real time. This talk presents recent developments in text analytics and natural language processing (NLP), where we have introduced our third generation, transformer-based model for event detection and role assignment. We will also describe how NLP-derived intelligence is combined with technical intelligence to discover emerging threats, including attacker attribution and victim identification.