Seminar in Computational Linguistics

This talk describes the design and implementation of a very large scale system for internet monitoring. The Recorded Future system harvests and analyzes tens of millions of documents per day from all parts of the internet, turns unstructured text into structured data using natural language processing, and allows human analysts to efficiently monitor threats against their organizations. The system does semantic analysis in more than 10 different languages, and the talk will focus on linguistic challenges in making sense of the world using text in different languages from a wide range of sources.