Designing an Intelligent System to Map Global Connections

This study develops a knowledge graph from the Common Crawl News Dataset to provide situational awareness and answer information requirements for national security leaders. We map dynamic global interactions between entities using news data collected over a defined period. We developed a data pipeline to extract semantic content from the Common Crawl News feed and filter it to content related to national security. We build a knowledge graph using national security-related content using Named Entity Recognition, Named Entity Disambiguation, and Named Entity Linking. After developing this intelligent system, we demonstrate its value in an excursion on a Russian language knowledge graph focused on the Russian invasion of Ukraine. The knowledge graphs developed in this study provide decision-makers an understanding of the entities and links in our highly interconnected world.