Knowledge Graph

Everyday, events are occurring all around us. The news content describing these events is filled with connections between people, places, and things; yet, typical news aggregators tend to ignore these connections in favor of simply displaying the news.

We believe that harvesting and presenting these connections helps to tell a better, more complete story - one where it's easy to discover multiple viewpoints on an event and understand the background context of what led to the participation of these people, places, and things in each event.

Black Cape Knowledge Graph (BCKG) is our flagship Artificial Intelligence (AI)/Machine Learning (ML) platform that automatically extracts, geolocates, connects, and organizes entities and insights found in unstructured news content. This allows users to identify non-obvious relationships and discover hidden insights, enabling actionable intelligence.

BCKG is a cloud-native system that orchestrates a number of custom machine-learning and automation micro-services.

Some of the foundational technologies in BCKG include:

• geo-registration and geolocation to make implied locations into explicit locations
• natural language processing (NLP) to add structure and extract meaning and from text
• named entity recognition (NER) to highlight places, people, and groups
• topic modeling to associate content with machine discovered categories
• advanced geospatial visualization to support Exploratory Data Analysis (EDA)
• fulltext search allowing users to dive into raw content.

Trying to discover trends, understand emerging threats, or see the big picture? Black Cape Knowledge Graph delivers global insights at the speed of mission.

Three Pillars

Collection

	Our system collects news and other forms of textual content from a vast array of sources on the web. It then performs data conditioning to clean up the raw information so that it is ready for machine learning.

Machine Learning

	Data from every source are enriched by machine learning algorithms that automatically extract entities of interest using Named Entity Recognition, geolocate content, and categorize data by topic.

Insight Discovery

	This machine learning derived intelligence from open sources can be navigated through our web application, which helps analysts discover insights based on location, topic, keyword, or relationship.