Attensity Triples

Transforming Text into Actionable Intelligence

Today’s modern analyst monitors and processes vast amounts of open source, along with vetted and unvetted internal information to discover connections and extract relevant information to tell a story. These information sources can include HumINT reports, “suspicious activity reports”, witness statements, field notes, investigation reports, arrest reports, seizure documents, and open source intelligence.

To process all of this information and make it easier to search and understand, the old hybrid approach of extraction (whereby you used multiple noun extractors to do the same task to increase precision) is giving way to a “sequential” approach whereby you add additional tags from other engines that specialize in differing competencies (i.e. run for nouns, then run for links, etc.). Attensity Triples provide the next layer to add to your semantic ingestion process for advanced, comprehensive event and relationship extraction.

ATTENSITY TRIPLES - The next generation of natural language processing

Attensity’s “Triples” technology automatically finds the actions, behaviors, tactics, and relationships that tie things together - who did what to whom - or this thing is related to this other thing.

The Attensity parsing engine “understands” the grammar of language, making intelligent guesses about how to process text based on sophisticated linguistic principles, including not only part of speech, but syntactic function and other cues.

Attensity then uses this information to generate a principal form of relationship metadata called a Triple. Attensity’s engine is able to accommodate a wide variety of Triple types, which describes the people, places and things in a document and their relationships to each other.
 
Unlike structured entity or targeted fact extraction, the Triples engine finds, tags, and extracts a wide variety of information – not just those that conform to a known Entity Type or specified Event– and helps illuminate their relationships to each other. 

Attensity provides these technologies in an open, commercial-class server architecture framework, with robust APIs that make it simple to integrate Attensity into broader initiatives, including tying it together with other technologies such as entity extraction. Attensity also provides packaged applications for discovery and exploration that make it easy to explore trends, connections, and search for information.

By intelligently parsing text and its structure, Attensity’s unique Triples engine makes it possible for the first time to extract 100% of relevant information, while providing tools for culling that information down to precisely the information needed.