Veracity Enriched Event Extraction
Ning Yu, Graham Horwood · 2018
To support intelligence analysis, lots of effort goes into the development of information extraction (IE) systems to ingest massive collections of both traditional sensor data and streaming data from both public domain and classified sources, classify them, distill evidence from them, and store that evidence as event and entity tuples in a knowledge base. This paper focuses on an unmet challenge of this effort: determining the veracity of extracted information, in order to help analysts make balanced, rational judgments.