Intelligent Web search and autonomous agents responding to free text need detailed information about the entities mentioned and the relations they participate in. We have developed a number of tools to assign fine-grained entity types, link entities to Freebase, and extract relations between entities, including:
  • FIGER fine-grained entity recognizer assigns over 100 semantic types
  • VINICULUM identifies entity mentions in text and maps them to Freebase entities
  • MultiR learns relation extractors using distant supervision from Freebase
  • Information Omnivore learns relation extractors from a combination of crowd sourcing and distant supervision
  • A system to learn common-sense attributes objects (e.g. size & shape) through text mining & global inference.
  • NewsSpike, a system that automatically discovers and extracts events from news text.
We have also developed GoReCo , a new gold standard evaluation for relation extraction consisting of exhaustive annotations of the 128 documents from ACE 2004 newswire for 48 relations. The code and data are available at https://github.com/mitchellkoch/GoReCo. It requires the original ACE 2004 download from the Linguistics Data Consortium (https://catalog.ldc.upenn.edu/LDC2005T09).