The Rosette entity linking endpoint links entity mentions in input text to entities in a knowledge base.
The following two passages illustrate entity linking where we have mentions of two different people with the same name.
|The first men to reach the moon – Mr. Armstrong and his co-pilot, Col. Edwin E. Aldrin, Jr. of the Air Force – brought their ship to rest on a level, rock-strewn plain near the southwestern shore of the arid Sea of Tranquility.
|The International Cycling Union announced Monday that it would not appeal the United States Anti-Doping Agency’s ruling to bar Mr. Armstrong for life from Olympic sports for doping and for playing an instrumental role in the team-organized doping on his Tour de France-winning cycling squads.
Unstructured text contains mentions of entities that occur in the world. Knowing the name of an entity, however, does not provide you with its identity.
LINKING THE IDENTITY OF NAME MENTIONS INVOLVES REASONING ABOUT:
- Names are inherently ambiguous: many entities share the same name and can be referenced in exactly the same way. For example, the two mentions of Mr. Armstrong above, although identical in surface form, refer to different entities.
- A single entity may also be referred to by a variety of name mentions. In moon.txt, the astronaut Buzz Aldrin is referred to as Col. Edwin E. Aldrin, Jr. In other texts he is likely to be referred to as Buzz Aldrin, Buzz, and/or Dr. Aldrin.