Surfacing unique facts for entities

Abstract

Systems and methods identify and provide interesting facts about an entity. An example method includes selecting documents associated with at least one unique fact trigger, the documents being from a document repository. The method also includes generating entity-sentence pairs from the documents and, for a first entity of the entities represented by the entity-sentence pairs, clustering the entity-sentence pairs for the first entity using salient terms occurring in the sentence. The method also includes determining a representative sentence for each of the clusters and providing at least one of the representative sentences in response to a query that identifies the first entity. Another example method includes determining that a query relates to an entity in a knowledge base, determining that the entity has an associated unique fact list, and providing at least one of the unique facts in the list in response to the query.

What is claimed is: 
     
         1 . A system comprising:
 a data store storing unique fact triggers; and   a query handling system that includes:
 at least one processor, and 
 a memory storing instructions that, when executed by the at least one processor, cause the query handling system to perform operations including:
 selecting, from a document repository, documents associated with at least one unique fact trigger stored in the data store, 
 generating entity-sentence pairs for an entity by:
 extracting sentences from the selected documents and, 
 for each sentence of at least some of the sentences: 
  identifying a reference to the entity in the sentence, 
  determining whether the sentence matches a structured fact pattern, and 
  in response to determining that the sentence does not match the structured fact pattern, storing the sentence and an identifier for the document from which the sentence was extracted as an entity-sentence pair for the entity, and 
 
 providing at least one stored sentence from the entity-sentence pairs in response to a query that identifies the entity. 
 
   
     
     
         2 . The system of  claim 1 , wherein the structured fact pattern includes a stored pattern. 
     
     
         3 . The system of  claim 1 , wherein the structured fact pattern includes a regular expression. 
     
     
         4 . The system of  claim 1 , wherein the instructions further cause the query handling system to generate the structured fact pattern based on facts stored in a knowledge base. 
     
     
         5 . The system of  claim 1 , wherein the structured fact pattern includes the entity followed by predetermined words. 
     
     
         6 . The system of  claim 1 , wherein the at least one unique fact trigger includes a whitelisted trigger phrase that identifies the documents as including unique facts. 
     
     
         7 . The system of  claim 1 , wherein the instructions further cause the query handling system to:
 generate at least one main unique fact cluster for the entity by clustering the entity-sentence pairs on salient terms; and   determine a representative sentence for the at least one main unique fact cluster,   wherein the at least one stored sentence is the representative sentence.   
     
     
         8 . The system of  claim 7 , wherein a sentence that matched the structured fact pattern is not included in the entity-sentence pairs that are clustered to generate the at least one main unique fact cluster. 
     
     
         9 . A non-transitory medium having code segments stored thereon, the code segments, when executed by a processor, cause a system to:
 select, from a document repository, documents associated with at least one unique fact trigger,   generate entity-sentence pairs for an entity by:
 extracting sentences from the selected documents and, 
 for each sentence of at least some of the sentences:
 identifying a reference to the entity in the sentence, 
 determining whether the sentence matches a structured fact pattern, and 
 in response to determining that the sentence does not match the structured fact pattern, storing the sentence and an identifier for the document from which the sentence was extracted as an entity-sentence pair for the entity, and 
 
   provide at least one stored sentence from the entity-sentence pairs in response to a query that identifies the entity.   
     
     
         10 . The non-transitory medium of  claim 9 , wherein the structured fact pattern includes a stored pattern. 
     
     
         11 . The non-transitory medium of  claim 9 , wherein the structured fact pattern includes a regular expression. 
     
     
         12 . The non-transitory medium of  claim 9 , wherein the code segments further cause the system to generate the structured fact pattern based on facts stored in a knowledge base. 
     
     
         13 . The non-transitory medium of  claim 9 , wherein the structured fact pattern includes the entity followed by predetermined words. 
     
     
         14 . The non-transitory medium of  claim 9 , wherein the code segments further cause the system to:
 generate at least one main unique fact cluster for the entity by clustering the entity-sentence pairs on salient terms; and   determine a representative sentence for the at least one main unique fact cluster, the at least one stored sentence being the representative sentence,   wherein a sentence that matched the structured fact pattern is not included in the entity-sentence pairs that are clustered to generate the at least one main unique fact cluster.   
     
     
         15 . A method comprising:
 selecting, from a document repository, documents associated with at least one unique fact trigger,   generating entity-sentence pairs for an entity by:
 extracting sentences from the selected documents and, 
 for each sentence of at least some of the sentences:
 identifying a reference to the entity in the sentence, 
 determining whether the sentence matches a structured fact pattern, and 
 in response to determining that the sentence does not match the structured fact pattern, storing the sentence and an identifier for the document from which the sentence was extracted as an entity-sentence pair for the entity, and 
 
   providing at least one stored sentence from the entity-sentence pairs in response to a query that identifies the entity.   
     
     
         16 . The method of  claim 15 , wherein the structured fact pattern includes a stored pattern. 
     
     
         17 . The method of  claim 15 , wherein the structured fact pattern includes a regular expression. 
     
     
         18 . The method of  claim 15 , wherein the method further includes generating the structured fact pattern based on facts stored in a knowledge base. 
     
     
         19 . The method of  claim 15 , wherein the structured fact pattern includes the entity followed by predetermined words. 
     
     
         20 . The method of  claim 15 , further comprising:
 generating at least one main unique fact cluster for the entity by clustering the entity-sentence pairs on salient terms; and   determining a representative sentence for the at least one main unique fact cluster,   wherein the at least one stored sentence is the representative sentence.