Web-scale entity relationship extraction
Techniques for displaying a relationship graph are described herein. In one example, a search term may be used to obtain a plurality of documents from a network, such as the Internet. A plurality of entities, and relationships between at least some of those entities, may be extracted from the documents. In an example user interface, representations of a plurality of entities may be displayed, such as by shapes (e.g., circles) labeled to identify people or organizations. Edges (e.g., lines) may be used to connect different representations of entities and to thereby indicate a relationship between the connected entities. In a particular example, input from movement of a cursor over an edge may result in display of a description of a relationship between the connected entities. In a further particular example, size of each entity may be related to a number of connections each has with others.
1. A system, comprising:
one or more processors;
memory, in communication with the one or more processors;
an output module, defined in the memory, executable by the one or more processors, and configured to graph relationship data, the graphed relationship data comprising:
representations of individual ones of a plurality of entities, the plurality of entities extracted from documents obtained based at least in part on a search term, each of the plurality of entities related to, but distinct from, the search term, wherein sizes of the representations of individual ones of the plurality of entities and the representation of the search term are not all equal;
edges between at least some of the representations of the plurality of entities, each edge forming a pair of two entities and indicating a relationship between entities of the pair; and
a representation of the search term having an edge connected to each of the representations of the plurality of entities.
2. The system as recited in claim 1 , wherein:
size of any particular representation is based at least in part on a number of edges connected to the particular representation.
3. The system as recited in claim 1 , wherein:
a size of the representation of the search term is larger than a size of any of the representations of the plurality of entities.
4. The system as recited in claim 1 , wherein the graphed relationship data additionally comprises:
a relationship display, configured to be visible in response to receipt of user input selecting an edge and configured to display an indication of a type of relationship associated with the selected edge.
5. The system as recited in claim 1 , additionally comprising:
a relationship clustering module, configured to use relationship data to group together representations of related entities.
6. The system as recited in claim 1 , additionally comprising a relation-tuple extraction module, configured to:
identify related-entity pairs;
detect keywords that indicate relationships of the identified related-entity pairs; and
graph relationship data based at least in part on the identified related-entity pairs and the indicated relationships.
7. A method of graphing relationship data, the method performed by one or more processors, the method comprising:
obtaining a plurality of documents based at least in part on a search term;
extracting a plurality of entities from the obtained documents;
generating for display representations of individual ones of the plurality of entities, each of the plurality of entities related to, but distinct from, the search term, wherein sizes of the representations of individual ones of the plurality of entities and the representation of the search term are not all equal;
generating for display edges between at least some of the representations of the plurality of entities, each edge forming a pair of two entities and indicating a relationship between entities of the pair; and
generating for display a representation of the search term having an edge connected to each of the representations of the plurality of entities.
8. The method as recited in claim 7 , wherein size of any particular representation is based at least in part on a number of edges connected to the particular representation.
9. The method as recited in claim 7 , wherein a size of the representation of the search term is larger than a size of any of the representations of the plurality of entities.
10. The method as recited in claim 7 , further comprising:
visually indicating a type of relationship associated with one of the edges in response to receipt of user input selecting the one edge.
11. The method as recited in claim 7 , further comprising:
grouping together representations of related entities based on relationship data.
12. The method as recited in claim 7 , further comprising:
identifying related-entity pairs;
detecting keywords that indicate relationships of the identified related-entity pairs; and
graphing relationship data based at least in part on the identified related-entity pairs and the indicated relationships.
13. A processing device, comprising:
one or more processors;
memory, in communication with the one or more processors;
an output module, defined in the memory, executable by the one or more processors, and configured to graph relationship data, the graphed relationship data comprising:
representations of individual ones of a plurality of entities, the plurality of entities extracted from documents obtained based at least in part on a search term, each of the plurality of entities related to, but distinct from, the search term, wherein sizes of the representations of individual ones of the plurality of entities and the representation of the search term are not all equal;
edges between at least some of the representations of the plurality of entities, each edge forming a pair of two entities and indicating a relationship between entities of the pair; and
a representation of the search term having an edge connected to each of the representations of the plurality of entities.
14. The processing device as recited in claim 13 , wherein:
size of any particular representation is based at least in part on a number of edges connected to the particular representation.
15. The processing device as recited in claim 13 , wherein:
a size of the representation of the search term is larger than a size of any of the representations of the plurality of entities.
16. The processing device as recited in claim 13 , wherein the graphed relationship data additionally comprises:
a relationship display, configured to be visible in response to receipt of user input selecting an edge and configured to display an indication of a type of relationship associated with the selected edge.
17. The processing device as recited in claim 13 , additionally comprising:
a relationship clustering module, configured to use relationship data to group together representations of related entities.
18. The processing device as recited in claim 13 , additionally comprising a relation-tuple extraction module, configured to:
identify at least on related-entity pair;
detect at least one keyword that indicates a relationship of the identified related-entity pair; and
graph relationship data based at least in part on the identified related-entity pair and the indicated relationship.
19. The processing device as recited in claim 18 , wherein the relation-tuple extraction module is further configured to:
generate for display the detected at least one keyword that indicates a relationship of the identified related-entity pair.
20. The system as recited in claim 6 , further comprising:
generating for display the detected keywords that indicate relationships of the identified related-entity pairs.