US20180046717A1

US20180046717A1 - Related entities

Info

Publication number: US20180046717A1
Application number: US15/798,175
Authority: US
Inventors: Peter Jin Hong; Pravir K. Gupta; Nathaniel J. Gaylinn; Ramakrishnan Kazhiyur-Mannar; Kavi J. Goel; Omer Bar-or; Jack W. Menzel; Christina R. Dhanaraj; Jared L. Levy; Shashidhar A. Thakur; Grace Chung; Benson Tsai
Original assignee: Google LLC
Current assignee: Google LLC
Priority date: 2012-02-22
Filing date: 2017-10-30
Publication date: 2018-02-15
Also published as: WO2013126808A1; CN108388582B; KR101994987B1; KR20140128443A; CN108388582A; US9830390B2; US9275152B2; AU2013222184B2; CN104428767B; CN118467593A; US20130238594A1; AU2013222184A1; CN104428767A; US20230205828A1; US20160179958A1

Abstract

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for identifying entities that are related to an entity to which a search query is directed. One of the methods includes receiving a search query, wherein the search query has been determined to relate to a first entity of a first entity type, and wherein one or more entities of a second entity type have a relationship with the first entity; receiving search results for the search query; determining that a count of search results identifying a resource containing a reference to the first entity satisfies a first threshold value; determining that a count of search results identifying a resource having the second entity type as a relevant entity type satisfies a second threshold value; and transmitting information identifying the one or more entities of the second entity type as part of the response to the search query.

Description

CROSS REFERENCE TO RELATED APPLICATIONS

This application is a continuation of application Ser. No. 15/055,427, filed on Feb. 26, 2016, which is a continuation of application Ser. No. 13/774,896, filed on Feb. 22, 2013, now U.S. Pat. No. 9,275,152, entitled “Related Entities,” which claims the benefit under 35 U.S.C. §119(e) of U.S. patent application Ser. No. 61/601,975, filed Feb. 22, 2012, entitled “Related Entities,” U.S. patent application Ser. No. 61/647,977, filed May 16, 2012, entitled “Related Entities,” and U.S. patent application Ser. No. 61/660,637, filed Jun. 15, 2012, entitled “Related Entities,” all of which are incorporated by reference herein in their entirety.

BACKGROUND

This specification relates to Internet search systems.
Internet search engines aim to identify Internet resources, e.g., web pages, images, text documents, or multimedia content, that are relevant to a user's needs and to present information about the resources in a manner that is most useful to the user. Internet search engines return a set of search results in response to a user submitted query. Internet search engines generally include one or more services that can classify particular received queries. Such services may include services that classify queries as one or more of: a query that is pornographic, i.e., is seeking pornographic results or for which a large number of search results identifying resources that have been classified as pornographic are returned; a query that is navigational to a particular resource, i.e., is seeking that particular resource; a query that is a local query, i.e., is seeking information about a business located near the user; or a query that is seeking a particular item of information, e.g., is looking for an item of information that is an answer to a question posed in the query.

SUMMARY

This specification describes technologies relating to identifying entities that are related to an entity to which a search query is directed.
In general, one innovative aspect of the subject matter described in this specification can be embodied in methods that include the actions of receiving a first search query from a user device, wherein the first search query has been determined to relate to a first entity of a first entity type, and wherein one or more entities of a second entity type have a predetermined relationship with the first entity; receiving search results for the first search query provided by a search engine, wherein each of the search results identifies a respective resource; determining that a count of search results identifying a resource containing a reference to the first entity satisfies a first threshold value; determining that a count of search results identifying a resource having the second entity type as a relevant entity type satisfies a second threshold value; and transmitting information identifying the one or more entities of the second entity type to the user device as part of the response to the first search query.
Other embodiments of this aspect include corresponding computer systems, apparatus, and computer programs recorded on one or more computer storage devices, each configured to perform the actions of the methods. A system of one or more computers can be configured to perform particular operations or actions by virtue of having software, firmware, hardware, or a combination of them installed on the system that in operation causes or cause the system to perform the actions. One or more computer programs can be configured to perform particular operations or actions by virtue of including instructions that, when executed by data processing apparatus, cause the apparatus to perform the actions.
These and other embodiments can optionally include one or more of the following features. Each of the search results can include a respective title and a respective snippet of text extracted from the respective resource identified by the search result, and determining that the count of search results identifying a resource containing a reference to the first entity satisfies the first threshold value can include: determining that a count of search results that include a reference to the first entity in the respective title or the respective snippet of text included in the search result satisfies the first threshold value.
The information identifying the one or more entities of the second entity type can include a respective image corresponding to each of the one or more entities, and the method can further include: obtaining, for each of the one or more entities of the second entity type, the respective image corresponding to the entity from an image search engine in response to a search query derived from the name of the entity.
Obtaining the image for a particular entity of the one or more entities of the second type can include: determining that a particular search query including the name of the particular entity is ambiguous, comprising determining, from search results provided for the particular search query by the search engine, that the particular search query either does not relate to any entity in an index that maps each of a plurality of resources to a specific entity of a specific type or relates to more than one entity in the index; generating a second search query that includes the name of the particular entity and at least one of: a reference to the first entity of the first entity type or a reference to the second entity type; obtaining image search results for the second search query from the image search engine; and selecting the image for the particular entity from images identified by the image search results for the second search query.
The method can further include: determining that the second search query is not ambiguous.
Selecting the image for the particular entity from images identified by the image search results can include: selecting the image based at least in part on an aspect ratio of the image.
The method can further include: determining from the search results for the first search query that the first search query relates to the first entity of the first entity type, including: determining, using an index that maps each of a plurality of resources to a specific entity of a specific type, that a count of search results that identify a resource that is mapped to the first entity exceeds a third threshold value.
The method can further include: obtaining data that classifies the search query as not being any of a query that is pornographic, a query that is navigational to a particular resource, a local query, or a query that is seeking a particular item of information.
The information identifying the one or more entities can include information identifying a name of each of the one or more entities and the second entity type.
The method can further include: determining a respective ranking score for each of the entities of the second entity type; and ordering the entities of the second entity type according to the ranking scores.
The ranking score for a particular entity of the second entity type can be based at least in part on how frequently a recognized reference to the particular entity co-occurs with a recognized reference to the first entity in resources indexed by an indexing engine.
The ranking score for a particular entity of the second entity type can be based at least in part on how frequently the particular entity is searched for by users after submitting a search query directed to the first entity.
The ranking score for a particular entity of the second entity type can be based at least in part on a global popularity of the particular entity.
The ranking score for a particular entity of the second entity type can be based at least in part on how frequently a recognized reference to the particular entity co-occurs in a same previously submitted search query as a recognized reference to the first entity.
The method can further include: accessing data that indicates that two or more of the entities of the second entity type are members of a set of entities that has a specified order; and adjusting the ordering of the two or more entities of the second entity type to match the specified order.
The method can further include: accessing data that indicates that two or more of the entities of the second entity type are better known as being part of a broader entity; and replacing the two or more entities of the second entity type with the broader entity in the ordering of the entities of the second entity type.
The subject matter described in this specification can be implemented in particular embodiments so as to realize one or more of the following advantages. Users can easily view information about entities that have a particular relationship with an entity to which their search query is directed. In particular, users can easily obtain information identifying entities that are related to an entity of interest by submitting a search query that identifies the entity of interest to a search engine. Additionally, a user can easily submit another search query to obtain more information about the related entities. A user can learn about an entity to which their search query is directed by viewing information about the entities that are related to the entity. For example, a user can learn that a particular person is an author by submitting a query that includes the name of the particular person to a search engine and being presented with information identifying books written by the particular person. Additionally, questions that users will likely have about an entity after submitting a query directed to the entity can be predicted and information about entities that are answers to those questions can be provided to the user as part of a response to the query, e.g., without the user having to submit another search query or navigate to another resource to seek out the answers. For example, users submitting queries directed to an author may frequently subsequently look for information about particular books written by the author. Information identifying those particular books can be presented to the user as part of a response to the search query without the user having to submit additional queries or navigate to resources identified by the search results for the query. By submitting a search query directed to an entity of a particular type, a user can easily obtain information about, and submit queries directed to, other entities of the particular type that relate to the entity.
The details of one or more embodiments of the subject matter of this specification are set forth in the accompanying drawings and the description below. Other features, aspects, and advantages of the subject matter will become apparent from the description, the drawings, and the claims.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 shows an example search results page.

FIG. 2 is a block diagram of an example search system.

FIG. 3 is a flow diagram of an example process for identifying one or more related entities to be identified as part of a response to a search query.

FIG. 4 is a flow diagram of an example process for determining whether a search query is directed to a particular entity.

FIG. 5 is a flow diagram of an example process for determining whether related entities should be identified in a response to a particular search query.

FIG. 6 is a flow diagram of an example process for building indices to be used in selecting relevant entities.

FIG. 7 is a flow diagram of an example process for ordering related entities.

FIG. 8 is a flow diagram of an example process for mapping an entity to related entities of the same type.

Like reference numbers and designations in the various drawings indicate like elements.

DETAILED DESCRIPTION

FIG. 1 shows an example search results page 100 for a search query 102 “roald dahl.” The search results page 100 includes two search results 104 and 106 and names of related entities 108. The search results 104 and 106 and the names of related entities 108 are generated by a search system in response to the search query 102. The search results 104 and 106 each identify a respective resource and include respective titles 120 and 122 and respective text snippets 124 and 126 that are extracted from the resources identified by the search results. The search system generates the search results 104 and 106 using conventional search techniques.
The search system classifies the search query 102 as being directed to a particular entity, i.e., the author Roald Dahl, and returns names of related entities 108 that have a predetermined relationship with the particular entity, i.e., that are books authored by Roald Dahl, for presentation in the search results page 100. In the illustrated example, the search system may classify the search query 102 as being directed to the author Roald Dahl because one or both of the search results 104 and 106 identify a resource that has been determined to be an authoritative resource for the author Roald Dahl. For example, one or both of the resources identified by the search results 104 and 106, i.e., the official web site of the author Roald Dahl and the Wikipedia page for the author Roald Dahl, may have been determined to be an authoritative resource for the author Roald Dahl. In response to the search query 102, the search system selects the names of the related entities that are to be returned, e.g., using an index that stores data identifying entities that have a relationship with the author Roald Dahl.
Each name in the names of related entities 108, e.g., “Charlie and the Chocolate Factory” and “James and the Giant Peach,” is presented in the form of a link by which a user can obtain search results for a query derived from the name of the related entity. For example, the query derived from the name of the related entity can include one or more of the name of the related entity, e.g., “Charlie and the Chocolate Factory,” the text of the search query 102, e.g., “Roald Dahl,” and the name of the entity type to which the name belongs, e.g., “book.” Each name is presented with an image 114 that corresponds to the name, e.g., an image of the front cover of the book. Each of the images may also be presented in the form of a link by which a user can get search results for the query derived from the name of the related entity to which the image corresponds. In some implementations, in response to a user hovering a cursor of an input device over one of the images 114, additional information about the related entity to which the image corresponds is displayed. For example, in response to a user hovering over one of the images 114, the year that the book to which that image corresponds was published could be displayed to the user.
The search results page also includes information 110 identifying the type of the related entities, in this case “books,” named in the search results page 100 and information 112 identifying the entity to which the search query 102 was determined to relate, in this case, “Roald Dahl.”
FIG. 2 is a block diagram of an example search system 214. The search system 214 is an example of an information retrieval system implemented as computer programs on one or more computers in one or more locations, in which the systems, components, and techniques described below can be implemented.
A user 202 can interact with the search system 214 through a user device 204. For example, the user device 204 can be a computer coupled to the search system 214 through a data communication network 212, e.g., local area network (LAN) or wide area network (WAN), e.g., the Internet, or a combination of networks. In some cases, the search system 214 can be implemented on the user device 204, for example, if a user installs an application that performs searches on the user device 204. The user device 204 will generally include a memory, e.g., a random access memory (RAM) 206, for storing instructions and data and a processor 208 for executing stored instructions. The memory can include both read only and writable memory.
A user 202 can use the user device 204 to submit a query 210 to a search system 214. A search engine 230 within the search system 214 performs a search to identify resources matching the query. When the user 202 submits a query 210, the query 210 may be transmitted through the network 212 to the search system 214. The search system 214 includes an index database 222 and the search engine 230. The search system 214 responds to the query 210 by generating search results 228, which are transmitted through the network to the user device 204 for presentation to the user 202, e.g., as a search results web page to be displayed by a web browser running on the user device 204.
In this specification, the term “database” will be used broadly to refer to any collection of data: the data does not need to be structured in any particular way, or structured at all, and it can be stored on storage devices in one or more locations. Thus, for example, the index database 222 can include multiple collections of data, each of which may be organized and accessed differently. Similarly, in this specification the term “engine” will be used broadly to refer to a software based system or subsystem that can perform one or more specific functions. Generally, an engine will be implemented as one or more software modules or components, installed on one or more computers in one or more locations. In some cases, one or more computers will be dedicated to a particular engine; in other cases, multiple engines can be installed and running on the same computer or computers.
When the query 210 is received by the search engine 230, the search engine 230 identifies resources that satisfy the query 210. The search engine 230 will generally include an indexing engine 220 that indexes resources, an index database 222 that stores the index information, and a ranking engine 252 or other software that generates scores for the resources that satisfy the query 210 and that ranks the resources according to their respective scores.
The search system 214 also includes or can communicate with a related entities engine 240 that determines, from the search results 228, whether the query 210 is directed to a particular entity, i.e., whether the query should be classified as being directed to the particular entity. Determining whether a query is directed to a particular entity will be described in more detail below with reference to FIG. 4. If the query is directed to a particular entity, the related entities engine 240 identifies entities that have a predetermined relationship to the particular entity. Identifying entities that have a predetermined relationship to the particular entity will be described in more detail below with reference to FIG. 3. Once the related entities are identified, the search system 214 can transmit information identifying the related entities to the user device 204 as part of a response to the search query 210, e.g., with the search results 228 or in place of the search results 228.
In order to determine whether the query is directed to a particular entity and to identify the entities that are related to the particular entity, the related entities engine 240 can communicate with a related entities index database 250. The related entities index database 250 includes two indices, one that maps each entity of a group of entities to one or more related entities and identifies a relationship between the entity and the one or more related entities, and another that maps, to each entity of the group of entities, one or more authoritative resources for the entity. Building these indices is described below with reference to FIG. 6.
FIG. 3 is a flow diagram of an example process 300 for identifying one or more related entities to be identified as part of a response to a search query. For convenience, the process 300 will be described as being performed by a system of one or more computers located in one or more locations. For example, a search system, e.g., the search system 214 of FIG. 2, appropriately programmed, can perform the process 300.
The system receives a search query from a user device (step 302) and obtains search results for the search query from a search engine (step 304).
The system determines that the search query is about a principal entity (step 306). The system does so by analyzing the search results obtained for the search query. Determining that a search query is directed to an entity from an analysis of search results for the search query is described in more detail below with reference to FIG. 4.
The system determines that one or more other entities have a relationship with the principal entity (step 308). The system determines that the one or more other entities have a relationship with the principal entity by accessing an index, e.g., an index in the related entities index database 250 of FIG. 2, that maps entities to other entities that have a pre-determined relationship with the entity. The index also identifies the type of the one or more related entities and, optionally, the relationship between the entity to which the search query is directed and the related entities. The related entities can be of the same type as the entity or of a different type, depending on the pre-determined relationship. For example, for a particular movie, the related entities may be actors who starred in the movie, the producer of the movie, the director of the movie, and so on. Alternatively, the related entities may be other movies that were produced by the same producer, that were directed by the same director, or that share one or more actors with the movie. The index can also identify the type of the related entities, e.g., “actor,” and the relationship between the related entities and the principal entity, e.g., “acted in.”
The system determines that related entities should be identified in a response to the search query (step 310). Determining whether related entities should be identified in a response to a search query will be described below with reference to FIG. 5.
The system transmits information identifying the related entities as part of a response to the search query (step 312). The information identifying the related entities can be, e.g., included in a search results web page and transmitted to the user device for presentation to a user. The information identifying the related entities can include the name of each entity presented to the user in the form of a link that, when selected by a user, submits a search query derived from the name of the entity to a search engine, e.g., the search engine 230 of FIG. 2. The search query derived from the name of the entity can include only the name of the related entity or the name and, e.g., one or more of the type of the related entity and the name of the principal entity, depending on whether the query consisting of only the name of the entity is determined to be ambiguous.
A search query can be determined to be ambiguous if the search results for the search query indicate that the search query is not directed to any one particular entity. That is, if, after an analysis of the search results for the search query, it is determined that the search query is not directed to any entity or is directed to more than one entity, the search query is determined to be ambiguous. Determining that a search query is directed to an entity from an analysis of search results for the search query is described in more detail below with reference to FIG. 4. If the search results are ambiguous, the link, when selected by the user, submits a query that includes the name of the entity and one or more of the type of the related entity and the name of the principal entity.
The information identifying the related entities can optionally include, instead of or in addition to the names of the related entities, an image that corresponds to the related entity. Like the names of the entities, each image can be presented in the form of a link, that when selected by a user, submits a search query derived from the name of the entity to which the image corresponds to the search engine. The system can obtain the corresponding image for a related entity by submitting a search query derived from the name of the entity to an image search engine and selecting an image from the images identified by image search results for the search query, e.g., by selecting the image identified by a highest-ranked image search result. The query derived from the name of the entity can be a query that has been determined to not be ambiguous, e.g., using the technique described above. Further, in some implementations, the system prefers images that have particular predetermined properties, e.g., that have an aspect ratio that falls within a predetermined range of aspect ratios. That is, the system can select only images having properties that match the predetermined properties, provided that the image search result that identifies the image has a ranking that satisfies a predetermined threshold value or has a score that satisfies a predetermined threshold score.
Alternatively, the system can maintain an index that provides images for entities. The information identifying the related entities can also optionally include metadata that is associated with the related entities in the index. For example, for an entity of the type “book,” the metadata can identify the year the book was originally published.
FIG. 4 is a flow diagram of an example process 400 for determining whether a search query is directed to a particular entity. For convenience, the process 400 will be described as being performed by a system of one or more computers located in one or more locations. For example, a search system, e.g., the search system 214 of FIG. 2, appropriately programmed, can perform the process 400.
The system obtains search results for a search query from a search engine (step 402).
The system determines whether a sufficient number of resources identified by the search results are authoritative resources for a particular entity (step 404). For example, the system may determine whether a count of resources that are authoritative resources for the particular entity exceeds a threshold value. In determining which resources are authoritative resources, the system may optionally consider resources identified by a specified number of highest-ranked search results or search results having a score assigned to them by the search engine that exceeds a threshold value.
The system determines whether a resource identified by a search result is an authoritative resource for any entities by accessing an index, e.g., an index included in the related entities database 250 of FIG. 2, that maps authoritative resources to entities. An authoritative resource for an entity is a resource whose occurrence in search results has been determined to be a strong indicator that the search query is directed to the entity. Determining which entities are authoritative and building the index is described below with reference to FIG. 6.
If the number of resources that are authoritative resources for the particular entity is sufficient, the system classifies the search query as being directed to the particular entity (step 406).
If an insufficient number of resources are authoritative resources for the particular entity, the system classifies the search query as not being directed to the particular entity (step 408).
FIG. 5 is a flow diagram of an example process 500 for determining whether related entities should be identified in a response to a particular search query. For convenience, the process 500 will be described as being performed by a system of one or more computers located in one or more locations. For example, a search system, e.g., the search system 214 of FIG. 2, appropriately programmed, can perform the process 500.
The system obtains search results for a search query from a search engine (step 502). The search query is a query that has been determined to be directed to a particular entity that is related to one or more entities of a particular entity type, e.g., by performing the process described above with reference to FIG. 4. The system also obtains data that, for at least some of the resources identified by the search results, identifies one or more entity types that are relevant to each of the resources. For example, the data may identify the entity types “car” and “movie” as being relevant to one resource identified by one search result, the entity types “actor” and “author” as being relevant to another resource identified by another search result, and so on.
The system determines whether more than a threshold number of resources identified by the search results contain references to the particular entity (step 504). In determining which resources contain references to the particular entity, the system may optionally consider only resources identified by a pre-determined number of highest-ranked search results or by search results having a score assigned to them by the search engine that exceeds a threshold value. For example, the system can determine whether the proportion of highest-ranking search results that include at least one recognized reference to the particular entity, e.g., a known name for the particular entity, in the title or the text snippet extracted from the resource identified by the search result exceeds a threshold value. For example, the system may determine whether two of the top five highest-ranking search results, three of the top ten highest-ranking search results, or thirty of the top one hundred highest-ranking search results include a recognized reference to the particular entity.
If an insufficient number of resources identified by the search results contain references to the particular entity, the system determines that related entities should not be identified in a response to the search query (step 510).
If a sufficient number of resources identified by the search results contain references to the particular entity, the system determines whether the entity type of the related entities is relevant to more than a threshold number of resources identified by the search results (step 506) using the data about relevant entity types obtained from the search engine. In determining which resources have relevant entity types that match the type of the related entities, the system may optionally consider only resources identified by a specified number of highest-ranked search results or search results having a score assigned to them by the search engine that exceeds a threshold value. For example, for a search query determined to be directed to an entity of the type “author,” one or more entities of the type “book” may have a relationship to the author. The system can then check whether a sufficient number of resources identified by a predetermined number of highest-ranked search results have a relevant entity type of “book” before determining to return names of the books that relate to the author as part of a response to the search query.
If a sufficient number of resources identified by the search results have a relevant entity type that matches the type of the related entities, the system determines that related entities should be identified in a response to the search query (step 508).
If an insufficient number of resources identified by the search results have a relevant entity type that matches the type of the related entities, the system determines that related entities should not be identified in a response to the search query (step 510).
In some implementations, the system has access to information that characterizes the search query as either belonging to or not belonging to one or more special classes. The system can obtain the information from, e.g., one or more services included in the search engine from which the search results are obtained. For example, the system can obtain data that characterizes the search query as potentially being one or more of pornographic, navigational, or local, or that characterizes the search query as a query that is seeking a particular item of information. In such implementations, even if the other criteria for identifying related entities in a response to the search query are satisfied, the system can refrain from returning information about related entities as part of a response to the search query if the information indicates that the search query belongs to one or more of the special cases, e.g., because returning names of related entities could either be inappropriate or undesirable to the user.
FIG. 6 is a flow diagram of an example process 600 for building indices to be used in selecting relevant entities. For convenience, the process 600 will be described as being performed by a system of one or more computers located in one or more locations. For example, a search system, e.g., the search system 214 of FIG. 2, appropriately programmed, can perform the process 600.
The process 600 can be performed for multiple entities of multiple types.
The system obtains data identifying a particular entity of a particular type (step 602). The obtained data includes one or more names for the particular entity and identifies the type of the entity. The obtained data can also optionally identify one or more resources associated with the entity. For example, for the author J. R. R. Tolkien, the obtained data can include the name of the entity, “J. R. R Tolkien,” and identify the type of the entity, “author.” The data can also optionally include the resource locators for one or more resources associated with the author J. R. R. Tolkien, e.g., an online encyclopedia page directed to the author or an official webpage of the author. In some circumstances, the obtained data may identify more than one type for the particular entity. For example, the data could characterize Will Smith as being of the type “actor” and of the type “musician.”
The system can obtain data identifying entities from a variety of sources. For example, one source may be an online database of structured data that includes nodes that represent entities and identifies the type of each entity represented by a node. An example of an online database of structured data that exists is the FREEBASE database that is accessible on the Internet at http://www.freebase.com. Other sources can include online encyclopedias having pages directed to each of a group of entities and websites directed to particular types of entities, e.g., a website that includes resources directed to movies and movie-related entities, e.g., actors, directors, and producers.
The system identifies one or more resources as authoritative resources for the particular entity (step 604). If the obtained data identifying the entity also identifies resources associated with the entity, the system can select those resources as authoritative resources for the entity. For example, an online encyclopedia page for an entity may contain links to other resources that relate to the entity, e.g., the official website of the entity. Additionally, resources relating to the entity may be associated with the node representing a particular entity in an online database of structured data, e.g., by way of a link to another node or by way of a link representing a property of the entity.
If the data does not identify any associated resources for an entity, or to augment the resources identified for the entity by the data, the system can submit a search query derived from the name or names of the entity to a search engine and obtain search results for the search query. The system can then select as authoritative resources for the entity particular resources from the resources identified by the obtained search results. For example, the system can select a specified number of highest-scoring search results or each search result having a score that exceeds a threshold score.
The system associates the authoritative resources with the particular entity in an index (step 606). For example, for each authoritative resource for a particular entity, the system can generate a mapping from, i.e., data defining an association between, the resource locator of the resource to the particular entity. The index can be, e.g., one of the indices included in the related entities index database 250.
If the obtained data identifies more than one entity type for the particular entity, the system can select one of the types as the entity type for the entry for the particular entity in the index. For example, the system can obtain search results for a search query that is derived from the name of the particular entity. For each of a pre-determined number of highest-ranked search results, the system can obtain data identifying the relevant entity types for the resource identified by the search result, e.g., from a service that identifies the entity types that are relevant to resources. The system can then select one of the entity types for the particular entity based on the relevant entity types, e.g., select the entity type that is relevant to the most resources as the entity type for the particular entity.
Alternatively, the system can generate an entry in the index for each of the types for that entity, e.g., one entry for Will Smith the “actor” and one for Will Smith the “musician,” and associate one or more of the identified authoritative resources with the respective entry for each of the types. The system can select which of the authoritative resources are associated with each of the entries for the particular entity by obtaining data that, for each of the authoritative resources, identifies the relevant entity types for that authoritative resource. For each of the index entries for the particular entity, the system can then associate the authoritative resources having a related entity type that matches the entity type for that entry.
For example, in the case where Will Smith is identified as both an “actor” and a “musician,” the system may identify two authoritative resources for Will Smith: a web page directed to Will Smith from a web site about actors and movies and a web page directed to Will Smith from a social media web site for musicians. The system can obtain data identifying the most relevant entity types for each of the authoritative resources and associate the page from the web site about actors and movies with entity “Will Smith” having the type “actor”, e.g., because the obtained data indicated that the entity type “actor” is relevant to the page. The system can also associate the page from the social media web site with the entity “Will Smith” having the entity type “musician”, e.g., because the obtained data indicated that the entity type “musician” is relevant to the page.
The system obtains data identifying other entities having a relationship with the particular entity (step 608). If the particular entity has more than one type, the system obtains data identifying other entities for each of the types. The desired type of relationship can be predetermined, e.g., specified by a system administrator. That is, the system administrator can specify that, for entities of the type “author,” the obtained data should identify entities of the type “book” that were written by the author.
The system can obtain the data identifying the other entities and their relationship with the entity from, e.g., the same source from which the data identifying the entity was obtained or from additional sources. For example, the system can query the online database of structured data to obtain data identifying entities that relate to the entity. For example, for an entity representing an actor, the system can query the online database of structured data to obtain data identifying one or more movies in which the actor acted. Additionally, a page in an online encyclopedia directed to the entity may identify other entities that relate to the entity. For example, an online encyclopedia page for a musical artist can identify albums by the musical artist, popular songs of the musical artist, and so on.
The system orders the entities having a predetermined relationship with the particular entity (step 610). In some implementations, the system orders the related entities according to the order in which they were received by the system from a data source. However, in other implementations, the system reorders the related entities, e.g., to account for users submitting search queries that are directed to the particular entity being more interested in information about particular ones of the entities that are related to the particular entity. Ordering related entities is described below with reference to FIG. 7.
The system associates the related entities with the particular entity in an index (step 612). For example, the system can generate a mapping from the particular entity to each of the related entities. The mapping can also identify the type of the related entities and, optionally, the nature of the relationship between the particular entity and the related entities. The index can be, e.g., one of the indices included in the related entities index database 250. The system can generate the mapping in such a manner that, when the related entities are selected for presentation to a user in response to a search query directed to the particular entity, the related entities are presented in an order that matches the order generated by the system.
In some implementations, the system also obtains additional information about each of the related entities from one of the data sources and stores the additional information in an index, e.g., as metadata associated with the related entity, for later display to a user, e.g., as described above with reference to FIG. 1. The data obtained may depend on the type of the related entities and on the relationship between the related entities and the particular entity. For example, if the particular entity is an entity of the type “movie” and the related entities are entities of the type “actor” that acted in the movie, the additional information may include the name of the character played by each of the related entities. However, if the particular entity is an entity of the type “author” and the related entities are entities of the type “book” that are written by the author, the additional information may include the year that each of the related entities was first published. Thus, the additional information may be a value of a property that the related entity possesses by virtue of that entity's relationship with the particular entity.
FIG. 7 is a flow diagram of an example process 700 for ordering related entities. For convenience, the process 700 will be described as being performed by a system of one or more computers located in one or more locations. For example, a search system, e.g., the search system 214 of FIG. 2, appropriately programmed, can perform the process 500.
The system obtains data identifying related entities for a particular entity (step 702).
The system determines a ranking score for each of the related entities (step 704). The system determines the ranking scores by aggregating two or more of a variety of factor-specific scores, with each factor-specific score being computed based on a respective factor.
For example, the system may generate a score for each of the related entities based on how frequently a recognized reference to each related entity, e.g., a known name of the related entity, co-occurs with a recognized reference to the particular entity in resources indexed by an indexing engine, e.g., the indexing engine 220 of FIG. 2, e.g., so that related entities that co-occur more frequently with the particular entity have higher scores than related entities that co-occur less frequently with the particular entity.
As another example, the system may generate a score for each of the related entities based on how frequently each related entity is searched for by users after submitting a query directed to the particular entity. For example, the system may obtain data that identifies, for each of the related entities, how frequently users submit search queries that include a recognized reference to the related entity after submitting an initial search query that includes a recognized reference to the particular entity, e.g., immediately following submitting the initial search query. A later search query can be considered to be submitted immediately following an earlier search query if it is submitted within a pre-determined window of time of submitting the earlier search query and if, when the later search query is submitted, no additional search queries have been submitted by the user after the earlier search query was submitted. The system can generate the scores so that related entities that are searched for more frequently after submitting a query directed to the particular entity have higher scores than related entities that are searched for less frequently after submitting a query directed to the particular entity.
As another example, the system may generate a score for each of the related entities based on the global popularity of each of the related entities. The global popularity of a related entity can be measured based on how frequently a recognized reference to the related entity appears in resources indexed by the indexing engine, how frequently a recognized reference to the related entity appears in previously submitted search queries, i.e., search queries stored in a record of queries that have been submitted to a search engine by users, or both. Alternatively, the global popularity of a related entity can be based at least in part on how frequently authoritative resources for the related entity are identified in search results for previously submitted queries. For example, the global popularity may be based at least in part on the number of previously submitted queries for which an authoritative resource for the related entity is identified by one of a pre-determined number of highest-ranked search results. The system can generate the scores so that a related entity that has a higher global popularity than another entity will have a higher score than a related entity that has a lower global popularity.
As another example, the system may generate a score for each of the related entities based on how frequently a recognized reference to the related entity co-occurs in the same previously submitted search query as a recognized reference to the particular entity, e.g., so that related entities that co-occur more frequently with the particular entity in previously submitted search queries have higher scores than related entities that co-occur less frequently with the particular entity.
The system can then generate a ranking score for each related entity based on the factor-specific scores for the entity. The system can generate the ranking score for a given entity by, e.g., normalizing the factor-specific scores for the entity and then computing an average of the normalized scores, computing a sum of the normalized scores, computing a product of the normalized scores, or otherwise aggregating the normalized scores.
The system orders the related entities according to their ranking scores (step 706). In some implementations, the system can make further adjustments to the ordering of the related entities after ordering the entities according to their ranking scores. In particular, the system may have access to data that specifies an ordering for one or more sets of entities. For example, the data may specify that movies in a particular movie trilogy be ordered by their release date or that Presidents of the United States be ordered by the date their term began or ended. If the data indicates two or more of the related entities are members of a set of entities that has a specified order, the system can adjust the order of the members of the set to match the specified order. In some implementations, if the members of the set of entities are separated by other entities in the order of related entities, the system reorders the members of the set to match the specified order and places the reordered members of the set together in the order of related entities, e.g., beginning at the position of the highest-ranked entity in the set or at the average position of the entities in the set. In some other implementations, if the members of the set of entities are separated by other entities in the order of related entities the system refrains from adjusting the order of the members of the set to match the specified order.
As another example, the system may have access to data that identifies sets of entities that are better known as being part of a broader entity. For example, the data may indicate that the individual books that make up a popular book series are better known as being part of the series rather than as individual books. If the data indicates that some or all of the related entities are included in a broader entity, the system can replace those related entities in the order with the broader entity. For example, the related entities for an author may include books written by that author. If the author has written a well-known trilogy of books and those books are included in the related entities for the author, the system may replace the books in that trilogy with a single entity that represents the trilogy. In some implementations, if the entities in the set of entities that is better known as being part of a broader entity are separated by other entities in the order of the related entities, the system places the broader entity in the order at the position of the highest-ranking entity in the set. Alternatively, the system can place the broader entity in the order at a position that is an average of the positions of the entities in the set. In some other implementations, the system can refrain from replacing the entities in the set with the broader entity if the entities in the set are separated by other entities in the order of the related entities.
While the processes 600 and 700 describe obtaining data that identifies entities that relate to the particular entity and then ordering the related entities, in some circumstances, e.g., when the related entities are of the same type as the particular entity, the system may refine candidate entities identified by the obtained data in order to identify the entities that relate to the particular entity. For example, for a particular entity of the type “person,” the system may obtain data identifying a large number of other entities of the type “person” that are represented by nodes in the online database of structured data. The system can then refine the obtained entities in order to identify the entities of the type “person” that relate to the particular entity.
FIG. 8 is a flow diagram of an example process 800 for mapping an entity to related entities of the same type. For convenience, the process 800 will be described as being performed by a system of one or more computers located in one or more locations. For example, a search system, e.g., search system 214 of FIG. 2, appropriately programmed, can perform the process 800.
The process 800 can be performed for multiple entities of multiple types. For example, the process 800 may be performed for each type for which, in response to a search query that is about a given entity of the type, it is desirable to return information identifying other entities of the type that are related to the given entity, e.g., because users often submit additional search queries referencing other entities of the type after submitting an initial query referencing an entity of the type.
The system obtains data identifying a particular entity of a particular type (step 802). The obtained data includes one or more names for the particular entity and identifies the type of the entity. The system can obtain the data, e.g., from the online database of structured data.
The system identifies other entities of the particular type that are related to the particular entity (step 804). For example, the system can query the online database of structured data to obtain data identifying other entities of the particular type. The system can then identify the entities of the particular type that are related to the particular entity based on how frequently a recognized reference to each other entity co-occurs with a recognized reference to the particular entity in resources indexed by an indexing engine, e.g., indexing engine 220 of FIG. 2. For example, the system can select as related entities a predetermined number of other entities that co-occur most frequently with the particular entity. Alternatively, the system can select as related entities those other entities that co-occur with the particular entity at a frequency that exceeds a threshold frequency.
The system orders the related entities (step 806). The system may calculate a co-occurrence score for each related entity based on how frequently a recognized reference to each related entity co-occurs with a recognized reference to the particular entity in resources indexed by the indexing engine, e.g., so that related entities that co-occur more frequently with the particular entity have a higher co-occurrence score than related entities that co-occur less frequently with the particular entity.
The system may also calculate a subsequent query score for each related entity based on how frequently each related entity is searched for by users after submitting a search query directed to the particular entity. For example, the system may obtain data that identifies, for each of the related entities, how frequently users submit search queries that include a name of the related entity after submitting an initial search query that includes a name of the particular entity, e.g., within a pre-determined window of time of submitting the initial search query or immediately following submitting the initial search query. The system can then calculate the subsequent query scores, e.g., so that entities that are searched for more frequently after the initial search query is submitted have higher subsequent query scores than related entities that are searched for less frequently after the initial search query is submitted.
The system can then generate a ranking score for each related entity based on the co-occurrence score, the subsequent query score, or both and order the related entities in accordance with the ranking scores. The system can generate the ranking score for a given entity by, e.g., normalizing the co-occurrence score and the subsequent query score for the entity and then computing an average of the normalized scores, computing a sum of the normalized scores, computing a product of the normalized scores, or otherwise aggregating the normalized scores.
The system associates the related entities with the particular entity in an index (step 808). For example, the system can generate a mapping from the particular entity to each of the related entities. The mapping can also identify the type of the related entities and, optionally, the nature of the relationship between the particular entity and the related entities. The index can be, e.g., one of the indices included in the related entities index database 250. The system can generate the mapping in such a manner that, when the related entities are selected for presentation to a user in response to a search query directed to the particular entity, the related entities are presented in an order that matches the order generated by the system.
In some implementations, before associating the related entities of the particular type with the particular entity, the system determines whether, for each related entity, both the co-occurrence score and the subsequent query score for the related entity exceed respective threshold values. If, for a given related entity, either score does not exceed the threshold value, the system can refrain from associating the related entity with the particular entity in the index.
Additionally, in some implementations, the system determines whether sufficient data exists for the particular entity before associating any related entities of the particular type with the particular entity. For example, the system can determine whether a recognized reference to the entity occurs in more than a threshold number of search queries or in more than a threshold number of resources. If insufficient data exists for the particular entity, the system can determine not to associate any related entities of the particular type with the entity in the index.
Embodiments of the subject matter and the functional operations described in this specification can be implemented in digital electronic circuitry, in tangibly-embodied computer software or firmware, in computer hardware, including the structures disclosed in this specification and their structural equivalents, or in combinations of one or more of them. Embodiments of the subject matter described in this specification can be implemented as one or more computer programs, i.e., one or more modules of computer program instructions encoded on a tangible non-transitory program carrier for execution by, or to control the operation of, data processing apparatus. Alternatively or in addition, the program instructions can be encoded on an artificially-generated propagated signal, e.g., a machine-generated electrical, optical, or electromagnetic signal, that is generated to encode information for transmission to suitable receiver apparatus for execution by a data processing apparatus. The computer storage medium can be a machine-readable storage device, a machine-readable storage substrate, a random or serial access memory device, or a combination of one or more of them.
The term “data processing apparatus” refers to data processing hardware and encompasses all kinds of apparatus, devices, and machines for processing data, including by way of example a programmable processor, a computer, or multiple processors or computers. The apparatus can also be or further include special purpose logic circuitry, e.g., an FPGA (field programmable gate array) or an ASIC (application-specific integrated circuit). The apparatus can optionally include, in addition to hardware, code that creates an execution environment for computer programs, e.g., code that constitutes processor firmware, a protocol stack, a database management system, an operating system, or a combination of one or more of them.
A computer program (which may also be referred to or described as a program, software, a software application, a module, a software module, a script, or code) can be written in any form of programming language, including compiled or interpreted languages, or declarative or procedural languages, and it can be deployed in any form, including as a stand-alone program or as a module, component, subroutine, or other unit suitable for use in a computing environment. A computer program may, but need not, correspond to a file in a file system. A program can be stored in a portion of a file that holds other programs or data, e.g., one or more scripts stored in a markup language document, in a single file dedicated to the program in question, or in multiple coordinated files, e.g., files that store one or more modules, sub-programs, or portions of code. A computer program can be deployed to be executed on one computer or on multiple computers that are located at one site or distributed across multiple sites and interconnected by a communication network.
The processes and logic flows described in this specification can be performed by one or more programmable computers executing one or more computer programs to perform functions by operating on input data and generating output. The processes and logic flows can also be performed by, and apparatus can also be implemented as, special purpose logic circuitry, e.g., an FPGA (field programmable gate array) or an ASIC (application-specific integrated circuit).
Computers suitable for the execution of a computer program include, by way of example, can be based on general or special purpose microprocessors or both, or any other kind of central processing unit. Generally, a central processing unit will receive instructions and data from a read-only memory or a random access memory or both. The essential elements of a computer are a central processing unit for performing or executing instructions and one or more memory devices for storing instructions and data. Generally, a computer will also include, or be operatively coupled to receive data from or transfer data to, or both, one or more mass storage devices for storing data, e.g., magnetic, magneto-optical disks, or optical disks. However, a computer need not have such devices. Moreover, a computer can be embedded in another device, e.g., a mobile telephone, a personal digital assistant (PDA), a mobile audio or video player, a game console, a Global Positioning System (GPS) receiver, or a portable storage device, e.g., a universal serial bus (USB) flash drive, to name just a few.
Computer-readable media suitable for storing computer program instructions and data include all forms of non-volatile memory, media and memory devices, including by way of example semiconductor memory devices, e.g., EPROM, EEPROM, and flash memory devices; magnetic disks, e.g., internal hard disks or removable disks; magneto-optical disks; and CD-ROM and DVD-ROM disks. The processor and the memory can be supplemented by, or incorporated in, special purpose logic circuitry.
To provide for interaction with a user, embodiments of the subject matter described in this specification can be implemented on a computer having a display device, e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor, for displaying information to the user and a keyboard and a pointing device, e.g., a mouse or a trackball, by which the user can provide input to the computer. Other kinds of devices can be used to provide for interaction with a user as well; for example, feedback provided to the user can be any form of sensory feedback, e.g., visual feedback, auditory feedback, or tactile feedback; and input from the user can be received in any form, including acoustic, speech, or tactile input. In addition, a computer can interact with a user by sending documents to and receiving documents from a device that is used by the user; for example, by sending web pages to a web browser on a user's device in response to requests received from the web browser.
Embodiments of the subject matter described in this specification can be implemented in a computing system that includes a back-end component, e.g., as a data server, or that includes a middleware component, e.g., an application server, or that includes a front-end component, e.g., a client computer having a graphical user interface or a Web browser through which a user can interact with an implementation of the subject matter described in this specification, or any combination of one or more such back-end, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication, e.g., a communication network. Examples of communication networks include a local area network (“LAN”) and a wide area network (“WAN”), e.g., the Internet.
The computing system can include clients and servers. A client and server are generally remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other.
While this specification contains many specific implementation details, these should not be construed as limitations on the scope of any invention or of what may be claimed, but rather as descriptions of features that may be specific to particular embodiments of particular inventions. Certain features that are described in this specification in the context of separate embodiments can also be implemented in combination in a single embodiment. Conversely, various features that are described in the context of a single embodiment can also be implemented in multiple embodiments separately or in any suitable subcombination. Moreover, although features may be described above as acting in certain combinations and even initially claimed as such, one or more features from a claimed combination can in some cases be excised from the combination, and the claimed combination may be directed to a subcombination or variation of a subcombination.
Similarly, while operations are depicted in the drawings in a particular order, this should not be understood as requiring that such operations be performed in the particular order shown or in sequential order, or that all illustrated operations be performed, to achieve desirable results. In certain circumstances, multitasking and parallel processing may be advantageous. Moreover, the separation of various system modules and components in the embodiments described above should not be understood as requiring such separation in all embodiments, and it should be understood that the described program components and systems can generally be integrated together in a single software product or packaged into multiple software products.
Particular embodiments of the subject matter have been described. Other embodiments are within the scope of the following claims. For example, the actions recited in the claims can be performed in a different order and still achieve desirable results. As one example, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In certain implementations, multitasking and parallel processing may be advantageous.

Claims

What is claimed is:

1-20. (canceled)

21. A method performed by one or more computers, the method comprising:

obtaining data identifying a first entity of a first entity type;

identifying a plurality of second entities of the first entity type that are related to the first entity;

determining a respective ranking score for each of the second entities of the plurality of second entities of the first entity type;

ordering the second entities of the first entity type according the ranking scores; and

generating a mapping from the first entity to each of the second entities based on the ordering of the second entities.

22. The method of claim 21, further comprising:

receiving a search query from a user device, wherein the search query has been determined to relate to the first entity of the first entity type;

in response to the search query, transmitting, to the user device, i) search results responsive to the search query and ii) information identifying an order of the second entities of the first entity type that matches the ordering.

23. The method of claim 22, wherein the mapping defines the order of the second entities of the first entity type in response to received search queries that match the ordering.

24. The method of claim 21, wherein identifying the plurality of second entities further comprises:

identifying the plurality of second entities based on each of the second entities having a frequency of co-occurrence with the first entity in resources of an index of resources greater than a threshold.

25. The method of claim 24, further comprising:

determining, for each second entity, a co-occurrence score for the second entity based on the frequency of co-occurrence with the first entity in resources of the index of resources,

wherein the ranking score for each of the second entities is based on the respective co-occurrence score for the respective second entity.

26. The method of claim 21, further comprising:

determining, for each second entity, a subsequent query score for the second entity based a frequency of queries for the second entity after a submission of a query directed to the first entity,

wherein the ranking score for each of the second entities is based on the respective subsequent query score for the respective second entity.

27. The method of claim 24, further comprising:

determining, for each second entity, a co-occurrence score for the second entity based on the frequency of co-occurrence with the first entity in resources of an index of resources,

wherein the ranking score for each of the second entities is based on the respective co-occurrence score and the respective subsequent query score for the respective second entity.

28. A system comprising one or more computers and one or more storage devices storing instructions that when executed by the one or more computers cause the one or more computers to perform operations comprising:

obtaining data identifying a first entity of a first entity type;

29. The system of claim 28, the operations further comprising:

30. The system of claim 29, wherein the mapping defines the order of the second entities of the first entity type in response to received search queries that match the ordering.

31. The system of claim 28, wherein identifying the plurality of second entities further comprises:

32. The system of claim 31, the operations further comprising:

33. The system of claim 28, the operations further comprising:

34. The system of claim 31, the operations further comprising:

35. One or more non-transitory computer storage media storing instructions that when executed by one or more computers cause the one or more computers to perform operations comprising:

obtaining data identifying a first entity of a first entity type;

36. The computer storage media of claim 35, the operations further comprising:

37. The computer storage media of claim 36, wherein the mapping defines the order of the second entities of the first entity type in response to received search queries that match the ordering.

38. The computer storage media of claim 35, wherein identifying the plurality of second entities further comprises:

39. The computer storage media of claim 38, the operations further comprising:

40. The computer storage media of claim 35, the operations further comprising: