JP7003020B2 - Information processing equipment, information processing methods, and programs - Google Patents

Description

The present invention relates to an information processing apparatus, an information processing method, and a program.

Conventionally, a crawler that collects data (documents, images, etc.) from the Web and automatically creates a database of the collected data has been known. Crawlers collect data from web pages with various IP addresses by following links in web pages. The data collected by the crawler is stored in the web information database. On the other hand, when the search engine receives the search word input by the user (hereinafter, also referred to as "search query"), the information on the web page or application page related to the received search word (for example, URL: Uniform Resource Locator). Is acquired from the web information database, and the list of acquired information is output as a search result.

Search engines may be operated to output data related to a search query entered by a user together with search results in order to improve user satisfaction. In order to realize this, a database in which text data and an entity are stored in association with each other may be used (see Patent Document 1). When text data associated with a particular entity is entered in the database as a search query, information about that particular entity is output along with the search results.

Japanese Unexamined Patent Publication No. 2017-76403

The search query entered by the user may include other names such as the nickname (nickname), abbreviation, common name, and self-name of the entity in addition to the official name of the entity. If these aliases and the entity are not associated, the information about the entity may not be acquired, or the information about the inappropriate entity that the user does not intend may be acquired and output together with the search result.

The present invention has been made in consideration of such circumstances, and an object of the present invention is to provide an information processing device, an information processing method, and a program capable of collecting information on aliases associated with an entity. I will do it.

One aspect of the present invention includes a generation unit that generates a candidate for an alias of an entity, a calculation unit that calculates first information indicating the degree of association between the entity and a candidate for another name generated by the generation unit, and the above. When the first information is input, the generator is based on a learning model trained to output information indicating whether or not the candidate alias associated with the first information is an alias of an entity. It is an information processing apparatus including a determination unit for determining whether or not the generated alias candidate is an alias of the entity.

According to one aspect of the invention, alias information associated with an entity can be collected.

It is a figure which shows the use environment and the configuration of the knowledge data server which concerns on this embodiment. It is a figure which shows an example of the knowledge graph which concerns on this embodiment. It is a figure which shows an example of the web page of "baseball player A" which concerns on this embodiment. It is a figure which shows an example of the search result web page which concerns on this embodiment. It is a figure explaining an example of the redirect process of the encyclopedia web page in this embodiment. It is a figure which shows an example of the redirect information of the encyclopedia web page in this embodiment. It is a figure which shows an example of the encyclopedia web page collected by the collection part in this embodiment. It is a figure which shows an example of the encyclopedia web page collected by the collection part in this embodiment. It is a figure which shows an example of the alias candidate information in this embodiment. It is a figure which shows an example of the alias list in this embodiment. It is a flow chart which shows an example of the learning operation of the knowledge data server of this embodiment. It is a figure explaining an example of the redirect process of the encyclopedia web page in this embodiment. It is a flow chart which shows an example of the alias determination operation of the knowledge data server of this embodiment. It is a flow chart which shows an example of the online operation of the knowledge data server of this embodiment.

Hereinafter, embodiments of the information processing apparatus, information processing method, and program of the present invention will be described with reference to the drawings. In the present embodiment, the information processing apparatus will be described as constituting a part of the knowledge data server. The knowledge data server is, for example, a server that collects data to be collected (for example, image or text data) and generates a knowledge graph based on the collected data. In the present embodiment, the knowledge graph is data in which an entity, a class, and a property are associated with the location information of a web page in which related data related to the entity is described.

An entity may represent, for example, an entity of an object (eg, an object that exists in the real world), or may be defined in the concept of an object (eg, in the real world or virtual world). It may represent a concept). For example, when the object is the concept of "building", the entity may represent an entity such as "○○ tower" or "○○ building". Further, for example, when the object is the concept of "economics", the entity may represent an insubstantial abstract concept such as "microeconomics" or "macroeconomics". Hereinafter, embodiments will be described in detail.

<Embodiment>
<1-1. Knowledge data server usage environment>
FIG. 1 is a diagram showing a usage environment and configuration of the knowledge data server 100 according to the present embodiment. The knowledge data server 100 is communicably connected to the terminal device 200, the web server 300, the crawl target device DV, and the network NW. The network NW means the World Wide Web, and is a system using HTML documents and the like that are standardly used on the Internet and intranets. The network NW may further include a radio base station, a provider device, a dedicated line, and the like.

The terminal device 200 is a device used by a user, and is, for example, a mobile phone such as a smartphone, a tablet computer, a notebook computer, a desktop computer, or the like. The web server 300 is a server that uses a search engine to generate a web page for displaying search results and provides it to the terminal device 200.

The knowledge data server 100 includes, for example, a control unit 110 and a storage unit 120. The storage unit 120 is realized by, for example, a RAM (Random Access Memory), an HDD (Hard Disk Drive), a flash memory, or a hybrid storage device in which a plurality of these are combined. Further, a part or all of the storage unit 120 may be an external device such as NAS (Network Attached Storage) or an external storage server that can be accessed by the knowledge data server 100. The storage unit 120 stores, for example, the knowledge graph D1, the alias candidate information D2, the determination model D3, and the alias list D4.

FIG. 2 is a diagram showing an example of Knowledge Graph D1 according to the present embodiment. The things described in Knowledge Graph D1 are defined by the ontology. An ontology is a definition of a class and a property of an object, and is a collection of constraints that hold between the class and the property.

A class is a group of things that have the same properties in an ontology. What the nature of an object is, that is, which class the object belongs to, is determined by the properties described below.

For example, things that have a beak, an oviparous vertebrate, and forelimbs that are wings are classified in the "bird" class. Also, in the class of "birds", things that have the property of not being able to fly are classified into lower classes such as "penguins" and "ostriches". In this way, the class system may have a hierarchical structure having a higher-lower relationship. The nature of the upper class is inherited by the lower class. In the example above, the "bird" class's "beak-bearing, oviparous vertebrate with winged forelimbs" property is a lower class property of "penguins" and "ostriches". Will also be included. For example, if the class name for identifying the class is a class "bird", the class name may be represented by the character string "bird". The class name does not necessarily have to represent a meaning, and for example, even in the class "bird", a character string indicating mere identification information such as "information 1" or "C1" may be assigned. .. The above-mentioned entity, that is, an entity, belongs to one of the classes included in the class system defined by the ontology.

Properties are attributes that describe the nature and characteristics of things and the relationships between classes. For example, a property may be an attribute that has the property of "having ... as a component of the body" or the property of "living in", or "a class is a higher class and a certain class is a lower class". It may be an attribute indicating the relationship between the upper and lower levels of "class". As for the property name for identifying the property, the property name itself may or may not represent the meaning, as in the class name described above.

The Knowledge Graph D1 is a directed graph in which the above-mentioned classes are represented as nodes and the above-mentioned properties are represented as labeled and directional edges. With such a graph structure, information about things can be discriminated by nodes, and relationships between things can be discriminated by edges.

In the knowledge graph D1 of the example shown in FIG. 2, the entity E1 and the entity E2 belong to a class having a property name of "affiliated team". In the present embodiment, each entity is associated with information that can identify each entity (hereinafter, "entity identification information EID"), an entity name, and location information of a web page of the entity.

The location information is information for specifying a position on the Web, and is, for example, a URL. The web page indicated by the location information contains the relevant data of the entity. In the knowledge graph D1 shown in FIG. 2, the entity E1 is the location of a web page in which the entity identification information EID "0001", the entity name "baseball player A", and related data related to the entity are described. It is associated with the URL "http: // encyclopedia web page / baseball player A" which is information. Further, the entity E2 is the location information of the web page in which the entity identification information EID "0002", the entity name "baseball team B", and the related data related to the entity are described, "http: //". It is associated with the URL "Encyclopedia Web Page / Baseball Team B". In the following description, the related data related to the entity will be described as the related data of the entity. In addition, the web page in which the data related to the entity is described is also described as the web page of the entity.

FIG. 3 is a diagram showing an example of a web page of “baseball player A” according to the present embodiment. The web page of "baseball player A" contains data related to "baseball player A".

<1-2. Search result web page ＞
FIG. 4 is a diagram showing an example of a search result web page according to the present embodiment. As shown in FIG. 4, the display unit 210 of the terminal device 200 displays a search result web page including a query input area 211, a knowledge panel 220, and a search result 230. The knowledge panel 220 includes the entity 221 associated with the search query, the image 222 associated with the entity 221 and the relevant information 223 associated with the entity 221. The user uses the terminal device 200 to input a search query on the search page displayed on the display unit 210 of the terminal device 200. A search query is a search word or a combination of a plurality of search words. The terminal device 200 sends the search query entered by the user to the web server 300.

The web server 300 transmits the search query received from the terminal device 200 to the knowledge data server 100. The knowledge data server 100 sends information related to the entity corresponding to the received search query to the web server 300. The knowledge data server 100 transmits, for example, the location information associated with the entity associated with the received search query to the web server 300 based on the knowledge graph D1. Based on the received location information, the web server 300 extracts the relevant data of the entity from the web pages indicated by the location information. The web server 300 generates a search result web page including the related data of the extracted entity (in this example, the information shown in the knowledge panel 220).

In the example shown in FIG. 4, "baseball player A" is input as a search query, and an image and various information related to "baseball player A" are displayed on the knowledge panel 220. On the knowledge panel 220, for example, an image 222 of "baseball player A" and related information 223 such as the date of birth and birthplace of "baseball player A" are displayed. In this way, by displaying not only the search result 230 but also the knowledge panel 220 on the display unit 210, the user satisfaction with the search result can be improved.

<1-3. Configuration of Knowledge Data Server 100>
Returning to FIG. 1, the control unit 110 is realized by, for example, a hardware processor such as a CPU (Central Processing Unit) executing a program (software). The control unit 110 includes, for example, a communication I / F (Interface) 111, a collection unit 112, an alias candidate extraction unit 113 (generation unit), a feature extraction unit 114 (calculation unit), a determination unit 115, and an alias list. It includes a generation unit 116, a location information acquisition unit 117, and a learning unit 118. In addition, some or all of these components (excluding the contained storage unit) are LSI (Large Scale Integration), ASIC (Application Specific Integrated Circuit), FPGA (Field-Programmable Gate Array), GPU (Graphics Processing). It may be realized by hardware (circuit unit; including circuitry) such as Unit), or it may be realized by the cooperation of software and hardware. The communication I / F 111 communicates with each functional unit via the network NW and other devices (for example, a crawl target device (hereinafter, “crawl target device DV”), a terminal device 200, and a web server 300). Mediate.

The collection unit 112 collects data to be collected for a predetermined web page from the device DV to be crawled via the network NW. The collection target data to be collected by the collection unit 112 is, for example, an encyclopedia web page (hereinafter, “encyclopedia web page”) that can be edited by a user, a search engine search log, or the like.

The collecting unit 112 is realized by a so-called crawler program. The data to be collected includes data that is on the network NW (in the storage area of the device DV to be crawled) and can be viewed by a browser. The data to be collected is not limited to the browser, and may be data reproduced by the application program. The data to be collected includes, for example, a web page containing text data indicating a source of HTML (HyperText Markup Language). The collection unit 112 collects all of the encyclopedia web pages, for example, if there is a web page containing the name of each entity included in the Nagilegraph D1 and the corresponding text data.

The data to be collected also includes, for example, redirect information about the encyclopedia web page. FIG. 5 is a diagram illustrating an example of the redirect process of the encyclopedia web page in the present embodiment. FIG. 6 is a diagram showing an example of redirect information of the encyclopedia web page in the present embodiment. For example, on an encyclopedia web page, when a user operates a terminal device 200 and inputs "alias a" (second word), which is another name for "baseball player A" (first word), as a search query, the terminal device. In 200, the web page P1 (second page) associated with the input "alias a" is not displayed, and the redirect to the web page P2 (first page) associated with "baseball player A" (1st page). (Transfer) is performed, and this web page P2 is displayed on the terminal device 200. That is, the web page P1 is a redirect source page, and is a page in which the web page P2 is set as the redirect destination page (a transfer page in which the URL of the web page P2 is set as the redirect destination). As shown in FIG. 6, the redirect information includes a word associated with the redirected page as described above and a list associated with the word associated with the redirected page. FIG. 6 includes, for example, data in which the redirect source word “alias a” and the redirect destination word “baseball player A” are associated with each other.

It should be noted that, for example, the operator of the knowledge data server 100 manually collects the data to be collected by operating the browser of the operation terminal without performing the crawl process by the collection unit 112, and the input unit provided in the knowledge data server 100 ( Data to be collected may be input via (not shown).

The redirect source word is a synonym for the redirect destination word, and it is assumed that these two words represent the same entity. That is, the redirect source word is assumed to be another name for the redirect destination word. Therefore, the alias candidate extraction unit 113 extracts the redirect source word and the redirect destination word from the redirect information included in the collection target data collected by the collection unit 112, and extracts the redirect source word and the redirect destination word. It is registered in the alias candidate information D2 as a candidate for another name (hereinafter, "alias candidate").

Further, the alias candidate extraction unit 113 extracts guidance information from the text data indicating the HTML source of the encyclopedia web page included in the collection target data collected by the collection unit 112, and the alias candidate extraction unit 113 extracts guidance information based on the extracted guidance information. Is extracted and registered in the alias candidate information D2. FIG. 7 is a diagram showing an example of an encyclopedia web page collected by the collection unit 112 in the present embodiment. The encyclopedia web page shown in FIG. 7 is a page associated with "Character C" (first entity) and associated with each of "Character C (animation) T1" and "Character C (television drama) T2". Contains information to guide you to the page. Both of these "character C (animation)" and "character C (television drama)" (second entity) can be entities. In this case, the word "character C" may be another name (for example, abbreviation) for "character C (animation)" and "character C (television drama)" as entity. Therefore, the alias candidate extraction unit 113 uses a word associated with a certain web page (“character C” in the example shown in FIG. 7) as a word included in the guidance information included in the web page (shown in FIG. 7). In the example, it is registered in the alias candidate information D2 as an alias candidate of "character C (animation)" and "character C (television drama)").

Further, the alias candidate extraction unit 113 extracts the alias candidate based on the search log of the search engine included in the collection target data collected by the collection unit 112, and registers the alias candidate in the alias candidate information D2. For example, the alias candidate extraction unit 113 collects information on the usage status such as a click log of a link included in a search result web page. The alias candidate extraction unit 113 includes the search query used for the search and the name of the entity associated with the linked page (for example, the encyclopedia web page) clicked by the user among the links included in the search result web page. If not, extract this search query as an alias candidate for the entity. The alias candidate extraction unit 113 includes only the entities associated with the linked page of the link whose click count, click through rate (CTR), etc. are equal to or higher than a predetermined threshold among the links included in the search result web page. On the other hand, the search query may be set as an alias candidate. In the following, the case where the data to be collected is the redirect information related to the encyclopedia web page will be described as an example.

Returning to FIG. 1, the feature extraction unit 114 extracts feature information (first information) for each of the pair of the entity and the alias candidate extracted by the alias candidate extraction unit 113 with respect to the entity. The feature information is information indicating the degree of association between the entity and the alias candidate. The feature information is information that serves as an index for determining the certainty of the alias candidate. The feature information includes, for example, the first feature information based on the summary information of the entity included in the encyclopedia web page, the second feature information based on the usage status such as the click log of the link included in the search result web page, and the alias candidate. The third feature information and the fourth feature information based on the character string structure are included. The feature extraction unit 114 extracts, for example, at least one of the first to fourth feature information. It should be noted that the feature information is not limited to the above four types of the first to fourth feature information, and any information that is useful for showing the degree of association between the entity and the alias candidate can be arbitrarily added. can.

(1st feature information)
FIG. 8 is a diagram showing an example of an encyclopedia web page of “baseball player A” according to the present embodiment. The "Baseball Player A" encyclopedia web page contains information (summary information) about the profile of "Baseball Player A", and the information about this profile contains the string "nickname: BBB (N1)". include. The feature extraction unit 114 refers to, for example, this profile information (information (key) in the column "nickname" and "BBB" (value) indicating the content), and each of the alias candidates included in the alias candidate information D2. Determines if it contains page summary information for the associated entity's page. As described above, when the feature extraction unit 114 includes the alias candidate in the summary information, for example, "1" is recorded in the feature 1 column of the alias candidate information D2, and the alias candidate is included in the summary information. If not, "0" is recorded in the feature 1 column of the alias candidate information D2.

(Second feature information)
The feature extraction unit 114 is associated when each of the alias candidates included in the alias candidate information D2 is used as a search query based on the information on the usage status such as the click log of the link included in the search result web page. The second feature information indicating the probability of transitioning to the page associated with the entity (probability that the alias candidate is associated with the corresponding entity) is calculated. For example, the feature extraction unit 114 calculates the click rate of the link included in the search result web page as the second feature information, and records it in the feature 2 column of the alias candidate information D2.

(Third feature information)
The identity extraction unit 114 calculates the third identity information indicating whether or not each character string of the alias candidate included in the alias candidate information D2 is a substring of the character string of the associated entity, and calculates the alias candidate. Register in the column of the identity 3 of the information D2. For example, if the character string of the alias candidate is "ABCD" and the character string of the entity is "Tokyo ABCD", the character string "ABCD" of the alias candidate is a substring of the character string "Tokyo ABCD" of the entity. Is. When each character string of the alias candidate is a substring of the character string of the associated entity, the identity extraction unit 114 records, for example, "1" in the identity 3 column of the alias candidate information D2, and the identity extraction unit 114 records the alias candidate. If each character string of is not a substring of the character string of the associated entity, "0" is recorded in the column of the identity 3 of the alias candidate information D2.

(4th feature information)
The identity extraction unit 114 calculates the edit distance (Levenshtein distance) between each character string of the alias candidate included in the alias candidate information D2 and the character string of the associated entity, and indicates the calculated edit distance. The identity information is recorded in the column of the identity 4 of the alias candidate information D2. For example, when the character string of the alias candidate is "ABE film maker" and the character string of the entity is "Abe film maker", the editing distance is "3".

Returning to FIG. 1, the determination unit 115 determines whether or not each of the alias candidates included in the alias candidate information D2 is an alias of the associated entity based on the determination model D3. The determination model D3 is modeled so as to output a determination result of whether or not the alias candidate related to the feature information is an alias of the entity when the feature information (for example, at least one of the features 1 to 4) is given. It was done. The determination result may be an alias indicating whether or not it is an alias (for example, "1" indicating that it is an alias or "0" indicating that it is not an alias), or a numerical value indicating the probability of being an alias. May be. Hereinafter, a case where the determination result indicates whether or not the determination result is an alias will be described as an example.

The decision model D3 is, for example, logistic regression analysis, decision tree analysis, support vector machine, recurrent neural network (RNN), multi-layered neural network (DNN), or convolutional neural network (Convolutional). It may be a model generated by arbitrary machine learning such as a technique based on deep learning using Neural Network (CNN).

When the determination unit 115 determines that the alias candidate is an alias of the associated entity, for example, records "1" in the determination result column of the alias candidate information D2 and determines that the alias candidate is not an alias. “0” is recorded in the determination result column of the candidate information D2.

FIG. 9 is a diagram showing an example of alias candidate information D2 in the present embodiment. In the example shown in FIG. 9, for the entity "baseball player A", the alias candidate "alias a", the feature 1 "1 (included in the summary information)", and the feature 2 "0.14 (click)". Rate) ”, feature 3“ 1 (substring) ”, feature 4“ 3 (editing distance) ”, and determination result“ 1 (alias) ”are associated with each other.

The alias list generation unit 116 has a determination result of "1" among the data included in the alias candidate information D2, that is, the "entity" and "entity" of the data for which the determination unit 115 determines that the alias candidate is an alias. The pair with "alias candidate" is extracted to generate the alias list D4. FIG. 10 is a diagram showing an example of the alias list D4 in the present embodiment. In the example shown in FIG. 10, alias 1 "alias a" and alias 2 "alias c" are associated with the entity "baseball player A". Instead of generating the alias list D4, the extracted "alias candidate" may be associated with the "entity" and registered in the knowledge graph D1.

The location information acquisition unit 117 acquires the location information associated with the entity corresponding to the search query received from the web server 300 based on the knowledge graph D1 and the alias list D4. The location information acquisition unit 117 determines whether or not the entity name matching the received search query exists in the entity name included in the knowledge graph D1. When it is determined that the entity name matching the received search query exists, the location information acquisition unit 117 acquires the location information associated with the entity.

On the other hand, when the location information acquisition unit 117 determines that the entity name matching the received search query does not exist in the knowledge graph D1, it matches the received search query among the aliases registered in the alias list D4. Determine if an alias exists. When the location information acquisition unit 117 determines that an alias matching the received search query exists in the alias list D4, the location information acquisition unit 117 acquires the location information associated with the entity associated with the alias. At this time, not only when the search query and the entity name or alias are exactly matched, but also when there is a partial match or a difference to the extent that they are considered to be semantically identical, it may be regarded as the corresponding entity name or alias. When the alias is registered in the knowledge graph D1, the location information acquisition unit 117 determines whether or not there is an alias that matches the received search query based on the alias registered in the knowledge graph D1. You may.

Next, the location information acquisition unit 117 transmits the acquired location information to the web server 300. When the location information acquisition unit 117 determines that the entity and alias matching the received search query do not exist, the location information acquisition unit 117 transmits information indicating that the location information does not exist to the web server 300.

Returning to FIG. 1, the learning unit 118 generates the judgment model D3 by performing machine learning using the learning data including the set of the entity name, the alias, and the identity information, and stores the generated judgment model D3. It is stored in the unit 120. The learning unit 118 may employ any technique such as logistic regression analysis, decision tree analysis, support vector machine, recurrent neural network, multi-layered neural network, or deep learning technique using a convolutional neural network. ..

<1-4. Learning behavior ＞
FIG. 11 is a flow chart showing an example of the learning operation of the knowledge data server 100 of the present embodiment. The knowledge data server 100 performs a learning operation by batch processing executed at an arbitrary timing such as daily or weekly.

First, the learning unit 118 generates learning data to be learned (S10). For example, the learning unit 118 extracts a pair of "entity" and "alias" already registered in the knowledge graph D1, and "feature information (for example, feature) for the extracted pair of" entity "and" alias ". At least one of 1 to 4) ”is calculated, and the set of“ entity ”,“ alias ”, and“ feature information ”is used as learning data.

In addition, the learning unit 118 may generate learning data based on the redirect information of the encyclopedia web page. FIG. 12 is a diagram illustrating an example of the redirect process of the encyclopedia web page in the present embodiment. In FIG. 12, the redirect 1 from the web page P1 associated with the “alias a” to the web page P2 associated with the “baseball player A” and the web page P1 associated with the “alias a” from the web page P1 to the “baseball team”. Redirect 2 to a specific location ("member information" column) in the web page P3 associated with "B" is shown. To redirect to a specific position in a web page, set an A tag that specifies the id attribute and name attribute at the location that receives the redirect in the web page, and set the name attribute specified above in the argument of the URL that specifies the redirect destination. It is realized by specifying the name of.

In this case, since the redirect 1 is a redirect to the web page P2 itself associated with the "baseball player A", the "alias a" associated with the redirect source web page P1 is the redirect destination web page P2. It is highly probable that it is another name for "baseball player A" associated with. On the other hand, since the redirect 2 is not the web page P3 itself associated with the "baseball team B" but a redirect to a specific position in the web page P3, the "alias a" associated with the redirect source web page P1. Is unlikely to be another name for "baseball team B" associated with the redirected web page P2. Therefore, the learning unit 118 adopts the set of the entity whose redirect destination is the web page itself and the alias among the data obtained based on the redirect information about the encyclopedia web page as the learning data. On the other hand, the learning unit 118 does not adopt the pair of the entity whose redirect destination is a specific position in the web page and the alias as learning data. The learning data may be generated by the operator of the knowledge data server.

Next, the learning unit 118 performs machine learning using the generated learning data, and generates a determination model D3 that learns the relationship between the entity, the alias of the entity, and the identity information of the entity (S12). The generated determination model D3 is stored in the storage unit 120 (S14). This completes the processing of this flowchart.

<1-5. Alias judgment operation>
FIG. 13 is a flow chart showing an example of the alias determination operation of the knowledge data server 100 of the present embodiment. The knowledge data server 100 performs an alias determination operation by, for example, a batch process executed at an arbitrary timing such as daily or weekly.

First, the alias candidate extraction unit 113 extracts the redirect source word and the redirect destination word from the redirect information included in the collection target data collected by the collection unit 112, and sets the redirect source word as the redirect destination word ( That is, it is extracted as an alias candidate of the entity name) (S20). The alias candidate extraction unit 113 records the set of the extracted entity and the alias candidate in the alias candidate information D2.

Next, the feature extraction unit 114 extracts feature information (for example, at least one of features 1 to 4) for each of the set of the entity included in the alias candidate information D2 and the alias candidate (S22). .. The feature extraction unit 114 records the extracted feature information in the alias candidate information D2.

Next, the determination unit 115 determines whether or not each of the alias candidates is an alias of the associated entity based on the alias candidate information D2 and the determination model D3 (S24). The determination unit 115 records the determination result in the alias candidate information D2.

Next, the alias list generation unit 116 determines that the determination result is "1" among the data included in the alias candidate information D2 (that is, the determination unit 115 determines that the alias candidate is an alias of the associated entity. The pair of the "entity" and the "alias candidate" of the generated data) is extracted to generate the alias list D4 (S26). This completes the processing of this flowchart.

<1-6. Online operation>
FIG. 14 is a flow chart showing an example of online operation of the knowledge data server 100 of the present embodiment. First, the location information acquisition unit 117 receives a search query from the web server 300 (S30).

Next, the location information acquisition unit 117 determines whether or not the entity name matching the received search query exists in the entity name included in the knowledge graph D1 (S32). When the location information acquisition unit 117 determines that an entity name matching the received search query exists, the location information acquisition unit 117 acquires the location information associated with the entity, transmits it to the web server 300 (S34), and processes this flowchart. To finish.

On the other hand, when the location information acquisition unit 117 determines that the entity name matching the received search query does not exist in the knowledge graph D1, the received search query is included in the alias registered as the alias in the alias list D4. It is determined whether or not a matching alias exists (S36). When the location information acquisition unit 117 determines that an alias matching the received search query exists in the alias list D4, the location information acquisition unit 117 acquires the location information associated with the entity associated with the alias and transmits it to the web server 300. (S34). This completes the processing of this flowchart.

On the other hand, when the location information acquisition unit 117 determines that the alias matching the received search query does not exist, the location information acquisition unit 117 transmits information indicating that the location information does not exist to the web server 300 (S38). This completes the processing of this flowchart.

As described above, the knowledge data server 100 of the present embodiment has a generation unit (alias candidate extraction unit 113) that generates a candidate for an alias of an entity, the entity, and a candidate for an alias generated by the generation unit. When the calculation unit (identity extraction unit 114) that calculates the first information indicating the degree of association with the first information and the first information are input, is the candidate for the alias associated with the first information an alias of the entity? Based on a learning model trained to output information indicating whether or not, a determination unit (determination unit 115) that determines whether or not the alias candidate generated by the generation unit is an alias of the entity. , Equipped with. This allows you to collect alias information associated with the entity.

Although the embodiments for carrying out the present invention have been described above using the embodiments, the present invention is not limited to these embodiments, and various modifications and substitutions are made without departing from the gist of the present invention. Can be added.

100 ... Knowledge data server 110 ... Control unit 111 ... Communication I / F
112 ... Collection unit 113 ... Alias candidate extraction unit 114 ... Identity extraction unit 115 ... Judgment unit 116 ... Alias list generation unit 117 ... Location information acquisition unit 118 ... Learning unit 120 ... Storage unit 200 ... Terminal device 210 ... Display unit 211 ... Query Input area 220 ... Knowledge panel 221 ... Entity 222 ... Image 223 ... Related information 230 ... Search result 300 ... Web server D1 ... Knowledge graph D2 ... Alias candidate information D3 ... Judgment model D4 ... Alias list R1 ... Redirect information

Claims (12)

A generator that generates candidate aliases for an entity,
A calculation unit that calculates the first information indicating the degree of association between the entity and the alias candidate generated by the generation unit.
When the first information is input, the generation unit is based on a learning model trained to output information indicating whether or not the candidate for the alias associated with the first information is an alias of the entity. A determination unit that determines whether or not the alias candidate generated by is an alias of the entity.
Equipped with
The first information includes information on the usage status of a link included in a search result web page obtained when the candidate of the alias is used as a search query.
Information processing equipment. The first information includes information indicating whether or not the encyclopedia web page associated with the entity contains potential aliases for the entity.
The information processing apparatus according to claim 1. The first information includes information indicating whether or not the character string of the candidate for the alias is a substring of the character string of the entity.
The information processing apparatus according to claim 1 or 2 . The first information includes information indicating an edit distance between the character string of the candidate of the alias and the character string of the entity.
The information processing apparatus according to any one of claims 1 to 3 . A learning unit that generates the learning model is further provided.
The information processing apparatus according to any one of claims 1 to 4 . A generator that generates candidate aliases for an entity,
A calculation unit that calculates the first information indicating the degree of association between the entity and the alias candidate generated by the generation unit.
When the first information is input, the generation unit is based on a learning model trained to output information indicating whether or not the candidate for the alias associated with the first information is an alias of the entity. A determination unit that determines whether or not the alias candidate generated by is an alias of the entity.
The learning unit that generates the learning model and
Equipped with
Among the aliases obtained based on the redirect information of the encyclopedia web page, the learning department does not adopt the alias obtained based on the redirect information at a specific position in the web page as the learning data.
Information processing equipment . A generator that generates candidate aliases for an entity,
A calculation unit that calculates the first information indicating the degree of association between the entity and the alias candidate generated by the generation unit.
When the first information is input, the generation unit is based on a learning model trained to output information indicating whether or not the candidate for the alias associated with the first information is an alias of the entity. A determination unit that determines whether or not the alias candidate generated by is an alias of the entity.
Equipped with
The generator is the second word associated with the second page, which is the redirect source to the first page associated with the first word indicating the name of the first entity, based on the redirect information of the encyclopedia web page. Is a candidate for another name of the first entity.
Information processing equipment . A generator that generates candidate aliases for an entity,
A calculation unit that calculates the first information indicating the degree of association between the entity and the alias candidate generated by the generation unit.
When the first information is input, the generation unit is based on a learning model trained to output information indicating whether or not the candidate for the alias associated with the first information is an alias of the entity. A determination unit that determines whether or not the alias candidate generated by is an alias of the entity.
Equipped with
If the first page associated with the first word indicating the name of the first entity of the encyclopedia web page contains guidance information to the second page, the generator may use the first word as the second page. Candidate for a second entity alias as indicated by the second word associated with
Information processing equipment . A generator that generates candidate aliases for an entity,
A calculation unit that calculates the first information indicating the degree of association between the entity and the alias candidate generated by the generation unit.
When the first information is input, the generation unit is based on a learning model trained to output information indicating whether or not the candidate for the alias associated with the first information is an alias of the entity. A determination unit that determines whether or not the alias candidate generated by is an alias of the entity.
Equipped with
If the search query used for the search and the name of the entity associated with the linked page of the clicked link among the links included in the search result web page are different, the generation unit may execute the search query. Candidates for entity aliases,
Information processing equipment . A generator that generates candidate aliases for an entity,
A calculation unit that calculates the first information indicating the degree of association between the entity and the alias candidate generated by the generation unit.
When the first information is input, the generation unit is based on a learning model trained to output information indicating whether or not the candidate for the alias associated with the first information is an alias of the entity. A determination unit that determines whether or not the alias candidate generated by is an alias of the entity.
Equipped with
The first information includes information indicating whether or not the summary information of the encyclopedia web page associated with the entity includes candidates for aliases of the entity.
Information processing equipment . The computer
Generate candidate aliases for an entity
First information indicating the degree of association between the entity and the generated candidate for the alias is calculated.
When the first information is input, it is generated based on a learning model trained to output information indicating whether or not the candidate alias associated with the first information is an alias of an entity. Determining whether the candidate for the alias is an alias for the entity,
It is an information processing method
The first information includes information on the usage status of a link included in a search result web page obtained when the candidate of the alias is used as a search query.
Information processing method. On the computer
Generate a candidate for an alias for an entity
First information indicating the degree of association between the entity and the generated candidate for the alias is calculated.
When the first information is input, it is generated based on a learning model trained to output information indicating whether or not the candidate alias associated with the first information is an alias of an entity. To determine whether the candidate for the alias is an alias for the entity.
It ’s a program,
The first information includes information on the usage status of a link included in a search result web page obtained when the candidate of the alias is used as a search query.
program.

Images