CN112507715B

CN112507715B - Method, device, equipment and storage medium for determining association relation between entities

Info

Publication number: CN112507715B
Application number: CN202011377531.4A
Authority: CN
Inventors: 张峥; 徐伟建; 罗雨
Original assignee: Beijing Baidu Netcom Science and Technology Co Ltd
Current assignee: Beijing Baidu Netcom Science and Technology Co Ltd
Priority date: 2020-11-30
Filing date: 2020-11-30
Publication date: 2024-01-16
Anticipated expiration: 2040-11-30
Also published as: CN112507715A

Abstract

The application discloses a method, a device, equipment and a storage medium for determining association relations between entities, which are applied to the technical fields of natural language processing, knowledge maps and deep learning. The specific implementation scheme is as follows: acquiring target association information, wherein the target association information comprises a first word representing a first entity, a second word representing a second entity and a third word representing a target association relationship; determining first semantic features according to the target association relationship and similar text segments aiming at the target association information in preset text Duan Ku; determining a description text aiming at the target associated information according to the target associated information and a preset knowledge graph; determining a second semantic feature according to the target associated information and the descriptive text; and determining the confidence coefficient of the target association relationship between the first entity and the second entity according to the first semantic feature and the second semantic feature.

Description

Method, device, equipment and storage medium for determining association relation between entities

Technical Field

The present application relates to the field of artificial intelligence technology, and in particular, to the field of natural language processing, knowledge graph and deep learning technology, and more particularly, to a method, apparatus, device and storage medium for determining an association relationship between entities.

Background

With the development of artificial intelligence technology, electronic management of information is gradually rising. In each field, in order to improve information acquisition efficiency, information having an association relationship can be associated and managed when the information is electronically managed. Wherein, the association relation between the information can be determined by the association relation between the entities described by the information.

In the related art, when information association management is performed, two entities with association relationship are usually recalled from text data in each field by adopting a natural language processing method, and the method has high quality requirement on the text data. Moreover, for a specific field, the construction cost of some specific association relations is high, the information density of the specific association relations in the text is low, and the recall rate of recalling the entity with the specific association relations from the text by adopting a natural language processing method is limited.

Disclosure of Invention

Provided are a method, a device, equipment and a storage medium for improving the accuracy of determining association relations and facilitating the reduction of the construction cost of target association relations.

According to a first aspect, there is provided a method of determining an association between entities, comprising: acquiring target association information, wherein the target association information comprises a first word representing a first entity, a second word representing a second entity and a third word representing a target association relationship; determining first semantic features according to the target associated information and similar text segments aiming at the target associated information in preset text Duan Ku; determining a description text aiming at the target associated information according to the target associated information and a preset knowledge graph; determining a second semantic feature according to the target associated information and the descriptive text; and determining the confidence coefficient of the target association relationship between the first entity and the second entity according to the first semantic feature and the second semantic feature.

According to a second aspect, there is provided an apparatus for determining an association between entities, comprising: the information acquisition module is used for acquiring target association information, wherein the target association information comprises a first word representing a first entity, a second word representing a second entity and a third word representing a target association relationship; the first feature determining module is used for determining first semantic features according to the target associated information and similar text segments aiming at the target associated information in preset text Duan Ku; the text determining module is used for determining a description text aiming at the target associated information according to the target associated information and a preset knowledge graph; the second feature determining module is used for determining second semantic features according to the target associated information and the descriptive text; and the confidence determining module is used for determining the confidence coefficient of the target association relationship between the first entity and the second entity according to the first semantic feature and the second semantic feature.

According to a third aspect, there is provided an electronic device comprising: at least one processor; and a memory communicatively coupled to the at least one processor; the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method for determining associations between entities provided herein.

According to a fourth aspect, there is provided a non-transitory computer readable storage medium storing computer instructions for causing a computer to perform the method of determining an association between entities provided herein.

It should be understood that the description in this section is not intended to identify key or critical features of the embodiments of the disclosure, nor is it intended to be used to limit the scope of the disclosure. Other features of the present disclosure will become apparent from the following specification.

Drawings

The drawings are for better understanding of the present solution and do not constitute a limitation of the present application. Wherein:

fig. 1 is an application scenario schematic diagram of a method, an apparatus, a device and a storage medium for determining an association relationship between entities according to an embodiment of the present application;

FIG. 2 is a flowchart of a method for determining an association relationship between entities according to an embodiment of the present application;

FIG. 3 is a schematic diagram of determining a first semantic feature according to an embodiment of the present application;

FIG. 4 is a schematic diagram of determining a first semantic feature according to another embodiment of the present application;

FIG. 5 is a schematic diagram of determining descriptive text for target association information according to an embodiment of the present application;

FIG. 6 is a schematic diagram of determining an association relationship between entities according to an embodiment of the present application;

FIG. 7 is a block diagram of an apparatus for determining association between entities according to an embodiment of the present application; and

fig. 8 is a block diagram of an electronic device for implementing a method for determining association between entities according to an embodiment of the present application.

Detailed Description

Exemplary embodiments of the present application are described below in conjunction with the accompanying drawings, which include various details of the embodiments of the present application to facilitate understanding, and should be considered as merely exemplary. Accordingly, one of ordinary skill in the art will recognize that various changes and modifications of the embodiments described herein can be made without departing from the scope and spirit of the present application. Also, descriptions of well-known functions and constructions are omitted in the following description for clarity and conciseness.

The application provides a method for determining association relations between entities. The method may first obtain target association information including a first word representing a first entity, a second word representing a second entity, and a third word representing a target association relationship. Then determining first semantic features according to the target associated information and similar text segments aiming at the target associated information in preset text Duan Ku; and determining a description text aiming at the target associated information according to the target associated information and the preset knowledge graph, and determining a second semantic feature according to the target associated information and the description text. And finally, according to the first semantic features and the second semantic features, determining the confidence coefficient of the target association relationship between the first entity and the second entity.

An application scenario of the method and apparatus provided in the present application will be described below with reference to fig. 1.

Fig. 1 is an application scenario diagram of a method, an apparatus, a device, and a storage medium for determining an association relationship between entities according to an embodiment of the present application.

As shown in fig. 1, the application scenario 100 of this embodiment may include, for example, a terminal device 110 and a user 120.

Terminal device 110 may be, for example, a variety of electronic devices capable of providing an interactive interface and having processing functionality, including but not limited to smartphones, tablets, laptop portable computers, desktop computers, and the like.

According to the embodiment of the present application, the terminal device 110 may process the two entity words according to a pre-trained processing model, for example, so as to obtain an association relationship between two entities represented by the two entity words. For example, the terminal device 110 may be provided with an interactive interface, through which a user may input two entity words of which association relationships are to be determined, and select an estimated association relationship of the two entity words. Through the processing of the terminal device 110, the terminal device 110 may output a confidence that the two entity words have a predicted association relationship.

According to an embodiment of the present application, as shown in fig. 1, the application scenario may further include a server 130, where the terminal device 110 and the server 130 may communicate through a network. For example, terminal device 110 may obtain a pre-trained processing model from server 130 via a network.

The server 130 may also be, for example, various servers that provide support for applications running in the terminal device 110, for example. The server 130 may also receive, for example, via a network, target association information sent by the terminal device according to the two entity words input by the user and the selected estimated association relationship, and determine, according to the target association information and the processing model, a confidence level of the estimated association relationship between the two entities represented by the two entity words. For example, the server 130 may be an application server, a server of a distributed system, or a server incorporating a blockchain, for example. Alternatively, the server may be a virtual server or a cloud server.

According to an embodiment of the present application, as shown in fig. 1, the application scenario of the embodiment may further include a first database 140 and a second database 150. The first database 140 stores a large number of text paragraphs (hereinafter referred to as text paragraphs), and the second database 150 stores a pre-constructed Knowledge Graph (KG). The server 130 or the terminal device 110 may access the first database 140 and the second database 150 through a network to obtain a text segment similar to the target association information from the first database 140, and extract an associated entity word having an association relationship with the two entity words from the knowledge graph. It is understood that the first database 140 and the second database 150 may be two different databases or may be different memory partitions in the same database. In an embodiment, the first database 140 and the second database 150 may be, for example, two storage partitions in the server 130, or either of the first database 140 and the second database 150 may be a storage disk or a cloud database.

According to the embodiment of the application, similar text segments and associated entity words can be comprehensively considered, so that the characteristics of the context information and the like of the two entity words in the text segments are extracted, and the confidence degree of the estimated association relationship between the two entities is determined according to the extracted characteristics.

According to an embodiment of the present application, the terminal device 110 or the server 130 may also have a text recognition function, for example, to recognize from text an entity word and an association relationship between two entity words. And then, confidence prediction can be carried out on the association relationship between any two recognized entity words so as to verify the accuracy of the text recognition function.

It should be noted that, the method for determining the association relationship between the entities provided in the embodiments of the present application may be generally performed by the terminal device 110, or may also be performed by the server 130. Accordingly, the apparatus for determining the association relationship between entities provided in the embodiments of the present application may be generally disposed in the terminal device 110, or may also be disposed in the server 130.

It should be understood that the types of terminal devices, servers, first databases, and second databases in fig. 1 are merely illustrative. There may be any type of terminal device, server, first database, and second database, as desired for implementation.

The following describes the application scenario described in connection with fig. 1 in detail by using fig. 2 to 6 to describe the method for monitoring driving status provided in the embodiment of the present application.

Fig. 2 is a flowchart of a method for determining an association relationship between entities according to an embodiment of the present application.

As shown in fig. 2, the method 200 of determining an association relationship between entities of this embodiment may include operation S210, operation S230, operation S250, operation S270, and operation S290. The method 200 may be performed, for example, by the terminal device described above, or may be performed by a server.

In operation S210, target association information is acquired.

According to an embodiment of the present application, the target association information includes a first word representing a first entity, a second word representing a second entity, and a third word representing a target association relationship.

The first word, the second word, and the third word may each be obtained in response to an input operation by the user, for example. The first word, the second word, and the third word may constitute a main predicate-Object (SPO) triplet and use the triplet as target association information. Wherein S, O indicates two words representing entities, respectively, and P indicates a third word representing an association relationship between the two entities.

The first word and the second word may be extracted by entity recognition of the text, for example. The third word may be a preset word or may be obtained by learning semantic links between two entity words representing two entities in the text. After the first word and the second word are extracted from the text, the third word can be combined into a triplet SPO, and the triplet is used as target association information.

According to an embodiment of the present application, the target association relationship represented by the third word may include any one of the following, for example: upper and lower relationships, causal relationships, concomitant relationships, concurrency relationships, timing relationships, and the like. The target association relationship can be set according to an actual application scene, which is not limited in the application. For example, in the medical field, since the construction cost of the context between entities representing the disease is high, the target association relationship may be the context in order to construct the context of the disease.

In operation S230, a first semantic feature is determined according to the target association information and similar segments for the target association information in the preset text Duan Ku.

According to the embodiment of the application, a large number of text segments are stored in the preset text Duan Ku, and the large number of text segments can be extracted from text information acquired from various information propagation channels. The text information acquired by the information dissemination channel may include, for example, journal articles, professional books, newspaper advertisement text, news report text, and the like.

For example, the text information may be collected in advance and electronically by scanning or the like. The electronic text information is then identified using optical character recognition (Optical Character Recognition, OCR) techniques or the like to segment the text information. Each paragraph in the preset text Duan Ku may record information of one paragraph in the text information. It is understood that there may be different presets Duan Ku for different fields. In order to improve accuracy and recall of similar paragraphs and to improve accuracy of similar-paragraph-expression information, the embodiment may use a text Duan Ku including a plurality of paragraphs describing expertise in the field of the first entity and the second entity as the preset text Duan Ku.

For example, after obtaining a large number of segments, for ease of reference and recall, for example, the large number of segments may be indexed, and the large number of segments may be sequentially stored in the first database described above according to the index that is created.

According to an embodiment of the present application, operation S230 may first extract a similar segment to the target association information from the preset document Duan Ku as a similar segment to the target association information. And then, carrying out information fusion on the target associated information and the similar text segments to extract the characteristics of the context information representing the target associated information, and taking the characteristics as first semantic characteristics.

For example, a named entity recognition (Named Entity Recognition, NER) tool may be used to perform text recognition on each segment in the preset text Duan Ku, so as to obtain entity words included in each segment. The identified entity words are then extracted from the plurality of segments as similar segments including the first word and the second word. It will be appreciated that the foregoing description of the tool for text recognition is by way of example only to facilitate an understanding of the present application, and is not intended to limit the present application.

Illustratively, each of the text Duan Ku may also be processed using a term frequency-inverse text frequency (TF-IDF) statistical tool, where the importance of the first word and the second word in each of the text Duan Ku is obtained by statistics. And then extracting the text segments with the importance degrees of the first word and the second word being greater than a predetermined degree from the plurality of text segments as similar text segments.

According to the embodiment of the application, after the similar text is obtained, the target associated information and the similar text can be used as input of a first semantic feature extraction model, and the first semantic feature is obtained after processing through the semantic feature extraction model. The first semantic feature extraction model may be, for example, a Long Short-Term Memory (LSTM) model, a semantic representation model (Embeddings from Language Models, ELMo), a bi-directional transform encoder (Bidirectional Encoder Representation from Transformer, BERT), and the like.

According to the method and the device for extracting the first semantic features, the first semantic features are obtained by fusing the target associated information and the similar text, so that the extracted first semantic features can represent semantic information, lexical information, syntax information and the like of the target associated information in the text, and the richness and the accuracy of the extracted first semantic features are improved conveniently.

In operation S250, a description text for the target association information is determined according to the target association information and the preset knowledge-graph.

According to the embodiment of the application, the preset knowledge graph may be constructed according to a large amount of data in the field to which the first entity and the second entity belong, for example. For example, entities, relationships, and entity attribute information may be extracted from a large amount of data. And then, taking the entities and the entity attributes as nodes, and performing edge connection on the nodes according to the relation between the entities and the corresponding relation between the entities and the entity attributes to obtain a knowledge graph.

For example, where the first entity and the second entity belong to the medical field, the entity may include, for example, a medical term such as a disease name, a medication name, a department name, etc., and the entity attribute may include, for example, a clinical symptom of the disease, a symptom for which the medication is directed, etc. Relationships may include, for example, attribution relationships between diseases and therapeutic departments, concomitant relationships between symptoms, concurrent relationships between diseases, and the like.

According to an embodiment of the present application, operation S250 may first determine two nodes in the knowledge-graph indicating two entities represented by the first word and the second word in the target association information. And extracting the nodes connected with the two nodes in the knowledge graph to obtain the associated nodes. And finally, splicing entity words indicated by the associated nodes to form a description text.

For example, when extracting the associated node from the knowledge graph, a node indicating that the first word represents the entity may be taken as a starting point, and a node having an association relationship with the starting point may be determined as the first node according to a connection edge of the starting point. And then, determining a node with an association relation with the first node as a second node by taking the first node as a starting point according to the connecting edge of the first node. And the like, until the (m+1) th node which can establish the association relation with the node indicating the first word representation entity through the m nodes is obtained. Finally, the first to (m+1) th nodes are set as related nodes. Wherein m is a natural number, and the value of m can be set according to actual requirements, which is not limited in the application.

In operation S270, a second semantic feature is determined according to the target association information and the descriptive text.

According to an embodiment of the present application, the operation S270 may output the target association information and the description text as inputs of a second semantic feature extraction model similar to the first semantic feature extraction model described previously, to obtain a second semantic feature vector. The first semantic feature extraction model and the second semantic feature extraction model are different in that training samples adopted in training are different.

The second semantic feature extraction model may be, for example, BERT, LSTM, word representation global vector model (Global Vectors for word representation, gloVe), etc.

In operation S290, a confidence level of a target association relationship between the first entity and the second entity is determined according to the first semantic feature and the second semantic feature.

According to an embodiment of the present application, the operation S290 may, for example, use the first semantic feature and the second semantic feature as inputs of a prediction model at the same time, and obtain, via the prediction model, a confidence that two entities indicated by the first word and the second word have a target association relationship.

The predictive model may be, for example, a classifier for classification. Confidence vectors of the first word and the second word aiming at a preset number of incidence relations can be output and obtained through the classifier. The confidence vector comprises a plurality of values, each value corresponds to one incidence relation in a preset number of incidence relations, and the value of each value is the confidence coefficient of the one incidence relation between two entities indicated by the first word and the second word. And then determining the value of the corresponding target association relation in the confidence coefficient vector as a target value. And taking the value of the target value as the finally determined confidence degree of the target association relationship between the first entity and the second entity.

Compared with the technical scheme that the target association relationship is directly extracted from the text in the related art, the method for determining the association relationship between the entities can fully utilize the constructed structured information and unstructured information by extracting the similar text from the text library, obtaining the description text based on the knowledge graph, and finally integrating the similar text and the description text to determine the relationship confidence between the two entities. Therefore, the accuracy of the determined relation confidence coefficient can be effectively improved, accurate electronic knowledge can be conveniently established, and the accuracy of information recall is improved.

It is understood that the execution sequence of the foregoing operation S230 and operation S250 is not limited. For example, the operation S230 and the operation S250 may be performed simultaneously, or the operation S230 may be performed before the operation S250 is performed.

Fig. 3 is a schematic diagram of determining a first semantic feature according to an embodiment of the present application.

According to embodiments of the present application, a continuously learned semantic understanding framework (Enhanced Representation through Knowledge Integration, ernie) based on flying-oar open sources that is capable of learning a complete semantic representation may be employed to determine the first semantic features.

For example, when determining the first semantic feature, for example, the target association information may be split, a first word S representing the first entity and a third word P representing the target association relationship are converted into a question sentence through a template, and a second word O representing the second entity is replaced by a preset word, so as to obtain the target association information in which the second word is masked by the preset word. And then taking target associated information of the masked second word formed by splicing the problem statement and the preset word and the similar text as input of a pre-trained Ernie model, and outputting the first semantic feature through the Ernie model. The preset word may be, for example, any word indicating the type of the second word O, and may be, for example, "option" or "hyponym", etc. The template can be set according to the target association relationship and the actual requirement, which is not limited in the application. For example, the target association represented by the third word is a context, and the template may be "is the context of a? ". When the first word is "ill," the question sentence converted via the template may be "ill hyponym is? ".

According to the embodiment of the application, when the target associated information and the similar text segment of the second word are used as the input of the first semantic feature extraction model to obtain the first semantic feature, the output vector of the Ernie model can be subjected to dimension reduction processing, and the vector after the dimension reduction processing is used as the first semantic feature. Therefore, the expression capacity of the first semantic features can be improved, and the accuracy of the confidence coefficient determined according to the first semantic features can be improved conveniently.

Illustratively, as in the embodiment 300 shown in fig. 3, when the first semantic feature is obtained, the target association information 310 may be preprocessed to replace the second word with a preset word, so as to obtain the masked information 320 of the masked second word. The masked information 320 and the acquired similar segments 330 are then used as inputs to a first semantic vector model 340 (Ernie model). After processing via the first semantic vector model 340, an embedded vector 350 is output. Then, the embedded vector 350 is taken as an input of the attention neural network model 360, so that the embedded vector 350 is subjected to dimension reduction processing, and the vector output by the attention neural network model 360 is taken as a first semantic feature 370.

The attention neural network model 360 may be, for example, a multi-layer collaborative attention (c _o -attitution). The attention neural network model 360 may be composed of, for example, a plurality of attention mechanism layers and a fully connected layer. The embedded vectors 350 are respectively taken as inputs of the plurality of attention mechanism layers, and then the outputs of the plurality of attention mechanism layers are simultaneously input to the fully connected layer, and the first semantic features 370 are obtained through the fully connected layer outputs. It can be understood that the number of the attention mechanism layers can be set according to actual requirements, which is not limited in the present application.

According to embodiments of the present application, the Ernie model may, for example, directly employ an existing pre-training model. Or after the existing pre-training model is obtained, massive texts in the fields to which the first entity and the second entity belong can be obtained first, words or entities in the massive texts are shielded (mask) to obtain training samples, and finally model parameters in the existing training model are adjusted by adopting the training samples to obtain a first semantic feature extraction model capable of being used for extracting features.

Fig. 4 is a schematic diagram of determining a first semantic feature according to another embodiment of the present application.

According to the embodiment of the application, when the first semantic feature is determined, the text obtained from the preset text Duan Ku according to the target association information can be screened, and the text with high similarity with the target association information is used as the similar text. By the method, large noise brought by the text with low similarity as the feature extraction can be avoided, and therefore accuracy of the first semantic features obtained by extraction can be improved.

Illustratively, as in the embodiment 400 shown in fig. 4, after the target association information is obtained, the first word 411, the second word 412 and the third word 413 in the target association information may be first combined to obtain the search term. Here, for example, a character string obtained by splicing the first word 411, the second word 412, and the third word 413 in random order may be used as the search term. The preset document Duan Ku 430 with a large number of segments is then retrieved according to the retrieval statement 420. A segment including all or part of the words in the search sentence 420 is obtained from the preset text Duan Ku 430 as an alternative similar segment 440 for the target association information. Subsequently, the similarity between each of the candidate similar paragraphs 440 and the search term 420 is determined, and the paragraphs having the similarity with the search term 420 satisfying the first preset condition are determined as the similar paragraphs 450 for the target association information. Finally, a first semantic vector 480 is obtained by taking the similar segment 450 and the target associated information of the occluded second word (i.e., the occluded information 460) as input to the first semantic feature extraction model 470.

Illustratively, in determining the similarity between each candidate similar term and the search term 420, a TF-IDF tool may be used, for example, to statistically obtain the frequencies with which the candidate similar terms and the search term each include words in the word stock, to convert the candidate similar terms and the search term, respectively, into word vectors. And finally, taking the similarity between the two word vectors as the similarity between the alternative similar text and the search sentence. The similarity may be embodied in any of the following parameters: cosine similarity, jacquard similarity coefficients, szeman correlation coefficients, and the like.

Illustratively, a Binary Independent Model (BIM) based BM25 model may also be employed to calculate the similarity between each alternative similar segment and the search statement 420. The BM25 model is a classical algorithm for calculating similarity scores of query sentences and documents in the field of information indexing.

For example, the similarity satisfying the first preset condition may be, for example, a similarity greater than the first preset value, or a similarity ranked before the first predetermined position according to a large-to-small ranking of the similarities. The first preset value and the first preset position may be set according to actual requirements, which is not limited in this application. For example, the first preset value may be any value greater than 0.5, and the first predetermined position may be any position after the first bit and before the fifth bit.

Fig. 5 is a schematic diagram of determining descriptive text for target association information according to an embodiment of the present application.

According to the embodiment of the application, the description text for the target association information may be, for example, text obtained by converting SPO. The embodiment may extract a plurality of SPO groups from the knowledge-graph according to the target association information as associated SPOs associated with entities represented by the target association information. The associated SPO is then converted to a plurality of text. And finally extracting texts with higher similarity with the text converted by the target associated information from a plurality of texts obtained by converting according to the associated SPO, and taking the texts as descriptive texts aiming at the target associated information. By extracting texts with high similarity from the texts to serve as description texts and determining the second semantic features according to the description texts, the accuracy of the determined second semantic features for representing the target associated information can be improved, and the accuracy of the determined confidence level can be improved conveniently.

According to an embodiment of the present application, as in embodiment 500 shown in fig. 5, when determining a description text for target association information, a first association word having an association relationship with a first word 511 may be acquired from a preset knowledge graph 520, and a second association word having an association relationship with a second word 512 may be acquired. The plurality of candidate association information 540 is then determined according to the association between the first word 511 and the first associated word and the association between the second word 512 and the second associated word.

Illustratively, the first related word may include a word indicated by a first node connected by an edge with a node indicating the first word 511 in the preset knowledge-graph 520. The first related word may also include, for example, a word indicated by a second node connected to the first node by an edge, and the like. The first associated word includes the word indicated by the associated node determined according to m as described above. After the first related word is obtained, a plurality of SPOs may be constructed according to the node indicating the first related word and the edge connecting the plurality of nodes in the node indicating the first word. Each SPO of the plurality of SPOs serves as a first association information 531, and a plurality of first association information is finally obtained. It will be appreciated that a method similar to the method of obtaining the first related term may be used to obtain the second related term and obtain the second related information 532. Finally, the first association information 531 and the second association information 532 are aggregated to obtain a plurality of candidate association information 540.

According to the embodiment of the application, after the candidate associated information is obtained, the description text of each candidate associated information and the description text of the target associated information can be determined according to the preset text template. And finally, determining the description text of which the similarity with the description text of the target associated information meets a second preset condition in the description text of each of the plurality of candidate associated information, and taking the description text as the description text of the target associated information.

Illustratively, as shown in fig. 5, S, P, O in the target association information 510 may be substituted into the text template 550 to obtain the target description text 560. S, P, O in each alternative association information 540 is substituted into the text template 550 to obtain alternative description text 570. The similarity between each candidate description text 570 and the target description text 560 is then calculated, resulting in a similarity for each candidate description text 570. Finally, the candidate description text for which the similarity satisfying the second preset condition is used as the finally determined description text 580 for the target association information.

For example, the similarity satisfying the second preset condition may be a similarity greater than the second preset value, or a similarity ranked before the second predetermined position according to a large-to-small ranking of the similarities. The second preset value and the second preset position may be set according to actual requirements, which is not limited in this application. For example, the second preset value may be any value greater than 0.6, and the second predetermined position may be any position after the second bit and before the eighth bit. It is understood that, according to the actual requirement, the second preset condition and the first preset condition described above may be the same condition, for example.

For example, the preset text template may be set according to actual requirements, for example, the preset text template may be "a has C association with B". When the target associated information is 'illness-upper and lower relationship-cold', the description text 'illness has upper and lower relationship with cold' can be obtained by substituting the target associated information into a preset text template. It should be understood that the above-mentioned preset text templates are merely examples to facilitate understanding of the present application, which is not limited thereto.

According to the embodiment of the application, after the description text aiming at the target associated information is obtained, the operation of replacing the second word by the preset word as described above can be performed on the target associated information, so that the target associated information of the masked second word is obtained. And finally, taking the target associated information of the second word and the description text aiming at the target associated information as the input of a second semantic feature extraction model, and outputting the second semantic feature after processing through the second semantic feature model. It will be appreciated that a second semantic feature may be obtained in a similar way to the first semantic feature, the two ways differing in that the text from which the second semantic feature is obtained is descriptive of the structured knowledge extracted from the knowledge graph.

According to the embodiment of the application, after the first semantic feature and the second semantic feature are obtained, the first semantic feature and the second semantic feature can be input into a prediction model at the same time, probability vectors are output after being processed through the prediction model, and the probability vectors comprise probability values of the masked second words serving as all entity words in a preset entity word bank. Finally, a probability value in the probability vector indicating that the masked second word is an entity word of the second entity may be determined, and the probability value is used as a confidence level that the first entity and the second entity have a target association relationship.

For example, the preset entity word library may include entity words extracted from expertise describing the areas to which the first entity and the second entity belong, and the entity word library includes the second words. For different fields, different entity word libraries can be pre-constructed. The determined probability value is a probability value of the probability vector for describing the entity word of the second entity in the preset entity word stock.

The prediction model may be, for example, a classification model, which may include, for example, a fully connected layer and an output layer. The full connection layer is used for fusing the first semantic features and the second semantic features, and the output layer is used for outputting probability vectors described above. Illustratively, the classification model may further include a softmax activation layer for normalizing the output probability vector.

Fig. 6 is a schematic diagram of determining an association relationship between entities according to an embodiment of the present application.

In this embodiment 600, the overall process of determining the association relationship between entities may include the process described below, as shown in fig. 6, according to an embodiment of the present application.

After the target associated information is acquired, the second word is used as an option 612, the first word and the third word are converted into questions 613 according to the template, the first word, the second word and the third word are spliced to form a search sentence, and a similar text 611 aiming at the target associated information is obtained through search. Then, the text 611 is taken as evidence, the option 612 and the question 613 are spliced and then input into the first semantic feature extraction model 620, and an embedded vector capable of expressing the text, the option and the question is obtained after being processed by the first semantic feature extraction model 620, and the embedded vector can be understood as a vector spliced by the Wen Duanbiao vector 631, the option expression vector 632 and the question expression vector 633. The embedded vector output by the first semantic feature extraction model 620 is then input into the attention neural network model 640 and output as the first semantic feature.

Meanwhile, the related words for the first word and the related words for the second word may be obtained from the knowledge graph 650, so as to obtain a plurality of combinations of S and O, where S and O are two words respectively indicated by two nodes having connecting edges in the knowledge graph. According to the association relationship indicated by the S and O and the edge connecting the two nodes, a plurality of SPOs may be obtained, so as to splice and form the association string 662. The associated string 662 is then converted to descriptive text using a preset text template. Next, target related information of the masked second word obtained by masking the second word is inputted as a target character string 661 and a descriptive text into the second semantic feature extraction model 670, and the second semantic feature is outputted after processing via the second semantic feature extraction model 670.

Finally, the first semantic feature and the second semantic feature are input into the prediction model 680 at the same time, and a confidence 690 with a target association relationship between the first entity and the second entity can be obtained according to the output of the prediction model 680.

According to the method for determining the association relationship between the entities, the existing structured knowledge and unstructured knowledge can be fully utilized to determine the association relationship confidence, and therefore the method can be used for recall of the association relationship between the entities. Compared with the prior art, the method and the device only recall the association relation according to the text, the accuracy of the association relation of recall can be improved. By applying the method to the prediction of the upper and lower relationship between the entities in the medical field, the construction cost of the upper and lower relationship of the diseases can be effectively reduced, and the intelligent electronic management of knowledge in the medical field is facilitated.

According to the embodiment of the application, after the confidence coefficient with the target association relationship between the first entity and the second entity is obtained, if the confidence coefficient is higher, the knowledge graph can be supplemented according to the target association information, so that more comprehensive and richer structured knowledge can be obtained in the subsequent confidence coefficient prediction, and the accuracy of the subsequent confidence coefficient prediction is improved.

According to an embodiment of the present application, the operation of determining the association relationship between the entities may further determine whether the confidence coefficient of the target association relationship between the first entity and the second entity is greater than a preset confidence coefficient. If the target correlation information is larger than the preset confidence coefficient, supplementing a preset knowledge graph according to the target correlation information.

For example, when the preset knowledge-graph is supplemented, it may be determined whether there are a node indicating the first word and a node indicating the second word in the preset knowledge-graph. If the node indicating the first word is not included, adding a node in the knowledge graph for indicating the first word. Similarly, a node indicating the second word may be added when the node indicating the second word is not included. After determining that the preset knowledge graph has the node indicating the first word and the node indicating the second word, determining whether a connecting edge exists between the node indicating the first word and the node indicating the second word, and if not, adding the edge between the two nodes to complete the supplement of the preset knowledge graph.

The preset confidence level may be determined, for example, from the predicted confidence level obtained by using the test sample after training each model. For example, the relationship between the confidence and the label of the test sample may be set such that the preset confidence value may enable the confidence of the test sample for the preset proportion to be greater than the preset confidence. The preset proportion may be a larger value such as 80%, 85%, 90%, etc., and the preset confidence may be a value greater than 0.5 such as 0.6, 0.8, etc. It should be understood that the foregoing values of the preset proportion and the preset confidence coefficient and the determining method are merely examples to facilitate understanding of the present application, which is not limited in this application.

According to the embodiment of the application, after the confidence that the target association relationship between the first entity and the second entity is determined to be greater than the preset confidence, confidence accuracy assessment can be performed by a professional, for example. And when the evaluation result is that the confidence coefficient is reliable, supplementing a preset knowledge graph according to the target association relation.

Based on the method for determining the association relationship between the entities, which is described above, the application also provides a device for determining the association relationship between the entities. The apparatus provided in this application will be described in detail below in conjunction with fig. 7.

Fig. 7 is a block diagram of an apparatus for determining association relationships between entities according to an embodiment of the present application.

As shown in fig. 7, the apparatus 700 for determining an association relationship between entities of this embodiment may include an information acquisition module 710, a first feature determination module 730, a text determination module 750, a second feature determination module 770, and a confidence determination module 790.

The information acquisition module 710 is configured to acquire target association information, where the target association information includes a first word indicating a first entity, a second word indicating a second entity, and a third word indicating a target association relationship. In an embodiment, the information obtaining module 710 may be used to perform the operation S210 described above, which is not described herein.

The first feature determining module 730 is configured to determine a first semantic feature according to the target associated information and a similar segment of the preset text Duan Ku for the target associated information. In an embodiment, the first feature determining module 730 may be used to perform the operation S230 described above, which is not described herein.

The text determining module 750 is configured to determine a description text for the target association information according to the target association information and a preset knowledge graph. In an embodiment, the text determining module 750 may be used to perform the operation S250 described above, which is not described herein.

The second feature determination module 770 is configured to determine a second semantic feature based on the target association information and the descriptive text. In an embodiment, the second feature determining module 770 may be used to perform the operation S270 described above, which is not described herein.

The confidence determining module 790 is configured to determine, according to the first semantic feature and the second semantic feature, a confidence level of a target association relationship between the first entity and the second entity. In an embodiment, the confidence determining module 790 may be used to perform the operation S290 described above, which is not described herein.

According to embodiments of the present application, an electronic device and a readable storage medium are also provided.

As shown in fig. 8, a block diagram of an electronic device is used to implement a method for determining an association relationship between entities according to an embodiment of the present application.

Electronic devices are intended to represent various forms of digital computers, such as, for example, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframes, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processing, cellular telephones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are meant to be exemplary only, and are not meant to limit implementations of the application described and/or claimed herein.

As shown in fig. 8, the electronic device 800 includes: one or more processors 801, memory 802, and interfaces for connecting the components, including high-speed interfaces and low-speed interfaces. The various components are interconnected using different buses and may be mounted on a common motherboard or in other manners as desired. The processor may process instructions executing within the electronic device, including instructions stored in or on memory to display graphical information of the GUI on an external input/output device, such as a display device coupled to the interface. In other embodiments, multiple processors and/or multiple buses may be used, if desired, along with multiple memories and multiple memories. Also, multiple electronic devices may be connected, each providing a portion of the necessary operations (e.g., as a server array, a set of blade servers, or a multiprocessor system). One processor 801 is illustrated in fig. 8.

Memory 802 is a non-transitory computer-readable storage medium provided herein. The memory stores instructions executable by the at least one processor to cause the at least one processor to perform the method for determining an association relationship between entities provided herein. The non-transitory computer readable storage medium of the present application stores computer instructions for causing a computer to perform the method of determining an association between entities provided by the present application.

The memory 802 is used as a non-transitory computer readable storage medium, and may be used to store non-transitory software programs, non-transitory computer executable programs, and modules, such as program instructions/modules (e.g., the information acquisition module 710, the first feature determination module 730, the text determination module 750, the second feature determination module 770, and the confidence determination module 790 shown in fig. 7) corresponding to a method for determining an association relationship between entities in an embodiment of the present application. The processor 801 executes various functional applications of the server and data processing, that is, a method of determining association between entities in the above-described method embodiment, by executing non-transitory software programs, instructions, and modules stored in the memory 802.

Memory 802 may include a storage program area that may store an operating system, at least one application program required for functionality, and a storage data area; the storage data area may store data created according to the use of an electronic device implementing a method of determining an association relationship between entities, etc. In addition, memory 802 may include high-speed random access memory, and may also include non-transitory memory, such as at least one magnetic disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, memory 802 may optionally include memory located remotely from processor 801, which may be connected via a network to an electronic device implementing the method of determining associations between entities. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.

The electronic device 800 implementing the method of determining an association relationship between entities may further include: an input device 803 and an output device 804. The processor 801, memory 802, input devices 803, and output devices 804 may be connected by a bus or other means, for example in fig. 8.

The input device 803 may receive input numeric or character information and generate key signal inputs related to user settings and function control of an electronic device implementing the method of determining associations between entities, such as a touch screen, a keypad, a mouse, a track pad, a touch pad, a pointer stick, one or more mouse buttons, a track ball, a joystick, etc. input devices. The output device 804 may include a display apparatus, auxiliary lighting devices (e.g., LEDs), and haptic feedback devices (e.g., vibration motors), among others. The display device may include, but is not limited to, a Liquid Crystal Display (LCD), a Light Emitting Diode (LED) display, and a plasma display. In some implementations, the display device may be a touch screen.

Various implementations of the systems and techniques described here can be realized in digital electronic circuitry, integrated circuitry, application specific ASIC (application specific integrated circuit), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include: implemented in one or more computer programs, the one or more computer programs may be executed and/or interpreted on a programmable system including at least one programmable processor, which may be a special purpose or general-purpose programmable processor, that may receive data and instructions from, and transmit data and instructions to, a storage system, at least one input device, and at least one output device.

These computing programs (also referred to as programs, software applications, or code) include machine instructions for a programmable processor, and may be implemented in a high-level procedural and/or object-oriented programming language, and/or in assembly/machine language. As used herein, the terms "machine-readable medium" and "computer-readable medium" refer to any computer program product, apparatus, and/or device (e.g., magnetic discs, optical disks, memory, programmable Logic Devices (PLDs)) used to provide machine instructions and/or data to a programmable processor, including a machine-readable medium that receives machine instructions as a machine-readable signal. The term "machine-readable signal" refers to any signal used to provide machine instructions and/or data to a programmable processor.

To provide for interaction with a user, the systems and techniques described here 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 a user; and a keyboard and pointing device (e.g., a mouse or trackball) by which a user can provide input to the computer. Other kinds of devices may also be used to provide for interaction with a user; for example, feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form, including acoustic input, speech input, or tactile input.

The systems and techniques described here can be implemented in a computing system that includes a background 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 user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such background, 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: local Area Networks (LANs), wide Area Networks (WANs), and the internet.

The computer system may include a client and a server. The client and server are typically 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.

It should be appreciated that various forms of the flows shown above may be used to reorder, add, or delete steps. For example, the steps described in the present application may be performed in parallel, sequentially, or in a different order, provided that the desired results of the technical solutions disclosed in the present application can be achieved, and are not limited herein.

The above embodiments do not limit the scope of the application. It will be apparent to those skilled in the art that various modifications, combinations, sub-combinations and alternatives are possible, depending on design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present application are intended to be included within the scope of the present application.

Claims

1. A method of determining an association between entities, comprising:

acquiring target association information, wherein the target association information comprises a first word representing a first entity, a second word representing a second entity and a third word representing a target association relationship;

determining a first semantic feature according to the target associated information and similar text segments aiming at the target associated information in preset text Duan Ku;

determining a description text aiming at the target associated information according to the target associated information and a preset knowledge graph;

determining a second semantic feature according to the target associated information and the description text; and

determining the confidence degree of the target association relationship between the first entity and the second entity according to the first semantic feature and the second semantic feature;

Wherein determining the first semantic feature comprises:

replacing the second word with a preset word to obtain target associated information of the second word shielded by the preset word;

the target associated information of the second word shielded by the preset word and the similar text segment are used as input of a first semantic feature extraction model, and the first semantic feature is obtained;

wherein determining the second semantic feature comprises:

taking target associated information of the second word shielded by a preset word and a description text aiming at the target associated information as input of a second semantic feature extraction model to obtain the second semantic feature;

wherein determining the confidence level of the target association relationship between the first entity and the second entity comprises:

taking the first semantic features and the second semantic features as inputs of a prediction model, and outputting masked second words as probability vectors of all entity words in a preset entity word bank; and

and determining the probability of the second masked word as the entity word of the second entity in the probability vector, and taking the probability as the confidence degree of the target association relationship between the first entity and the second entity.

2. The method of claim 1, wherein the obtaining the first semantic feature with the object associated information and the similar segments obscured by the second word as inputs to a first semantic feature extraction model comprises:

Taking the target associated information and the similar text segment which are shielded from the second word as the input of a first semantic feature extraction model, and outputting an embedded vector; and

and taking the embedded vector as an input of an attention neural network model, and outputting the first semantic feature.

3. The method of claim 1 or 2, wherein determining the first semantic feature further comprises:

combining the first word, the second word and the third word to obtain a search sentence;

searching the preset text Duan Ku according to the search statement to obtain an alternative similar text segment aiming at the target associated information; and

and determining the text segments, of which the similarity with the search statement meets a first preset condition, in the alternative similar text segments as similar text segments aiming at the target associated information.

4. The method of claim 1, wherein determining descriptive text for the target association information comprises:

acquiring a first association word with an association relation with the first word from the preset knowledge graph, and acquiring a second association word with an association relation with the second word;

determining a plurality of candidate association information according to the association relation between the first word and the first association word and the association relation between the second word and the second association word;

Determining the description text of the target associated information and the description text of each of the plurality of candidate associated information according to a preset text template; and

and determining the description text of which the similarity with the description text of the target associated information meets a second preset condition in the description text of each of the plurality of candidate associated information, and taking the description text as the description text aiming at the target associated information.

5. The method of claim 1, further comprising:

and supplementing the preset knowledge graph according to the target association information under the condition that the confidence coefficient of the target association relationship between the first entity and the second entity is larger than the preset confidence coefficient.

6. The method of claim 1, wherein the preset segment library comprises: a plurality of paragraphs describing expertise in the art to which the first entity and the second entity pertain.

7. An apparatus for determining an association between entities, comprising:

the information acquisition module is used for acquiring target association information, wherein the target association information comprises a first word representing a first entity, a second word representing a second entity and a third word representing a target association relationship;

the first feature determining module is used for determining first semantic features according to the target associated information and similar text segments aiming at the target associated information in preset text Duan Ku;

The text determining module is used for determining a description text aiming at the target associated information according to the target associated information and a preset knowledge graph;

the second feature determining module is used for determining a second semantic feature according to the target associated information and the description text; and

the confidence degree determining module is used for determining the confidence degree of the target association relationship between the first entity and the second entity according to the first semantic feature and the second semantic feature;

wherein the first feature determination module is further configured to:

wherein the second feature determination module is further configured to:

wherein the confidence determination module is further configured to:

8. An electronic device, comprising:

at least one processor; and

a memory communicatively coupled to the at least one processor; wherein,

the memory stores instructions executable by the at least one processor to enable the at least one processor to perform: the method of any one of claims 1 to 6.

9. A non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are configured to cause the computer to perform: the method of any one of claims 1 to 6.