CN111782763A

CN111782763A - Information retrieval method based on voice semantics and related equipment thereof

Info

Publication number: CN111782763A
Application number: CN202010440491.7A
Authority: CN
Inventors: 胡逸天; 李琪; 孟令成; 魏俊勇; 游志刚
Original assignee: Ping An Technology Shenzhen Co Ltd
Current assignee: Ping An Technology Shenzhen Co Ltd
Priority date: 2020-05-22
Filing date: 2020-05-22
Publication date: 2020-10-16
Also published as: WO2021135439A1

Abstract

The invention relates to artificial intelligence, and provides an information retrieval method based on voice semantics, which comprises the following steps: acquiring an input user query statement; replacing instance entities in the user query sentences with concept entities to obtain template query sentences; calculating the similarity between the template query sentence and each stock query sentence in the question sentence corpus; determining an inventory query statement matched with the template query statement and a retrieval logic formula corresponding to the inventory query statement according to the calculated similarity; updating the retrieval logic formula according to the example entity; generating a retrieval tree based on the updated retrieval logic formula; and carrying out information retrieval on the database according to the retrieval tree, and displaying the retrieved answer information. The invention can improve the accuracy of information retrieval.

Description

Information retrieval method based on voice semantics and related equipment thereof

Technical Field

The present application relates to artificial intelligence, and in particular, to a method and an apparatus for retrieving information based on speech semantics, a computer device, and a storage medium.

Background

With the development of artificial Intelligence, intelligent questions and answers are more and more widely applied to a BI (Business Intelligence) system. The intelligent question-answering relates to semantic analysis, voice recognition and the like in the field of artificial intelligence, and generally, a computer acquires a query instruction of a user for a certain thing, analyzes the query instruction, retrieves corresponding answer information and displays the answer information. In the intelligent question answering, the query content and the expression mode of the user are various and are difficult to limit, so that the query intention of the user is accurately understood, answer information is accurately and quickly retrieved, and the intelligent question answering method is a key for realizing the intelligent question answering.

In order to deal with uncertain input of a user, the conventional intelligent question-answering technology generally adopts keyword capture, namely, retrieval is carried out according to keywords in a query sentence of the user. However, it is difficult to retrieve answer information satisfying the user's intention by capturing only a complete question sentence input by the user through a keyword, and the accuracy of information retrieval is low.

Disclosure of Invention

An embodiment of the present application aims to provide an information retrieval method, an information retrieval device, a computer device, and a storage medium based on voice semantics, so as to solve the problem of low accuracy of information retrieval.

In order to solve the above technical problem, an embodiment of the present application provides an information retrieval method based on voice semantics, which adopts the following technical solutions:

acquiring an input user query statement;

analyzing the user query statement, and replacing an instance entity in the user query statement with a concept entity to obtain a template query statement; the concept entity is an entity type to which the instance entity belongs; calculating the similarity between the template query statement and each stock query statement in the question corpus;

determining an inventory query statement matched with the template query statement and a retrieval logic formula corresponding to the inventory query statement according to the calculated similarity;

updating a retrieval logic formula according to the instance entity;

generating a retrieval tree based on the updated retrieval logic formula;

and performing information retrieval on the database according to the retrieval tree, and displaying the retrieved answer information.

Further, the step of analyzing the user query statement, replacing an instance entity in the user query statement with a concept entity, and obtaining a template query statement specifically includes:

identifying an instance entity in the user query statement, and determining an entity type of the instance entity through semantic identification to obtain a concept entity representing the entity type;

querying a standard entity corresponding to the instance entity from a standard entity list;

and replacing the instance entity in the user query statement with the concept entity to obtain a template query statement, and storing the instance entity and the standard entity in an associated manner.

Further, the step of updating the search logic formula according to the instance entity specifically includes:

acquiring a standard entity stored in association with the instance entity;

and replacing the standard entity in the retrieval logic formula with the acquired standard entity.

Further, the step of generating the search tree based on the updated search logical expression specifically includes:

identifying a retrieval type of the retrieval logic formula;

when the retrieval type is single triple single medium retrieval, generating a single triple single medium retrieval tree;

and when the retrieval type is multi-triple multi-media retrieval, generating a multi-triple multi-media retrieval tree.

Further, the step of retrieving information from the database according to the retrieval tree and displaying the retrieved answer information specifically includes:

performing depth-first traversal on the retrieval tree to determine a retrieval strategy corresponding to the retrieval tree, and determining an information type based on the retrieval strategy;

performing information retrieval on the database according to the retrieval strategy to obtain answer information;

displaying the answer information according to the information type;

and uploading the answer information to a block chain.

Further, the step of displaying the answer information according to the information type specifically includes:

when the information type is single entity single attribute or entity relationship, the answer information is displayed by text;

when the information type is single entity multi-attribute or multi-entity single attribute, displaying the answer information by a histogram;

and when the information type is the attribute variation trend, displaying the answer information by a line graph.

Further, after the steps of retrieving information from the database according to the retrieval tree and displaying the retrieved answer information, the method further includes:

setting the template query statement as an inventory query statement to update the question corpus;

and correlating the newly added stock query sentences in the question corpus with the updated retrieval logic.

In order to solve the above technical problem, an embodiment of the present application further provides an information retrieval apparatus based on voice semantics, including:

the sentence acquisition module is used for acquiring an input user query sentence;

the entity replacing module is used for analyzing the user query statement, replacing an instance entity in the user query statement with a concept entity and obtaining a template query statement; the concept entity is an entity type to which the instance entity belongs;

the similarity calculation module is used for calculating the similarity between the template query statement and each stock query statement in the question corpus;

the statement determination module is used for determining the stock query statement matched with the template query statement and the retrieval logic formula corresponding to the stock query statement according to the calculated similarity;

the logic formula updating module is used for updating the retrieval logic formula according to the example entity;

the retrieval tree generation module is used for generating a retrieval tree based on the updated retrieval logic formula;

and the information retrieval module is used for retrieving information from the database according to the retrieval tree and displaying the retrieved answer information.

In order to solve the above technical problem, an embodiment of the present application further provides a computer device, which includes a memory and a processor, where the memory stores a computer program, and the processor implements the steps of the information retrieval method based on voice semantics when executing the computer program.

In order to solve the above technical problem, an embodiment of the present application further provides a computer-readable storage medium, where a computer program is stored, and when the computer program is executed by a processor, the steps of the information retrieval method based on the voice semantics are implemented.

Compared with the prior art, the embodiment of the application mainly has the following beneficial effects: the method comprises the steps of replacing instance entities in obtained user query sentences to obtain template query sentences, removing the user query sentences through the template query sentences in an individualized mode, calculating the similarity between the template query sentences and various inventory query sentences in a corpus, and determining inventory query sentences matched with the user query sentences and retrieval logic formulas thereof according to the similarity so as to improve the processing capacity of the user query sentences in various forms and ensure the accuracy and the usability of information retrieval; the retrieval tree is generated according to the retrieval logic formula, the retrieval tree indicates how to retrieve information from a plurality of databases, and the information aimed by the user query sentence can be accurately retrieved from the databases by retrieval based on the retrieval tree, so that the accuracy of information retrieval is further ensured.

Drawings

In order to more clearly illustrate the solution of the present application, the drawings needed for describing the embodiments of the present application will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present application, and that other drawings can be obtained by those skilled in the art without inventive effort.

FIG. 1 is an exemplary system architecture diagram in which the present application may be applied;

FIG. 2 is a flow diagram of one embodiment of a method for information retrieval based on speech semantics according to the present application;

FIG. 3 is a diagram of a single triple single media retrieval tree in one embodiment;

FIG. 4 is a diagram of a multi-triplet multimedia retrieval tree in one embodiment;

FIG. 5 is a flowchart of one embodiment of step S207 of FIG. 2;

FIG. 6 is a diagram illustrating answer information in a histogram, according to an embodiment;

FIG. 7 is a diagram illustrating answer information in a line chart, according to an embodiment;

FIG. 8 is a schematic diagram of an embodiment of an information retrieval device based on speech semantics according to the present application;

FIG. 9 is a schematic block diagram of one embodiment of a computer device according to the present application.

Detailed Description

Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs; the terminology used in the description of the application herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the application; the terms "including" and "having," and any variations thereof, in the description and claims of this application and the description of the above figures are intended to cover non-exclusive inclusions. The terms "first," "second," and the like in the description and claims of this application or in the above-described drawings are used for distinguishing between different objects and not for describing a particular order.

Reference herein to "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the application. The appearances of the phrase in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments. It is explicitly and implicitly understood by one skilled in the art that the embodiments described herein can be combined with other embodiments.

In order to make the technical solutions better understood by those skilled in the art, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings.

As shown in fig. 1, the system architecture 100 may include

terminal devices

101, 102, 103, a network 104, and a server 105. The network 104 serves as a medium for providing communication links between the

terminal devices

101, 102, 103 and the server 105. Network 104 may include various connection types, such as wired, wireless communication links, or fiber optic cables, to name a few.

The user may use the

terminal devices

101, 102, 103 to interact with the server 105 via the network 104 to receive or send messages or the like. The

terminal devices

101, 102, 103 may have various communication client applications installed thereon, such as a web browser application, a shopping application, a search application, an instant messaging tool, a mailbox client, social platform software, and the like.

The

terminal devices

101, 102, 103 may be various electronic devices having a display screen and supporting web browsing, including but not limited to smart phones, tablet computers, e-book readers, MP3 players (Moving picture experts Group Audio Layer III, mpeg compression standard Audio Layer 3), MP4 players (Moving picture experts Group Audio Layer IV, mpeg compression standard Audio Layer 4), laptop portable computers, desktop computers, and the like.

The server 105 may be a server providing various services, such as a background server providing support for pages displayed on the

terminal devices

101, 102, 103.

It should be noted that the information retrieval method based on the voice semantics provided by the embodiment of the present application is generally executed by a server, and accordingly, the information retrieval device based on the voice semantics is generally disposed in the server.

It should be understood that the number of terminal devices, networks, and servers in fig. 1 is merely illustrative. There may be any number of terminal devices, networks, and servers, as desired for implementation.

With continuing reference to FIG. 2, a flow diagram of one embodiment of a method for speech semantic based information retrieval according to the present application is shown. The information retrieval method based on the voice semantics comprises the following steps:

step 201, acquiring an input user query statement.

In this embodiment, an electronic device (for example, a server shown in fig. 1) on which the information retrieval method based on the voice semantics operates may communicate with the terminal through a wired connection manner or a wireless connection manner. It should be noted that the wireless connection means may include, but is not limited to, a 3G/4G connection, a WiFi connection, a bluetooth connection, a WiMAX connection, a Zigbee connection, a uwb (ultra wideband) connection, and other wireless connection means now known or developed in the future.

The user query statement may be a query statement input by a user.

Specifically, a user inputs a user query sentence in the information retrieval page in a text form, and the terminal displaying the information retrieval page sends the user query sentence to the server. The user can also ask questions through voice query, and the input voice is converted into a user query sentence in a text form through voice recognition. A user can perform voice query through an input method supporting voice input; the information retrieval page can also call an application program interface provided by a third party to convert the voice, or the terminal can send the voice to the server, and the server can convert the voice into characters.

Step 202, analyzing a user query statement, replacing an instance entity in the user query statement with a concept entity to obtain a template query statement; a conceptual entity is an entity type to which an instance entity belongs.

Wherein, the instance entity can be a named entity in the user query statement; the entity type may be a category attribute of the instance entity.

Specifically, the server parses the user query statement to identify an instance entity in the user query statement, and determines an entity type of the instance entity through semantic identification to determine a concept entity corresponding to the instance entity. The intent of the user query statement is to instruct the server to retrieve information related to the instance entity. Example entities may be named entities in a user query statement, including names of people, places, organizations, numbers, dates, currency, addresses, proper nouns, and so forth. The server replaces an instance entity in the user query statement with the concept entity to obtain a template query statement; at the same time, the server retains the replaced instance entities for subsequent operations to assemble new search logic.

For example, the user query statement is "when M is true? "the server identifies instance entity" M ", assuming" M "is an abbreviation for a company belonging to a certain organization, and the conceptual entity corresponding to" M "is" organization ". Replace "M" with "mechanism", get the template query statement "< mechanism > is true? ", while retaining the replaced instance entity" M ".

In one embodiment, the step of analyzing the user query statement, replacing an instance entity in the user query statement with a concept entity, and obtaining the template query statement specifically includes: identifying instance entities in the user query statement, and determining entity types of the instance entities through semantic identification to obtain concept entities representing the entity types; inquiring a standard entity corresponding to the example entity from the standard entity list; and replacing the instance entity in the user query statement with the concept entity to obtain a template query statement, and storing the instance entity and the standard entity in an associated manner.

Specifically, the server parses the user query statement, identifies a Named entity in the user query statement through Named entity identification (NER, also called proper name identification), takes the identified Named entity as an instance entity, and determines an entity type to which the instance entity belongs through semantic identification to determine a concept entity representing the entity type.

The instance entity in the user query statement may be abbreviated or unnormalized, while the information stored in the database exists in a standard descriptive manner. The standard entities are stored in a standard entity list.

The server obtains a pre-established standard entity list, and searches a standard entity corresponding to the instance entity in the standard entity list through fuzzy matching. And the server replaces the instance entity in the user query statement with the concept entity to obtain the template query statement, and stores the standard entity and the instance entity in an entity association table in an associated manner. The entity association table is used for storing the instance entities in the user query statement and the corresponding standard entities.

For example, the instance entity in the user query statement is "M", and "M" is short, and what is stored in the database is called "M shares company"; "M shares company" is a standard entity corresponding to "M". After template replacement, the user query statement becomes "< when the organization > is true? "and stores the instance entity" M "in association with the standard entity" M corporation "for subsequent assembly of a new search logic.

In the embodiment, an instance entity in a user query statement is identified, an entity type of the instance entity is determined, and a concept entity representing the entity type is determined; the standard entity corresponding to the instance entity is inquired, the instance entity in the user inquiry statement is replaced by the concept entity, the user inquiry statement is converted from diversification to standardization, personalized information in the user inquiry statement is reduced, the follow-up inquiry of stock inquiry statements through similarity is facilitated, and the accuracy of information retrieval is ensured; the instance entity and standard entity associations are stored for subsequent assembly of a new logical search.

Step 203, calculating the similarity between the template query sentence and each stock query sentence in the question sentence corpus.

Wherein the inventory query statement may be a statement stored in a question corpus; the search logic is another embodiment of an inventory query statement that is used to build a search tree and characterize the search logic. The inventory query statement corresponds to a search logical expression, and multiple inventory query statements may correspond to the same search logical expression.

Specifically, the server accesses a question corpus and converts each stock query statement and template query statement in the question corpus into a sentence vector. And calculating the similarity between the sentence vector of the template query sentence and the sentence vector of each stock query sentence through a preset similarity formula.

In one embodiment, the similarity may be calculated by cosine similarity, edit distance, jaccard coefficient, TFIDF coefficient (adding inverse document frequency IDF on the basis of the word frequency TF), and the like, where the cosine similarity is calculated according to the following formula (1):

where, quessiona may be a sentence vector of a template query statement, and quessionb may be a sentence vector of an inventory query statement.

And step 204, determining the inventory query statement matched with the template query statement and a retrieval logic formula corresponding to the inventory query statement according to the calculated similarity.

Specifically, the server compares the calculated similarity with a preset similarity threshold, and screens the inventory query statement corresponding to the maximum similarity from the similarities greater than the similarity threshold as the inventory query statement matched with the template query statement.

The server inquires a retrieval logic formula corresponding to the stock inquiry statement from the question corpus and establishes a mapping relation between the user inquiry statement, the template inquiry statement, the stock inquiry statement and the retrieval logic formula.

For example, the user query statement is "when M is true? ", when a template query statement" < organization > holds after instance entity replacement? ". The server queries, through similarity, that the stock query statement matched with the template query statement is "< establishment date of organization? ", and its corresponding search logical expression < V1: Unary (class ═ mechanism ', value ═ N') > < a: Binary (V1, registration date, a.

The server implements the following mapping relationships: when M is true? -a template query statement? Inventory query statement? Search logic formula: < V1: Unary (class ═ mechanism ', value ═ N') > < a: Binary (V1, registration date, a.

Step 205, updating the retrieval logic according to the instance entity.

Specifically, there is a standard entity in the retrieval logic formula, which is the standard entity retrieved in the previous retrieval. The server needs to update the retrieval logic formula according to the instance entity in the user query statement during the retrieval.

In one embodiment, the step of updating the search logic formula according to the instance entity specifically includes: acquiring a standard entity stored in association with the instance entity; and replacing the standard entity in the retrieval logic formula with the acquired standard entity.

Specifically, in the retrieval logic formula corresponding to the inventory query statement, the position of "N" is a variable, other parts in the retrieval logic formula are not variable, and "N" at value is variable. The former search may be "the established date of N", so the search logical expression after the search is "N", and the search is performed for "M" this time, so "N" is replaced by "M limited company", otherwise the generated search tree is for "N".

And the server acquires the standard entity associated with the instance entity in the user query statement from the entity association table, and replaces the standard entity in the retrieval logic formula with the acquired standard entity.

In the embodiment, the standard entity in the retrieval logic formula is replaced by the standard entity associated with the instance entity in the user query statement, and the replaced retrieval logic formula aims at the retrieval, so that the information related to the retrieval can be accurately acquired from the database.

And step 206, generating a retrieval tree based on the updated retrieval logic formula.

Wherein the retrieval tree may be a binary tree based storage structure.

Specifically, the retrieval logic formula notes the information that needs to be retrieved last in each retrieval. And when constructing the retrieval tree based on the updated retrieval logic formula, taking the information needing to be retrieved finally as a root node. Different retrieval formulas can correspond to different retrieval types, different retrieval types correspond to different retrieval tree structures, and the server fills the retrieval tree structures according to the retrieval logic formulas to generate the retrieval tree.

The search tree may be a binary tree, each internal node in the branches of the binary tree is information to be searched, the left and right branches of the node are search conditions, and the root node of the binary tree is information to be searched finally.

In one embodiment, the step of generating the search tree based on the updated search logical expression specifically includes: identifying a retrieval type of the retrieval logic formula; when the retrieval type is single triple single medium retrieval, generating a single triple single medium retrieval tree; and when the retrieval type is multi-triple multi-media retrieval, generating a multi-triple multi-media retrieval tree.

Wherein the retrieval type can be a type of retrieval, and is determined by the attribute of the retrieved object and the storage medium accessed during the retrieval; the storage medium may be a database storing information.

In particular, different retrievals may correspond to different retrieval types, including single triple single media retrieval and multi-triple multi-media retrieval.

When the retrieval type is single triple single medium retrieval, a single triple single medium retrieval tree is generated. For example, when a single entity single attribute value is retrieved, the logical form of the retrieval tree is < entry ═ E > < attr ═ a > < attr _ value? < CHEM > A

Where E denotes a standard entity, attr denotes an attribute of the standard entity, here attribute a, and attr _ value denotes an attribute value of attribute a.

The search tree structure includes a root node "attribute value", a left leaf node "entity E", and a right leaf node "attribute a", and is searched only once within a single storage medium.

For example, when retrieving the registration date of M, the logical form of the retrieval tree is:

< entry? < CHEM > A

The corresponding search tree structure includes a root node "attribute value", a left-leaf node "M shares limited company" and a right-leaf node "registration date", and the generated search tree is shown in fig. 3.

And when the retrieval type is multi-triple multi-media retrieval, generating a multi-triple multi-media retrieval tree. The single triple single medium retrieval tree and the multi-triple multi-medium retrieval tree are both binary trees, but the depths and the forms of the two trees are different. If the attribute values of the entities having a certain relationship with the instance entity are retrieved, the logical form of the retrieval tree is

<entity＝(<head_entity＝HE><relation＝'R'><tail_entity＝？>)><attr＝A><attr_value＝？>

Wherein, HE is a standard entity, HE is a head entity head _ entry in the search tree, relationship ═ R' indicates that HE is related to another entity as R, another entity is a tail entity tail _ entry in the search tree, attr ═ a indicates an attribute a of the tail entity, and attr _ value indicates an attribute value of the attribute a.

The search tree structure contains a root node 'attribute value', a left sub-tree (a left leaf node 'entity HE', a right leaf node 'relationship R') and a right leaf node 'attribute A', and the structure is searched once in two storage media respectively.

For example, when retrieving the registration date of M, the search tree is logically formed as

< entry > < investment > < tail _ entry > < attr _ value >, (< head _ entry > -M shares, ltd >) > < attr _ entry >? < CHEM > A

The corresponding search tree structure includes a root node "attribute value", a left sub-tree (left leaf node "M shares limited company", right leaf node "investment relation") and a right leaf node "registration date", and the generated search tree is shown in fig. 4.

In the embodiment, the retrieval tree corresponding to the retrieval type of the retrieval logic formula is generated, and indicates how to retrieve the information from the database, so that the information related to the query statement of the user can be accurately acquired from the database.

And step 207, performing information retrieval on the database according to the retrieval tree, and displaying the retrieved answer information.

Specifically, the nodes of the search tree are information to be searched, the left and right branches of each node are search conditions required for searching the node, and the root node of the binary tree is information to be searched finally. And the server performs depth-first traversal on the retrieval tree to perform feasibility verification on the retrieval tree and obtain a retrieval strategy.

For example, when the user query statement is "what is the registration date of Zhang III? If "the left leaf node of the search tree is" Zhang three ", the right leaf node is" registration date ". The server checks whether the node meets the grammar through depth-first traversal, and the third Zhang is a person name which is not matched with the registration date, namely the third Zhang does not have the feasibility of retrieving the registration date, and returns error prompt information. The depth-first traversal can check the feasibility of the search tree and determine the search steps in the database, namely, the left branch and the right branch of each node are searched first to obtain the related information of each node, and finally the related information of the root node is searched. And the determined retrieval step is a retrieval strategy, the server retrieves in each database according to the retrieval strategy, and returns answer information to the terminal for display after the answer information is retrieved.

Wherein Depth-First-Search is to reach a leaf node (i.e., a node that does not contain any branches) in the Search tree. When the depth-first retrieval is carried out on the retrieval tree, a single chain is completely searched, when no branch is left along the chain, the previous node is returned to continue exploring other chains in the retrieval tree, and when no other chain is selectable in the whole retrieval tree, the depth-first retrieval is finished.

In one embodiment, after the steps of retrieving information from the database according to the retrieval tree and presenting the retrieved answer information, the method further includes: setting the template query statement as an inventory query statement to update a question corpus; and correlating the newly added stock query sentences in the question corpus with the updated retrieval logic.

Specifically, after completing retrieval, the server adds a template query sentence obtained by replacing the user query sentence into a question corpus to obtain a new stock query sentence; and the retrieval logic formula updated according to the standard entity and the newly added inventory query statement are set to be associated with each other.

The newly added stock query sentences can participate in later retrieval so as to continuously enrich the question corpus and improve the system robustness and the processing capacity for different question sentences.

In the embodiment, the template query sentences are added into the question corpus and matched with the retrieval logic formula, so that the stock query sentences in the question corpus are enriched, and the system has improved processing capacity on various user query sentences.

In the embodiment, the instance entity in the obtained user query statement is replaced to obtain the template query statement, the template query statement carries out personalized removal on the user query statement, then the similarity between the template query statement and each stock query statement in the corpus is calculated, and the stock query statement matched with the user query statement and the retrieval logic formula thereof are determined according to the similarity, so that the processing capacity of the user query statements in various forms is improved, and the accuracy and the usability of information retrieval are ensured; the retrieval tree is generated according to the retrieval logic formula, the retrieval tree indicates how to retrieve information from a plurality of databases, and the information aimed by the user query sentence can be accurately retrieved from the databases by retrieval based on the retrieval tree, so that the accuracy of information retrieval is further ensured.

Further, as shown in fig. 5, step 207 may include:

step 2071, perform depth-first traversal on the search tree to determine the search strategy corresponding to the search tree, and determine the information type based on the search strategy.

The information type may be a type of information retrieved for a standard entity, including retrieving a single entity single attribute, an entity relationship, a single entity multiple attributes, a multiple entity single attribute, an attribute variation trend (including a single entity multiple attribute variation trend and a multiple entity single attribute variation trend), and the like.

Before retrieval, various information needs to be stored orderly. For triple data of the < entity-attribute value > type, real-time retrieval, analysis and screening are required to be met, and the triple data can be stored in an ElasticSearch distributed extensible database. In the elastic search, the retrieval frequency of the standard entity is obtained through big data or historical data, and the standard entity is inverted and indexed according to the retrieval frequency so as to retrieve the required information as soon as possible. Entity attributes (such as bulletins, news and other long text data) which are not used as retrieval conditions are stored in a traditional relational database PostgreSQL, so that the load of an ElasticSearch database is reduced. For triple data of the < head entity-relationship-tail entity > class, a graph database Neo4j stored in NoSQL (Not Only SQL, a non-relational database).

The server determines a retrieval policy by depth-first traversal, the retrieval policy indicating how to retrieve information from the database.

For example, for the search tree in fig. 3, the search strategy is: the ElasticSearch database is accessed and the registration date of M shares, Inc., is retrieved from the ElasticSearch database. For the search tree in fig. 4, the search strategy is: and (3) retrieving a tail entity having an investment relation with the M shares company from a Neo4j database, retrieving the registration date of the tail entity from an ElasticSearch database, and finally splicing the retrieved answer information in different databases based on the M shares company.

The type of information may be determined by the retrieval policy. For example, when the retrieval policy is to access the ElasticSearch database and retrieve the registration date of M shares company from the ElasticSearch database, only one attribute "registration date" of the entity "M shares company" needs to be retrieved, and the information type is a single entity and a single attribute. When the trade amount of six companies in 2019 of a certain industry needs to be retrieved, the same attribute 'trade amount' of six standard entities needs to be retrieved, and the information type is a multi-entity single attribute.

And 2072, performing information retrieval on the database according to the retrieval strategy to obtain answer information.

Specifically, the server accesses the database according to the determined retrieval strategy, extracts information from the database, and obtains answer information.

Step 2073, displaying the answer information according to the information type.

Specifically, the server determines a display mode of the answer information according to the information type, wherein the display mode comprises a mode of characters, charts and the like. And the server sends the answer information to the terminal, and the terminal displays the answer information according to the determined display mode.

In one embodiment, the step of displaying the answer information according to the information type specifically includes: when the information type is single entity single attribute or entity relation, the answer information is displayed by text; when the information type is single entity multi-attribute or multi-entity single attribute, displaying answer information by a histogram; and when the information type is the attribute variation trend, displaying the answer information by a line graph.

Specifically, when the information type determined by the search tree is a single entity and a single attribute, the answer information is displayed in a descriptive text. Taking fig. 3 as an example, the format of the descriptive text is: if the < attribute name > of < entity > is < attribute value >, there are: the registration date of M shares, Inc. is xxxx year xx month xx day, and the answer information dimension is 1 x 2.

And when the information type is single entity multi-attribute or multi-entity single attribute, displaying the answer information by using a histogram. For example, the answer information displayed when searching the trade amount of six companies in a certain industry is shown in fig. 6, and the data date and the name of each company may be further included in the answer information. The histogram shows answer information with answer dimensions of 1 × N (N >2) or N × 2, where N is a positive integer.

When the question contains keywords such as trend, change and the like or contains time sequence, the information type is attribute change trend, and answer information is displayed by a line graph. For example, when a trend of sales of M shares company No. 2019 in each quarter is searched, answer information is displayed as shown in fig. 7, and data date may be included in the answer information.

In the embodiment, answer information is displayed in a text mode, a graphic mode and the like according to the type of the retrieved information, and the intelligence of answer information display is improved.

Step 2074, upload the answer information to the blockchain.

Specifically, the corresponding digest information is obtained based on the answer information, and specifically, the digest information is obtained by hashing the answer information, for example, using the sha256s algorithm. Uploading summary information to the blockchain can ensure the safety and the fair transparency of the user. The user equipment can download the summary information from the blockchain so as to verify whether the answer information is tampered.

In the embodiment, the retrieval tree is subjected to depth-first traversal to obtain the retrieval strategy and the information type is determined based on the retrieval strategy, the server can more quickly and accurately obtain required information from the database according to the retrieval strategy, answer information can be intelligently displayed according to the information type, and the answer information is uploaded to the block chain to ensure the safety, the fairness and the transparency of the answer information.

It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by a computer program, which can be stored in a computer-readable storage medium, and can include the processes of the embodiments of the methods described above when the computer program is executed. The storage medium may be a non-volatile storage medium such as a magnetic disk, an optical disk, a Read-Only Memory (ROM), or a Random Access Memory (RAM).

It should be understood that, although the steps in the flowcharts of the figures are shown in order as indicated by the arrows, the steps are not necessarily performed in order as indicated by the arrows. The steps are not performed in the exact order shown and may be performed in other orders unless explicitly stated herein. Moreover, at least a portion of the steps in the flow chart of the figure may include multiple sub-steps or multiple stages, which are not necessarily performed at the same time, but may be performed at different times, which are not necessarily performed in sequence, but may be performed alternately or alternately with other steps or at least a portion of the sub-steps or stages of other steps.

With further reference to fig. 8, as an implementation of the method shown in fig. 2, the present application provides an embodiment of an information retrieval apparatus based on voice semantics, where the embodiment of the apparatus corresponds to the embodiment of the method shown in fig. 2, and the apparatus may be specifically applied to various electronic devices.

As shown in fig. 8, the information retrieval apparatus 300 based on speech semantics according to the present embodiment includes: a statement acquisition module 301, an entity replacement module 302, a similarity calculation module 303, a statement determination module 304, a logical formula update module 305, a search tree generation module 306, and an information search module 307, wherein:

a statement obtaining module 301, configured to obtain an input user query statement.

An entity replacement module 302, configured to analyze a user query statement, replace an instance entity in the user query statement with a concept entity, and obtain a template query statement; a conceptual entity is an entity type to which an instance entity belongs.

The similarity calculation module 303 is configured to calculate a similarity between the template query statement and each stock query statement in the question corpus.

And the statement determining module 304 is configured to determine, according to the calculated similarity, an inventory query statement matched with the template query statement and a retrieval logic expression corresponding to the inventory query statement.

And a logical formula updating module 305, configured to update the retrieval logical formula according to the instance entity.

And a search tree generating module 306, configured to generate a search tree based on the updated search logical expression.

And the information retrieval module 307 is configured to perform information retrieval on the database according to the retrieval tree, and display the retrieved answer information.

In some optional implementations of this embodiment, the entity replacing module 302 includes: statement parsing submodule, standard query submodule and entity replacement submodule, wherein:

and the statement parsing submodule is used for identifying the instance entity in the user query statement and determining the entity type of the instance entity through semantic identification to obtain the concept entity representing the entity type.

And the standard query submodule is used for querying the standard entity corresponding to the example entity from the standard entity list.

And the entity replacing submodule is used for replacing the instance entity in the user query statement with the concept entity to obtain the template query statement and storing the instance entity and the standard entity in a correlation manner.

In some optional implementations of this embodiment, the logic updating module 305 includes: an entity acquisition sub-module and a standard replacement sub-module, wherein:

and the entity acquisition submodule is used for acquiring the standard entity which is stored in association with the instance entity.

And the standard replacing submodule is used for replacing the standard entity in the retrieval logic formula with the acquired standard entity.

In some optional implementations of this embodiment, the search tree generating module 306 includes: the type identification submodule and the search tree generation submodule, wherein:

and the type identification submodule is used for identifying the retrieval type of the retrieval logic formula.

And the search tree generation submodule is used for generating the single triple single medium search tree when the search type is single triple single medium search.

And the search tree generation submodule is also used for generating the multi-triple multi-medium search tree when the search type is multi-triple multi-medium search.

In some optional implementations of this embodiment, the information retrieving module 307 includes: the system comprises a depth traversal submodule, an information retrieval submodule, an information display submodule and an information uploading submodule, wherein:

and the depth traversal submodule is used for performing depth-first traversal on the retrieval tree to determine a retrieval strategy corresponding to the retrieval tree and determine the information type based on the retrieval strategy.

And the information retrieval submodule is used for carrying out information retrieval on the database according to the retrieval strategy to obtain answer information.

And the information display submodule is used for displaying the answer information according to the information type.

And the information uploading submodule is used for uploading the answer information to the block chain.

In some optional implementation manners of this embodiment, the information display sub-module includes: text display element, histogram display element and line graph display element, wherein:

and the text display unit is used for displaying the answer information in a text when the information type is a single entity single attribute or an entity relation.

And the histogram display unit is used for displaying the answer information by using the histogram when the information type is single entity multi-attribute or multi-entity single attribute.

And the line chart display unit is used for displaying the answer information by using a line chart when the information type is the attribute change trend.

In some optional implementations of the present embodiment, the information retrieval apparatus 300 based on voice semantics further includes: statement update module and association module, wherein:

and the statement updating module is used for setting the template query statement as an inventory query statement so as to update the question corpus.

And the association module is used for associating the stock query sentences newly added in the question corpus with the updated retrieval logic formula.

In order to solve the technical problem, an embodiment of the present application further provides a computer device. Referring to fig. 9, fig. 9 is a block diagram of a basic structure of a computer device according to the present embodiment.

The computer device 4 comprises a memory 41, a processor 42, a network interface 43 communicatively connected to each other via a system bus. It is noted that only computer device 4 having components 41-43 is shown, but it is understood that not all of the shown components are required to be implemented, and that more or fewer components may be implemented instead. As will be understood by those skilled in the art, the computer device is a device capable of automatically performing numerical calculation and/or information processing according to a preset or stored instruction, and the hardware includes, but is not limited to, a microprocessor, an Application Specific Integrated Circuit (ASIC), a Programmable gate array (FPGA), a Digital Signal Processor (DSP), an embedded device, and the like.

The computer device can be a desktop computer, a notebook, a palm computer, a cloud server and other computing devices. The computer equipment can carry out man-machine interaction with a user through a keyboard, a mouse, a remote controller, a touch panel or voice control equipment and the like.

The memory 41 includes at least one type of readable storage medium including a flash memory, a hard disk, a multimedia card, a card type memory (e.g., SD or DX memory, etc.), a Random Access Memory (RAM), a Static Random Access Memory (SRAM), a Read Only Memory (ROM), an Electrically Erasable Programmable Read Only Memory (EEPROM), a Programmable Read Only Memory (PROM), a magnetic memory, a magnetic disk, an optical disk, etc. In some embodiments, the memory 41 may be an internal storage unit of the computer device 4, such as a hard disk or a memory of the computer device 4. In other embodiments, the memory 41 may also be an external storage device of the computer device 4, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a flash Card (FlashCard), and the like, which are provided on the computer device 4. Of course, the memory 41 may also include both internal and external storage devices of the computer device 4. In this embodiment, the memory 41 is generally used for storing an operating system installed in the computer device 4 and various types of application software, such as program codes of an information retrieval method based on voice semantics. Further, the memory 41 may also be used to temporarily store various types of data that have been output or are to be output.

The processor 42 may be a Central Processing Unit (CPU), controller, microcontroller, microprocessor, or other data Processing chip in some embodiments. The processor 42 is typically used to control the overall operation of the computer device 4. In this embodiment, the processor 42 is configured to execute the program code stored in the memory 41 or process data, for example, execute the program code of the information retrieval method based on the voice semantics.

The network interface 43 may comprise a wireless network interface or a wired network interface, and the network interface 43 is generally used for establishing communication connection between the computer device 4 and other electronic devices.

The computer device provided in this embodiment may perform the steps of the information retrieval method based on the voice semantics. Here, the steps of the information retrieval method based on the speech semantics may be the steps in the information retrieval method based on the speech semantics of the above embodiments.

The present application further provides another embodiment, which is to provide a computer-readable storage medium storing a speech-semantic-based information retrieval program, which is executable by at least one processor to cause the at least one processor to perform the steps of the speech-semantic-based information retrieval method as described above.

The block chain is a novel application mode of computer technologies such as distributed data storage, point-to-point transmission, a consensus mechanism, an encryption algorithm and the like. A block chain (Blockchain), which is essentially a decentralized database, is a series of data blocks associated by using a cryptographic method, and each data block contains information of a batch of network transactions, so as to verify the validity (anti-counterfeiting) of the information and generate a next block. The blockchain may include a blockchain underlying platform, a platform product service layer, an application service layer, and the like.

Through the above description of the embodiments, those skilled in the art will clearly understand that the method of the above embodiments can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware, but in many cases, the former is a better implementation manner. Based on such understanding, the technical solutions of the present application may be embodied in the form of a software product, which is stored in a storage medium (such as ROM/RAM, magnetic disk, optical disk) and includes instructions for enabling a terminal device (such as a mobile phone, a computer, a server, an air conditioner, or a network device) to execute the method according to the embodiments of the present application.

It is to be understood that the above-described embodiments are merely illustrative of some, but not restrictive, of the broad invention, and that the appended drawings illustrate preferred embodiments of the invention and do not limit the scope of the invention. This application is capable of embodiments in many different forms and is provided for the purpose of enabling a thorough understanding of the disclosure of the application. Although the present application has been described in detail with reference to the foregoing embodiments, it will be apparent to one skilled in the art that the present application may be practiced without modification or with equivalents of some of the features described in the foregoing embodiments. All equivalent structures made by using the contents of the specification and the drawings of the present application are directly or indirectly applied to other related technical fields and are within the protection scope of the present application.

Claims

1. An information retrieval method based on voice semantics is characterized by comprising the following steps:

acquiring an input user query statement;

analyzing the user query statement, and replacing an instance entity in the user query statement with a concept entity to obtain a template query statement; the concept entity is an entity type to which the instance entity belongs;

calculating the similarity between the template query statement and each stock query statement in the question corpus;

updating a retrieval logic formula according to the instance entity;

generating a retrieval tree based on the updated retrieval logic formula;

2. The information retrieval method based on the voice semantics as claimed in claim 1, wherein the step of parsing the user query sentence, replacing an instance entity in the user query sentence with a concept entity, and obtaining a template query sentence specifically includes:

3. The information retrieval method based on speech semantics as claimed in claim 2, wherein the step of updating the retrieval logic formula according to the instance entity specifically comprises:

acquiring a standard entity stored in association with the instance entity;

4. The information retrieval method based on speech semantics as claimed in claim 1, wherein the step of generating a retrieval tree based on the updated retrieval logic formula specifically comprises:

identifying a retrieval type of the retrieval logic formula;

5. The information retrieval method based on the voice semantics as claimed in claim 1, wherein the step of performing information retrieval on the database according to the retrieval tree and displaying the retrieved answer information specifically includes:

displaying the answer information according to the information type;

and uploading the answer information to a block chain.

6. The information retrieval method based on the voice semantics as claimed in claim 5, wherein the step of displaying the answer information according to the information type specifically comprises:

7. The information retrieval method based on voice semantics as claimed in any one of claims 1 to 6, wherein after the step of retrieving information from the database according to the retrieval tree and presenting the retrieved answer information, the method further comprises:

8. An information retrieval apparatus based on speech semantics, comprising:

the entity replacement module is used for analyzing the user query statement, replacing an instance entity in the user query statement with a concept entity and obtaining a template query statement; the concept entity is an entity type to which the instance entity belongs;

9. A computer device comprising a memory in which a computer program is stored and a processor which, when executing the computer program, carries out the steps of the method for information retrieval based on speech semantics of any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program; characterized in that the computer program realizes the steps of the method for information retrieval based on speech semantics according to any one of claims 1 to 7 when being executed by a processor.