WO2021159656A1

WO2021159656A1 - Method, device, and equipment for semantic completion in a multi-round dialogue, and storage medium

Info

Publication number: WO2021159656A1
Application number: PCT/CN2020/098846
Authority: WO
Inventors: 黄孟缘
Original assignee: 平安科技（深圳）有限公司
Priority date: 2020-02-12
Filing date: 2020-06-29
Publication date: 2021-08-19
Also published as: CN111325034A

Abstract

A method, device, and equipment for semantic completion in a multi-round dialogue, and a storage medium, grammar check is performed with respect to the multi-round dialogue via a preset corpus segmentation function and a preset analysis function, and semantically incomplete sentences are completed, thus increasing the accuracy of a semantic analysis result and the accuracy of searching for corresponding response information on the basis of the semantic analysis result. The method comprises: utilizing a preset corpus segmentation function and a preset analysis function to respectively perform grammar checks with respect to a first sentence and a second sentence to produce a first sentence check result and a second sentence check result; when the second sentence check result comprises a single entity and the second sentence is a question, completing a semantically missing part of the second sentence on the basis of the first sentence check result to produce a first completed sentence (104); and if the first completed sentence comprises an unclear word, then replacing the unclear word in the first completed sentence on the basis of the first sentence check result to produce a second completed sentence (106).

Description

Method, device, equipment and storage medium for complementing semantics in multiple rounds of dialogue

This application claims the priority of a Chinese patent application filed with the Chinese Patent Office on February 12, 2020, the application number is 202010088078.9, and the invention title is "Methods, Apparatus, Equipment, and Storage Media for Semantic Completion in Multi-round Dialogues". The entire content is incorporated into this application by reference.

Technical field

This application relates to the field of artificial intelligence, and in particular to methods, devices, equipment and storage media for semantic complement in multiple rounds of dialogue.

Background technique

Human-computer interaction (HCI) is a study of the interaction between systems and users. The system can be a variety of machines or computerized systems and software. Generally, the user communicates with the system through a visible human-computer interaction interface, and performs operations to realize the exchange of information between the user and the system to complete certain tasks. Human-computer interaction is an important part of the field of artificial intelligence, especially in terms of customer service or consulting acquisition. Through human-computer interaction, users can obtain the information they need.

In terms of customer service, currently, intelligent dialogue robots are used to communicate with users, and the feedback data of intelligent dialogue robots are used to meet the needs of users to obtain information.

However, the inventor realizes that because the intelligent dialogue robot obtains the user’s request, there is a situation of unclear semantics. If the user input request lacks the subject or object, the intelligent dialogue robot cannot accurately recognize the user’s intention and cannot accurately give feedback. The response information made according to the user's intention leads to the low accuracy of the intelligent dialogue robot's feedback.

Summary of the invention

The present application provides a method, device, device, and storage medium for semantic complement in multiple rounds of dialogue, which are used to solve the problem of incomplete sentence semantics based on multiple rounds of dialogue, improve the accuracy of semantic analysis results, and at the same time improve The accuracy of searching the corresponding response information based on the semantic analysis result is improved.

The first aspect of the embodiments of this application provides a method for semantic complement in multiple rounds of dialogue, including: using a preset corpus sentence segmentation function and a preset analysis function to perform grammatical detection on the first sentence input by the user, Obtain the first sentence detection result, the first sentence detection result is the target dependency relationship between each word in the first sentence; using the preset corpus segmentation function and the preset analysis function to obtain The second sentence input by the user in the new round is grammatically tested, and the second sentence detection result is obtained. The second sentence detection result is the target dependency relationship between each word in the second sentence; Whether the sentence detection result includes only a single entity and the second sentence is an interrogative sentence; when the second sentence detection result includes only a single entity and the second sentence is an interrogative sentence, the first sentence detection result Complete the semantic missing part of the second sentence to obtain the first supplementary sentence; determine whether the first supplementary sentence includes unidentified words, and the unidentified words include pronouns, quantifiers, and articles; if If the first supplementary sentence includes an unknown word, the unknown word in the first supplementary sentence is replaced according to the detection result of the first sentence to obtain a second supplementary sentence.

A second aspect of the embodiments of the present application provides a device for complementing semantics in multiple rounds of dialogue, including: a first acquiring unit, configured to use a preset corpus segmentation function and a preset analysis function to compare the acquired first input by the user A sentence is checked for grammar, and a first sentence detection result is obtained. The first sentence detection result is the target dependency relationship between each word in the first sentence; the second acquisition unit is configured to use the preset corpus The sentence segmentation function and the preset analysis function perform grammatical detection on the acquired second sentence input by the user in a new round to obtain a second sentence detection result, and the second sentence detection result is each of the second sentences The target dependency relationship between words; the first judgment unit is used to judge whether the detection result of the second sentence includes only a single entity and the second sentence is a question sentence; the first complement unit is used when the second sentence When the sentence detection result only includes a single entity and the second sentence is an interrogative sentence, according to the first sentence detection result, the semantic and semantic missing parts of the second sentence are filled in to obtain the first complementary sentence; the second judgment unit , Used to determine whether the first supplementary sentence includes unclear words, the unidentified words include pronouns, quantifiers, and articles; the second supplement unit is used if the first supplement If the sentence includes an unclear word, then the unclear word in the first supplementary sentence is replaced according to the detection result of the first sentence to obtain a second supplementary sentence.

The third aspect of the embodiments of the present application provides a device for complementing semantics in multiple rounds of dialogue, including a memory, a processor, and a computer program stored in the memory and running on the processor. The processing When the computer program is executed by the computer program, the method based on semantic completion in multiple rounds of dialogues described in any of the above embodiments is implemented. The input first sentence is grammatically checked, and the first sentence detection result is obtained. The first sentence detection result is the target dependency relationship between each word in the first sentence; The preset analysis function performs grammatical detection on the acquired second sentence inputted by the user in a new round to obtain the second sentence detection result. The second sentence detection result is the difference between each word in the second sentence Target dependency relationship; determine whether the second sentence detection result includes only a single entity and the second sentence is a question sentence; when the second sentence detection result only includes a single entity and the second sentence is a question sentence, according to the According to the detection result of the first sentence, the missing part of the semantics of the second sentence is added to obtain the first supplementary sentence; it is judged whether the first supplementary sentence includes an unclear word, and the unclear word refers to it. Including pronouns, quantifiers, and articles; if the first supplementary sentence includes an unknown word, replace the unknown word in the first supplementary sentence according to the first sentence detection result to obtain the first supplementary sentence Two complete sentences.

The fourth aspect of the embodiments of the present application provides a computer-readable storage medium that stores instructions in the computer-readable storage medium, which when run on a computer, causes the computer to execute the method described in the first aspect, For example, the following steps are implemented: using the preset corpus sentence segmentation function and the preset analysis function to perform grammatical detection on the acquired first sentence input by the user to obtain the first sentence detection result, and the first sentence detection result is the first sentence The target dependency relationship between each word in a sentence; using the preset corpus sentence segmentation function and the preset analysis function to perform a grammatical check on the acquired second sentence input by the user in a new round to obtain the second sentence The detection result, the second sentence detection result is the target dependency relationship between each word in the second sentence; it is judged whether the second sentence detection result only includes a single entity and the second sentence is a question sentence; when When the second sentence detection result includes only a single entity and the second sentence is an interrogative sentence, fill in the semantic missing part of the second sentence according to the first sentence detection result to obtain the first supplementary sentence; Whether the first supplementary sentence includes unclear words, the unclear words include pronouns, quantifiers, and articles; if the first supplementary sentence includes unclear words, it will be determined according to the According to the detection result of the first sentence, the unknown words in the first supplementary sentence are replaced to obtain the second supplementary sentence.

The embodiments of the present application improve the accuracy of the semantic analysis result, and at the same time also improve the accuracy of searching the corresponding response information according to the semantic analysis result.

Description of the drawings

FIG. 1 is a schematic diagram of an embodiment of a method for semantic completion in multiple rounds of dialogue in this application;

FIG. 2 is a schematic diagram of another embodiment of a method for semantic completion in multiple rounds of dialogue in this application;

FIG. 3 is a schematic diagram of an embodiment of a device for complementing semantics in multiple rounds of dialogue in this application;

FIG. 4 is a schematic diagram of another embodiment of a device for complementing semantics in multiple rounds of dialogue in this application;

Fig. 5 is a schematic diagram of an embodiment of a device for semantic completion in multiple rounds of dialogues of the present application.

Detailed ways

The present application provides a method of semantic complement in multiple rounds of dialogue, which is used to solve the problem of incomplete sentence semantics based on multiple rounds of dialogue, improve the accuracy of semantic analysis results, and also improve the search for correspondence based on semantic analysis results. The accuracy of the response information.

The technical solution of this application can be applied to the field of artificial intelligence or big data technology. For example, the technical solution of this application can be implemented by a data platform such as a cloud computing platform.

In order to enable those skilled in the art to better understand the solutions of the present application, the technical solutions in the embodiments of the present application will be described below in conjunction with the drawings in the embodiments of the present application.

The terms "first", "second", "third", "fourth", etc. (if any) in the description and claims of this application and the above-mentioned drawings are used to distinguish similar objects, and do not need to be used. To describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances so that the embodiments described herein can be implemented in a sequence other than the content illustrated or described herein. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not necessarily limited to those clearly listed. Steps or units, but may include other steps or units that are not clearly listed or are inherent to these processes, methods, products, or equipment.

Referring to FIG. 1, an embodiment of the method for semantic completion in multiple rounds of dialogue in the embodiment of the present application includes:

101. Use the preset corpus sentence segmentation function and the preset analysis function to perform grammatical detection on the first sentence input by the user, and obtain the first sentence detection result. The first sentence detection result is between each word in the first sentence The goal dependence relationship.

The server uses the preset corpus sentence segmentation function and the preset analysis function to perform grammatical detection on the first sentence input by the user, and obtain the first sentence detection result. The first sentence detection result is the difference between each word in the first sentence. Target dependency.

It should be noted that the server uses the preset corpus sentence segmentation function to segment the first sentence and uses the preset analysis function to perform grammatical check on the first input sentence, both of which are preparations for complementing the semantics of the second sentence. The server uses the preset corpus sentence segmentation function to preliminarily segment the first sentence. The sentence segmentation is the basis of natural language processing. The accuracy of the sentence segmentation directly determines the quality of the part-of-speech tagging, syntactic analysis, word vector, and text analysis of the sentence behind the server. For a simple first sentence, you can directly use a comma to segment the sentence. The server will mark the words in the first sentence according to their position in the sentence. For example: mark the sentence "I love Shenzhen because it is beautiful". The result is: "I N love N Shenzhen Y because N is N, it is N is very beautiful N", where Y stands for sentence segmentation, N stands for non-sentence segmentation, which can be regarded as the server's response to each word in the first sentence. The position between the two is classified, and judge whether the first sentence should be segmented.

It should be noted that the preset analysis function here refers to the dependency analysis function. The server uses the dependency analysis function to check the grammar of the first sentence. The dependency analysis function uses the dependency relationship between the words in the sentence to express the grammar, such as: a When a word modifies another word, it is said that the word depends on another word, and the relationship between the words is clarified in the sentence. The server can accurately determine the part of speech of each word in the sentence and the relationship between the words, which is a paired sentence Make full preparations to make up. Before performing dependency analysis, the server needs to tag the sentence part-of-speech. The server tags the words in the first sentence according to their position in the word. The BMES tagging method can be used, where B is a word beginning with a word, and M It is a word in the middle of a word, the number of words in the middle of a word may be multiple, E is the last word of a word, and S is a word composed of one word. For example, mark "Online Merchant Bank is the most important product of Ant Financial's microfinance division", the result after labeling is "BMMESBMMEBMMMESBMEBE", and the corresponding word segmentation result is "Online Merchant Bank/Yes/Ant Financial Services/WeChat Loan Division/The/Most Important/Product", search for the part of speech of the corresponding word segmentation in the preset vocabulary, and mark the word segmentation result with part of speech. When a single entity appears in the sentence (the sentence includes only the subject, predicate or object), Extract the single entity and save it in the first sentence detection result for future use.

102. Use the preset corpus sentence segmentation function and the preset analysis function to perform a grammatical check on the acquired second sentence input by the user in a new round, and obtain the second sentence detection result. The second sentence detection result is each of the second sentences. The goal dependence relationship between words.

The server uses the preset corpus sentence segmentation function and the preset analysis function to perform grammatical detection on the acquired second sentence input by the user in the new round, and obtain the second sentence detection result. The second sentence detection result is each word in the second sentence The goal dependence relationship between.

The second sentence here is the following sentence of the first sentence, and the first sentence serves as the basis for filling in the missing part of the second sentence. Here, the grammatical detection method for the second sentence is the same as the grammatical detection method for the first sentence, and the detection result is used as the second sentence detection result.

103. Determine whether the detection result of the second sentence includes only a single entity and the second sentence is an interrogative sentence.

The server determines whether the detection result of the second sentence includes only a single entity and the second sentence is an interrogative sentence.

It is understandable that part of the sentence may be omitted in the second sentence input by the user in the new round, such as omitting the predicate and object. At this time, the server needs to judge the second sentence input in the new round. If the semantic expression of the second sentence is not complete, the second sentence entered in the new round will be completed according to the context in the multiple rounds of dialogue. Therefore, the server first determines whether the analysis result of the second sentence contains only a single entity and the second sentence It is an interrogative sentence, which is used to determine whether the second sentence entered in the new round satisfies the condition of semantic incompleteness.

104. When the second sentence detection result includes only a single entity and the second sentence is an interrogative sentence, fill in the semantic missing part of the second sentence according to the first sentence detection result to obtain the first supplementary sentence.

When the second sentence detection result includes only a single entity and the second sentence is an interrogative sentence, the server fills in the semantic missing part of the second sentence according to the first sentence detection result to obtain the first supplementary sentence.

If the detection result of the second sentence contains a single entity and the second sentence is an interrogative sentence, the server cannot accurately identify the meaning of the second sentence and cannot feed back the corresponding preset response. Therefore, the server fills in the second sentence according to the acquired context , A complete semantics of the second sentence is obtained, and the server feeds back the corresponding answer according to the semantics of the second sentence. The server obtains the first sentence detection result and the second sentence detection result. Both detection results include part-of-speech analysis of the sentence, and extract the single entity in the second sentence as the first target word, where the single entity refers to It is a single word or a single phrase in a sentence, for example: the sentence "What about a cold?", where the word "a cold" is a single entity in the sentence. After extracting the single entity in the second sentence, select the single entity with the same grammatical structure as the first target word in the first sentence as the second target word. The same grammatical structure here refers to the first target word and the first target word. The two target words have the same part of speech, for example: the first target word is "cold", the second target word is "cancer", the first target word and the second target word are both noun parts of speech, it can be said that the two have the same grammatical structure. After the server filters out the second target words, in the first sentence, the server replaces the first target words with the second target words to obtain the first supplementary sentence.

105. Determine whether the first supplementary sentence includes unidentified words, which include pronouns, quantifiers, and articles.

The server determines whether the first supplementary sentence includes unclear words, which include pronouns, quantifiers, and articles.

It should be noted that when the server completes a new round of input of the second sentence and obtains the first supplementary sentence, it must also determine whether the first supplementary sentence includes unclear words, which will lead to The semantics of the second sentence is unclear. Generally speaking, unclear words include pronouns, quantifiers and articles. For example: the first sentence of the user is "Are both Ping An Fu and Fu Tianxin medical insurance?" The server replies to the first sentence as "Yes." The second sentence entered by the user in a new round is "The first type can be reimbursed." What?" In the second sentence entered by the user in the new round, there is an unknown term "the first type". If the server does not contact the context of multiple rounds of dialogue to supplement the second sentence entered by the user in the new round, then It will lead to a deviation in semantic understanding, the server will not be able to give the correct answer, and the user will not be able to proceed to the next step.

106. If the first supplementary sentence includes an unknown word, replace the unknown word in the first supplementary sentence according to the detection result of the first sentence to obtain a second supplementary sentence.

If the first supplementary sentence includes an unknown word, the server replaces the unknown word in the first supplementary sentence according to the detection result of the first sentence to obtain a second supplementary sentence.

If the server determines that the first supplementary sentence includes an unknown word, indicating that the first supplementary sentence is a sentence with ambiguous semantics, the server needs to supplement the first supplementary sentence again according to the context in multiple rounds of dialogue. First, the server extracts the unknown words from the second sentence as the third target word. Secondly, it selects words with the same part of speech as the third target word in the first sentence as the fourth target word. In the second sentence, replace the fourth target word with the third target word.

In the embodiment of the present application, by using the preset corpus sentence segmentation function and the preset analysis function to perform grammatical detection on multiple rounds of dialogue input by the user, and using the semantics of the multiple rounds of dialogue context to complement sentences with incomplete semantics, the results of semantic analysis are improved. At the same time, the accuracy of searching the corresponding response information based on the semantic analysis result is also improved.

Referring to FIG. 2, another embodiment of the method for semantic completion in multiple rounds of dialogue in the embodiment of the present application includes:

201. Obtain the first sentence input by the user.

The server obtains the first sentence entered by the user. In human-computer interaction, the server obtains the sentence input by the user, and uses the sentence input by the user as the first sentence to prepare for the completion of the second sentence below.

202. Use a preset corpus sentence segmentation function to segment the first sentence to obtain the first input sentence.

The server uses the preset corpus sentence segmentation function to segment the first sentence to obtain the first input sentence. Specifically, the server segments the first sentence to obtain a segmented sentence; the server matches the segmented corpus in the segmented sentence with a preset corpus, which is a corpus established in a preset intent rule base based on business data ; If the segmented corpus matches the preset corpus, the server will segment the segmented sentence before and after the segmented corpus to obtain the segmented sentence, and use the segmented sentence as the first input sentence; if the segmented corpus matches the preset corpus If it does not match, the server directly uses the segmented corpus as the first input sentence.

It should be noted that the server uses the preset corpus sentence segmentation function to segment the first sentence and uses the preset analysis function to perform grammatical check on the first input sentence, both of which are preparations for complementing the semantics of the second sentence. The server uses the preset corpus sentence segmentation function to preliminarily segment the first sentence. The sentence segmentation is the basis of natural language processing. The accuracy of sentence segmentation directly determines the quality of part-of-speech tagging, syntactic analysis, word vector and text analysis behind the server. For a simple first sentence, you can directly use a comma to segment the sentence. The server will mark the words in the first sentence according to their position in the sentence. For example: mark the sentence "I love Shenzhen because it is beautiful". The result is: "I N love N Shenzhen N Y because N is N, it is N is very N beautiful N", where Y stands for sentence segmentation, N stands for non-sentence segmentation. Make a two-class classification of the position between the two to determine whether the sentence should be segmented.

After segmenting the first sentence, the server obtains the separated segmented sentence, and then matches the segmented corpus in the segmented sentence with the preset corpus. The segmented sentence here is a sentence that cannot be segmented, so Further match the segmented corpus in the segmented sentence, and clarify whether the corresponding corpus needs to be segmented. The preset corpus here is the corpus built in the preset intent rule base based on business data. For example, the preset intent rule base is a vault, and the preset corpus can be an insurance policy, an ID number of the insured, etc., based on the server. Business data constitutes corpus. If the segmented corpus matches the preset corpus, the server will segment the segmented sentence before and after the segmented corpus to obtain the segmented sentence, and use the segmented sentence as the first input sentence; if the segmented corpus does not match the preset corpus Match, the server directly uses the segmented corpus as the first input sentence. For example: match the sentence "Can a cold be insured for e-life insurance?" After matching the segmented corpus, it is divided into "cold/can/insured/e-life insurance/?", where "cold", "insurance", "e-life insurance" All are the preset corpus in the preset intent rule library. The server does not need to segment the segmented corpus that matches the preset corpus, and directly divides the segmented corpus before and after the matched segmented corpus, saving the server to segment the sentence time.

203. Perform grammatical detection on the first input sentence by using a preset analysis function to obtain a first sentence detection result. The first sentence detection result is a target dependency relationship between each word in the first input sentence.

The server uses a preset analysis function to perform grammatical detection on the first input sentence to obtain a first sentence detection result. The first sentence detection result is the target dependency relationship between each word in the first input sentence. Specifically, the server performs part-of-speech tagging and entity extraction on the words in the first input sentence to obtain the first sentence tagging result; the server calculates the dependency probability between each word, the word is the word in the first sentence tagging result, and the dependency probability It is the frequency of occurrence of the preset dependence relationship; the server determines the target dependence relationship between each word, and the preset dependence relationship corresponding to the probability with the largest dependence probability weight is the target dependence relationship between the words; the server obtains the first sentence detection result , The first sentence detection result is the target dependency relationship between each word in the first input sentence.

It should be noted that the preset analysis function here refers to the dependency analysis function. The server uses the dependency analysis function to check the grammar of the first sentence. The dependency analysis function uses the dependency relationship between the words in the sentence to express the grammar, such as: a When a word modifies another word, it is said that the word depends on another word, and the relationship between the words is clarified in the sentence. The server can accurately determine the part of speech of each word in the sentence and the relationship between the words, which is a paired sentence Make full preparations to make up. Before the server performs dependency analysis, the sentence needs to be marked with part of speech. The server marks the words in the first sentence according to their position in the word. The BMES marking method can be used, where B is a word beginning with a word, and M is For a word in the middle of a word, the number of words in the middle of a word may be more than one. E is the last word of a word, and S is a word consisting of one word. For example, mark "Online Merchant Bank is the most important product of Ant Financial's microfinance division", the result after labeling is "BMMESBMMEBMMMESBMEBE", and the corresponding word segmentation result is "Online Merchant Bank/Yes/Ant Financial Services/WeChat Loan Division/The/Most Important/Product", search for the part of speech of the corresponding word segmentation in the preset vocabulary, and mark the word segmentation result with part of speech. When a single entity appears in the sentence (the sentence includes only the subject, predicate or object), Extract the single entity and save it in the first sentence detection result for future use.

204. Use the preset corpus sentence segmentation function and the preset analysis function to perform a grammatical check on the acquired second sentence input by the user in a new round to obtain the second sentence detection result, and the second sentence detection result is each of the second sentences. The goal dependence relationship between words.

205. Determine whether the detection result of the second sentence includes only a single entity and the second sentence is an interrogative sentence.

It is understandable that part of the sentence may be omitted in the second sentence input by the user in the new round, such as omitting the predicate and object. At this time, the server needs to judge the second sentence input in the new round. If the semantic expression of the second sentence is not complete, the second sentence entered in the new round will be completed according to the context in the multiple rounds of dialogue. Therefore, the server first determines whether the analysis result of the second sentence contains only a single entity and the second sentence It is an interrogative sentence to judge whether the second sentence entered in the new round satisfies the condition of semantic incompleteness.

206. When the second sentence detection result includes only a single entity and the second sentence is an interrogative sentence, fill in the semantic missing part of the second sentence according to the first sentence detection result to obtain the first supplementary sentence.

When the second sentence detection result includes only a single entity and the second sentence is a question sentence, the server completes the semantics of the second sentence according to the first sentence detection result to obtain the first supplementary sentence. Specifically, when the second sentence detection result includes only a single entity and the second sentence is an interrogative sentence, the server extracts the single entity in the second sentence detection result and uses the single entity as the first target word; the server detects in the first sentence In the result, single entities with the same grammatical structure as the first target word are selected to obtain the second target word; in the first sentence, the server replaces the second target word with the first target word to obtain the first supplementary sentence.

If the detection result of the second sentence contains only a single entity and the second sentence is a question sentence, the server cannot accurately identify the meaning of the second sentence, and cannot feed back the corresponding preset response. Therefore, the server will determine the meaning of the second sentence according to the acquired context. The semantic missing parts are filled in to obtain the complete semantics of the second sentence, and the server then feeds back the corresponding answer according to the semantics of the second sentence. The server obtains the first sentence detection result and the second sentence detection result. Both detection results include part-of-speech analysis of the sentence, extract the single entity in the second sentence as the first target word, and detect it in the first sentence In the result, a single entity with the same grammatical structure as the first target word is selected as the second target word. In the first sentence, the server replaces the first target word with the second target word to obtain the first supplementary sentence.

For example: the first sentence entered by the user is "Can I apply for insurance for a cold?", and the server performs part-of-speech tagging and dependency analysis on the first sentence input by the user, and obtains "Cold/nhd Yes/c Apply for insurance/vn e生保/ nbx what/y?” where nhd and nbx are related parts of speech of the preset corpus, for example: the preset intent rule base where the preset corpus is located is the vault, nhd represents the word of disease in the vault, and nbx represents insurance Words like insurance names in the library; c stands for auxiliary verbs; vn stands for gerunds; y stands for modal auxiliary words. The extracted single entity is: cold/nhd insurance/vn e生保/nbx. The user enters a new round of the second sentence as "Cancer?" The server parses this sentence and gets "Cancer/nhd which/y". It is obvious that the second sentence only includes a single entity and the second sentence is a rhetorical question. At the end of the tone, if the semantic incomplete condition of the second sentence is met, the server will supplement the second sentence. According to the detection result of the first sentence and the detection result of the second sentence, it is obtained that cold/nhd and cancer/nhd are the same type of words. In the first sentence, you can get "Can cancer be insured for e-life insurance?" and fill in the second sentence.

207. Determine whether the first supplementary sentence includes unclear words, which include pronouns, quantifiers, and articles.

The server determines whether the first supplementary sentence includes unclear words, which include pronouns, quantifiers, and articles. Specifically, the server obtains the detection result of the first supplementary sentence, the detection result of the first supplementary sentence is a combination of the second sentence detection result and the first sentence detection result; the server determines whether the detection result of the first supplementary sentence includes Pronouns, quantifiers, and articles.

It should be noted that when the server completes a new round of input of the second sentence and obtains the first supplementary sentence, it must also determine whether the first supplementary sentence includes unclear words, which will lead to The semantics of the second sentence is unclear. Generally speaking, unclear words include pronouns, quantifiers and articles. For example: the first sentence of the user is "Are both Ping An Fu and Fu Dan for medical insurance?" The server replies to the first sentence as "Yes." The second sentence entered by the user in a new round is "The first type can be reimbursed." What?" In the second sentence entered by the user in the new round, there is an unknown term "the first type". If the server does not contact the context of multiple rounds of dialogue to supplement the second sentence entered by the user in the new round, then It will lead to a deviation in semantic understanding, the server will not be able to give the correct answer, and the user will not be able to proceed to the next step.

208. If the first supplementary sentence includes an unknown word, replace the unknown word in the first supplementary sentence according to the detection result of the first sentence to obtain a second supplementary sentence.

If the first supplementary sentence includes an unknown word, the server replaces the unknown word in the first supplementary sentence according to the detection result of the first sentence to obtain a second supplementary sentence. Specifically, if the detection result of the first supplementary sentence includes an unknown word, the server extracts the unknown word, and uses the unknown word as the third target word; the server is in the first sentence detection result The words with the same grammatical structure as the third target word are selected to obtain the fourth target word; in the second sentence, the server replaces the third target word with the fourth target word to obtain the second supplementary sentence.

If the server determines that the first supplementary sentence includes an unknown word, indicating that the first supplementary sentence is a sentence with ambiguous semantics, the server needs to supplement the first supplementary sentence again according to the context in multiple rounds of dialogue. First, the server extracts the unknown words from the second sentence as the third target word. Secondly, it selects words with the same part of speech as the third target word in the first sentence as the fourth target word. Replace the fourth target word with the third target word in the second sentence.

For example: the first sentence of the user is "Ping An Fu and Fu Yuan, are they all medical insurance?", the server performs part-of-speech tagging and dependency analysis on the first sentence input by the user, and obtains "Ping An Fu/nbx, Fu Yuan/ nbx, is it /vMedical Insurance/nbx/y?” where nbx is the relevant part of speech of the preset corpus, for example: the preset intent rule library where the preset corpus is located is the vault, and nbx represents the name of the insurance in the vault. Class words; v stands for verbs; y stands for modal particles. The second sentence entered by the user in the new round is "What can the former be reimbursed?", the server performs part-of-speech tagging and dependency analysis on the second sentence entered by the user in the new round, and obtains "the former/r能/v reimbursement/vwhat/y "?" where r stands for pronouns; v stands for verbs; y stands for modal auxiliary words. In the second sentence input by the user in the new round, there is an unidentified word "former". The second sentence entered in the round is supplemented, combined with the detection result of the user's first sentence, the former here refers to "Ping An Fu", and "Ping An Fu" is directly brought into the second sentence to get "What is Ping An Fu reimbursed?" ?" Get a complete second supplementary sentence, so that the server can give feedback and answers based on the second supplementary sentence.

The above describes the method of semantic completion in multiple rounds of dialogue in the embodiment of this application. The following describes the device for semantic completion in multiple rounds of dialogue in the embodiment of this application. Please refer to FIG. 3. An embodiment of the device for semantic completion in the middle includes:

The first acquiring unit 301 is configured to perform grammatical detection on the acquired first sentence input by the user by using the preset corpus sentence segmentation function and the preset analysis function to obtain the first sentence detection result, and the first sentence detection result is the first sentence The target dependence relationship between each word in

The second acquisition unit 302 is configured to use the preset corpus sentence segmentation function and the preset analysis function to perform grammatical detection on the acquired second sentence inputted by the user in a new round to obtain the second sentence detection result, and the second sentence detection result is The target dependence relationship between each word in the second sentence;

The first judgment unit 303 is configured to judge whether the detection result of the second sentence includes only a single entity and the second sentence is a question sentence;

The first complementing unit 304 is used for when the second sentence detection result only includes a single entity and the second sentence is a question sentence, according to the first sentence detection result, the semantic missing part of the second sentence is complemented to obtain the first complementary sentence ；

The second judging unit 305 is used to judge whether the first supplementary sentence includes unclear words, which include pronouns, quantifiers, and articles;

The second supplementary unit 306 is configured to, if the first supplementary sentence includes an unknown word, replace the unknown word in the first supplementary sentence according to the first sentence detection result to obtain a second supplementary sentence.

In the embodiment of the present application, the first acquiring unit 301 uses the preset corpus sentence segmentation function and the preset analysis function to perform grammatical detection on the acquired first sentence input by the user to obtain the first sentence detection result, and the first sentence detection result is The target dependency relationship between each word in the first sentence; the second acquiring unit 302 uses the preset corpus sentence segmentation function and the preset analysis function to perform grammatical detection on the acquired second sentence input by the user in the new round, and obtain the first sentence The second sentence detection result, the second sentence detection result is the target dependency relationship between each word in the second sentence; the first judgment unit 303 judges whether the second sentence detection result only includes a single entity and the second sentence is a question sentence; The complementing unit 304, when the detection result of the second sentence only includes a single entity and the second sentence is an interrogative sentence, fills in the missing part of the semantics of the second sentence according to the detection result of the first sentence to obtain the first complementary sentence; the second judgment unit 305 Determine whether the first supplementary sentence includes unknown words, which include pronouns, quantifiers, and articles; the second supplement unit 306, if the first supplementary sentence includes unknown words, then According to the detection result of the first sentence, the unknown words in the first supplementary sentence are replaced to obtain the second supplementary sentence.

Referring to FIG. 4, another embodiment of the device for semantic complement in multiple rounds of dialogue in the embodiment of the present application includes:

The first acquiring unit 301 is configured to use the preset corpus sentence segmentation function and the preset analysis function to perform grammatical detection on the acquired first sentence input by the user to obtain the first sentence detection result, and the first sentence detection result is the first sentence The target dependence relationship between each word in

Optionally, the first obtaining unit 301 includes:

The obtaining module 3011 is used to obtain the first sentence input by the user;

The segmentation module 3012 is used to segment the first sentence by using a preset corpus sentence segmentation function to obtain the first input sentence;

The detection module 3013 is configured to perform grammatical detection on the first input sentence by using a preset analysis function to obtain a first sentence detection result. The first sentence detection result is the target dependency relationship between each word in the first input sentence.

Optionally, the segmentation module 3012 is specifically used for:

Segment the first sentence to obtain a segmented sentence;

Match the segmented corpus in the segmented sentence with the preset corpus. The preset corpus is the corpus built in the preset intent rule base based on business data;

If the segmented corpus matches the preset corpus, segment the segmented sentence before and after the segmented corpus to obtain the segmented sentence, and use the segmented sentence as the first input sentence;

If the segmented corpus does not match the preset corpus, the segmented corpus is directly used as the first input sentence.

Optionally, the detection module 3013 is specifically used for:

Perform part-of-speech tagging and entity extraction on the words in the first input sentence to obtain the first sentence tagging result;

Calculate the dependence probability between each word, the word is the word in the first sentence labeling result, and the dependence probability is the frequency of the preset dependence relationship;

Determine the target dependence relationship between each word, and the preset dependence relationship corresponding to the probability with the largest dependence probability weight is the target dependence relationship between words;

The first sentence detection result is obtained, and the first sentence detection result is the target dependency relationship between each word in the first input sentence.

Optionally, the first complementing unit 304 is specifically used for:

When the second sentence detection result includes only a single entity and the second sentence is an interrogative sentence, extract the single entity in the second sentence detection result, and use the single entity as the first target word;

Screening single entities with the same grammatical structure as the first target word in the first sentence detection result to obtain the second target word;

In the first sentence, replace the second target word with the first target word to obtain the first supplementary sentence.

Optionally, the second judgment unit 305 is specifically configured to:

Acquiring the detection result of the first supplementary sentence, the detection result of the first supplementary sentence being a combination of the second sentence detection result and the first sentence detection result;

Determine whether the detection result of the first supplementary sentence includes pronouns, quantifiers and articles.

Optionally, the second complementing unit 306 is specifically configured to:

If the detection result of the first supplementary sentence includes an unclear word, the unclear word is extracted, and the unclear word is used as the third target word;

Select words with the same grammatical structure as the third target word in the first sentence detection result to obtain the fourth target word;

In the second sentence, replace the third target word with the fourth target word to obtain the second supplementary sentence.

In the embodiment of the present application, the first obtaining unit 301 includes: an obtaining module 3011, configured to obtain the first sentence input by the user; Input sentence; detection module 3013, used for grammatical detection of the first input sentence using a preset analysis function, to obtain the first sentence detection result, the first sentence detection result is the target dependency relationship between each word in the first input sentence ; The second acquisition unit 302 uses the preset corpus sentence segmentation function and the preset analysis function to perform grammatical detection on the acquired second sentence input by the user in the new round, to obtain the second sentence detection result, and the second sentence detection result is the second The target dependency relationship between each word in the sentence; the first judgment unit 303 judges whether the detection result of the second sentence only includes a single entity and the second sentence is a question sentence; the first complement unit 304 when the detection result of the second sentence only includes a single When the entity and the second sentence is an interrogative sentence, according to the detection result of the first sentence, the semantic missing part of the second sentence is supplemented to obtain the first supplementary sentence; the second judgment unit 305 judges whether the first supplementary sentence includes unknown reference The words that refer to unknown words include pronouns, quantifiers, and articles; if the first supplementary sentence includes an unknown word, the second supplementary unit 306 will refer to the middle of the first supplementary sentence according to the detection result of the first sentence Replace the unknown words and get the second supplementary sentence.

The above Figures 3 to 4 describe in detail the semantic complementing device in the multi-round dialogue in the embodiment of the present application from the perspective of the modular functional entity, and the following is the semantic complement in the multi-round dialogue in the embodiment of the present application from the perspective of hardware processing. The equipment is described in detail.

The following is a detailed introduction to the various components of the semantically supplemented device in the multi-round dialogue in conjunction with Figure 5:

FIG. 5 is a schematic structural diagram of a device for semantic completion in a multi-round dialogue provided by an embodiment of the present application. The device 500 for semantic completion in a multi-round dialogue may have relatively large differences due to different configurations or performance, which may include One or more processors (central processing units, CPU) 501 (for example, one or more processors) and memory 509, and one or more storage media 508 (for example, one or more storage mediums 508 for storing application programs 507 or data 506) 501 (for example, one or more processors) Storage device). Among them, the memory 509 and the storage medium 508 may be short-term storage or persistent storage. The program stored in the storage medium 508 may include one or more modules (not shown in the figure), and each module may include a series of instruction operations in a device that complements semantics in multiple rounds of dialogue. Further, the processor 501 may be configured to communicate with the storage medium 508, and execute a series of instruction operations in the storage medium 508 on the device 500 with semantic completion in multiple rounds of dialogue.

The device 500 for supplementing semantics in multiple rounds of dialogue may also include one or more power sources 502, one or more wired or wireless network interfaces 503, one or more input and output interfaces 504, and/or, one or more operating systems 505, such as Windows Serve, Mac OS X, Unix, Linux, FreeBSD, etc. Those skilled in the art can understand that the device structure for semantic completion in multiple rounds of dialogue shown in FIG. 5 does not constitute a limitation on the device for semantic completion in multiple rounds of dialogue, and may include more or less than that shown in the figure. Components, or a combination of certain components, or different component arrangements.

The processor 501 is the control center of the device that completes semantics in multiple rounds of dialogue, and can perform processing in accordance with the method of semantic complement in multiple rounds of dialogue. The processor 501 uses various interfaces and lines to connect the various parts of the semantically supplemented device in the entire multi-round dialogue, by running or executing software programs and/or modules stored in the memory 509, and calling data stored in the memory 509 , Use preset corpus sentence segmentation function and preset analysis function to grammatically detect multiple rounds of dialogue input by users, and use the semantics of multiple rounds of dialogue context to complement sentences with incomplete semantics, which improves the accuracy of semantic analysis results. The storage medium 508 and the memory 509 are both carriers for storing data. In the embodiment of the present application, the storage medium 508 may refer to an internal memory with a small storage capacity but a fast speed, and the storage medium 509 may have a large storage capacity but a slow storage speed. External memory.

The memory 509 may be used to store software programs and modules. The processor 501 executes various functional applications and data processing of the device 500 with semantic complement in multiple rounds of dialogues by running the software programs and modules stored in the memory 509. The memory 509 may mainly include a storage program area and a storage data area. The storage program area may store an operating system, an application program required by at least one function, etc.; The created data, etc. In addition, the memory 509 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one magnetic disk storage device, a flash memory device, or other non-volatile solid-state storage devices. In the multi-round dialogue provided in the embodiment of the present application, the semantically supplemented program and the received data stream are stored in the memory, and the processor 501 is called from the memory 509 when it needs to be used.

When the computer program instructions are loaded and executed on the computer, the procedures or functions according to the embodiments of the present application are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. Computer instructions may be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, computer instructions may be transmitted from a website, computer, server, or data center through a cable (such as Coaxial cable, optical fiber, twisted pair) or wireless (such as infrared, wireless, microwave, etc.) to transmit to another website site, computer, server or data center. The computer-readable storage medium may be any available medium that can be stored by a computer or a data storage device such as a server or a data center integrated with one or more available media. The usable medium may be a magnetic medium (for example, a floppy disk, a hard disk, and a magnetic tape), an optical medium (for example, an optical disc), or a semiconductor medium (for example, a solid state disk (SSD)).

Optionally, the computer-readable storage medium (or storage medium) may be non-volatile or volatile.

Those skilled in the art can clearly understand that, for the convenience and conciseness of the description, the specific working process of the system, device and unit described above can refer to the corresponding process in the foregoing method embodiment, which will not be repeated here.

Above, the above embodiments are only used to illustrate the technical solutions of the present application, not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, a person of ordinary skill in the art should understand that it can still implement the foregoing various implementations. The technical solutions recorded in the examples are modified, or some of the technical features are equivalently replaced; these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

A method of semantic completion in multiple rounds of dialogue, which includes:

Use the preset corpus sentence segmentation function and the preset analysis function to perform grammatical detection on the acquired first sentence input by the user to obtain the first sentence detection result, and the first sentence detection result is each word in the first sentence The goal dependence relationship between;

Use the preset corpus sentence segmentation function and the preset analysis function to perform grammatical detection on the acquired second sentence inputted by the user in a new round to obtain a second sentence detection result, and the second sentence detection result is the The target dependence relationship between each word in the second sentence;

Judging whether the detection result of the second sentence includes only a single entity and the second sentence is a question sentence;

When the second sentence detection result includes only a single entity and the second sentence is an interrogative sentence, fill in the semantic missing part of the second sentence according to the first sentence detection result to obtain a first supplementary sentence;

Judging whether the first supplementary sentence includes unidentified words, the unidentified words including pronouns, quantifiers, and articles;

If the first supplementary sentence includes an unidentified word, the unidentified word in the first supplementary sentence is replaced according to the detection result of the first sentence to obtain a second supplementary sentence.
The method according to claim 1, wherein said using preset corpus sentence segmentation function and preset analysis function to perform grammatical detection on the acquired first sentence input by the user to obtain the first sentence detection result, said first sentence The sentence detection result is that the target dependency relationship between each word in the first sentence includes:

Get the first sentence entered by the user;

Use a preset corpus sentence segmentation function to segment the first sentence to obtain the first input sentence;

A preset analysis function is used to perform grammatical detection on the first input sentence to obtain a first sentence detection result. The first sentence detection result is a target dependency relationship between each word in the first input sentence.
3. The method according to claim 2, wherein said using a preset corpus sentence segmentation function to segment the first sentence to obtain the first input sentence comprises:

Segmenting the first sentence to obtain a segmented sentence;

Matching the segmented corpus in the segmented sentence with a preset corpus, the preset corpus is a corpus established in a preset intent rule base based on business data;

If the segmented corpus matches the preset corpus, segment the segmented sentence at the front and back positions of the segmented corpus to obtain a segmented sentence, and use the segmented sentence as the first input sentence;

If the segmented corpus does not match the preset corpus, the segmented corpus is directly used as the first input sentence.
3. The method according to claim 2, wherein the grammatical detection of the first input sentence is performed using a preset analysis function to obtain a first sentence detection result, and the first sentence detection result is the first input sentence The target dependence relationship between each word in the include:

Performing part-of-speech tagging and entity extraction on words in the first input sentence to obtain a first sentence tagging result;

Calculating the probability of dependence between each of the words, where the word is a word in the first sentence labeling result, and the dependence probability is the frequency of occurrence of a preset dependence relationship;

Determining the target dependency relationship between each of the words, and the preset dependency relationship corresponding to the probability with the largest dependency probability weight is the target dependency relationship between the words;

Obtain a first sentence detection result, where the first sentence detection result is the target dependency relationship between each of the words in the first input sentence.
The method according to claim 1, wherein when the second sentence detection result only includes a single entity and the second sentence is an interrogative sentence, the second sentence is determined according to the first sentence detection result. The missing part of semantics is completed, and the first completed sentence includes:

When the second sentence detection result includes only a single entity and the second sentence is an interrogative sentence, extract the single entity in the second sentence detection result, and use the single entity as the first target word;

Screening the single entity with the same grammatical structure as the first target word from the first sentence detection result to obtain a second target word;

In the first sentence, the first target word is used to replace the second target word to obtain a first supplementary sentence.
The method according to any one of claims 1 to 5, wherein said determining whether said first supplementary sentence includes unclear words, said unclear words including pronouns, quantifiers and articles including :

Acquiring a detection result of a first supplementary sentence, where the detection result of the first supplementary sentence is a combination of the second sentence detection result and the first sentence detection result;

It is determined whether the detection result of the first supplementary sentence includes pronouns, quantifiers and articles.
7. The method according to claim 6, wherein if the first supplementary sentence includes an unknown term, then the first supplementary sentence refers to the unknown term according to the first sentence detection result Word replacement, get the second complement sentence including:

If the detection result of the first supplementary sentence includes an unclear word, extract the unclear word, and use the unclear word as the third target word;

Screening words with the same grammatical structure as the third target word from the first sentence detection result to obtain the fourth target word;

In the second sentence, the fourth target word is used to replace the third target word to obtain a second supplementary sentence.
A device for complementing semantics in multiple rounds of dialogue, which includes:

The first acquisition unit is configured to use the preset corpus sentence segmentation function and the preset analysis function to perform grammatical detection on the acquired first sentence input by the user to obtain the first sentence detection result, and the first sentence detection result is the The target dependence relationship between each word in the first sentence;

The second acquisition unit is configured to use the preset corpus sentence segmentation function and the preset analysis function to perform grammatical detection on the acquired second sentence input by the user in a new round to obtain the second sentence detection result. The second sentence detection result is the target dependency relationship between each word in the second sentence;

The first judgment unit is configured to judge whether the detection result of the second sentence includes only a single entity and the second sentence is a question sentence;

The first complementing unit is used to fill in the missing semantic parts of the second sentence according to the first sentence detection result when the second sentence detection result only includes a single entity and the second sentence is a question sentence , Get the first supplementary sentence;

The second judging unit is used to judge whether the first supplementary sentence includes unclear words, and the unclear words include pronouns, quantifiers and articles;

The second supplementary unit is used to replace the unknown word in the first supplementary sentence according to the detection result of the first sentence if the first supplementary sentence includes an unknown word to obtain the first supplementary sentence. Two complete sentences.
A device for complementing semantics in multiple rounds of dialogue, which includes:

A memory and at least one processor, where instructions are stored in the memory, and the memory and the at least one processor are interconnected by wires;

The at least one processor invokes the instructions in the memory, so that the semantically-completed device in the multi-round dialogue executes the following steps:

Use the preset corpus sentence segmentation function and the preset analysis function to perform grammatical detection on the acquired first sentence input by the user to obtain the first sentence detection result, and the first sentence detection result is each word in the first sentence The goal dependence relationship between;

Use the preset corpus sentence segmentation function and the preset analysis function to perform grammatical detection on the acquired second sentence input by the user in a new round to obtain a second sentence detection result, and the second sentence detection result is the The target dependence relationship between each word in the second sentence;

Judging whether the detection result of the second sentence includes only a single entity and the second sentence is a question sentence;

When the second sentence detection result includes only a single entity and the second sentence is an interrogative sentence, fill in the semantic missing part of the second sentence according to the first sentence detection result to obtain a first supplementary sentence;

Judging whether the first supplementary sentence includes unidentified words, the unidentified words including pronouns, quantifiers, and articles;

If the first supplementary sentence includes an unidentified word, the unidentified word in the first supplementary sentence is replaced according to the detection result of the first sentence to obtain a second supplementary sentence.
9. The device according to claim 9, wherein the preset corpus segmentation function and the preset analysis function are used to perform grammatical detection on the acquired first sentence input by the user, and when the first sentence detection result is obtained, the following is specifically executed step:

Get the first sentence entered by the user;

Use a preset corpus sentence segmentation function to segment the first sentence to obtain the first input sentence;

A preset analysis function is used to perform grammatical detection on the first input sentence to obtain a first sentence detection result. The first sentence detection result is a target dependency relationship between each word in the first input sentence.
11. The device according to claim 10, wherein when said first sentence is segmented using a preset corpus sentence segmentation function to obtain the first input sentence, the following steps are specifically executed:

Segmenting the first sentence to obtain a segmented sentence;

Matching the segmented corpus in the segmented sentence with a preset corpus, the preset corpus is a corpus established in a preset intent rule base based on business data;

If the segmented corpus matches the preset corpus, segment the segmented sentence at the front and back positions of the segmented corpus to obtain a segmented sentence, and use the segmented sentence as the first input sentence;

If the segmented corpus does not match the preset corpus, the segmented corpus is directly used as the first input sentence.
10. The device according to claim 10, wherein the grammatical detection of the first input sentence is performed by using a preset analysis function, and when the first sentence detection result is obtained, the following steps are specifically executed:

Performing part-of-speech tagging and entity extraction on words in the first input sentence to obtain a first sentence tagging result;

Calculating the probability of dependence between each of the words, where the word is a word in the first sentence labeling result, and the dependence probability is the frequency of occurrence of a preset dependence relationship;

Determining the target dependency relationship between each of the words, and the preset dependency relationship corresponding to the probability with the largest dependency probability weight is the target dependency relationship between the words;

Obtain a first sentence detection result, where the first sentence detection result is the target dependency relationship between each of the words in the first input sentence.
The device according to claim 9, wherein when the second sentence detection result includes only a single entity and the second sentence is a question sentence, the second sentence is determined according to the first sentence detection result. When the missing part of semantics is completed and the first completed sentence is obtained, the following steps are specifically performed:

When the second sentence detection result includes only a single entity and the second sentence is an interrogative sentence, extract the single entity in the second sentence detection result, and use the single entity as the first target word;

Screening the single entity with the same grammatical structure as the first target word from the first sentence detection result to obtain a second target word;

In the first sentence, the first target word is used to replace the second target word to obtain a first supplementary sentence.
The device according to any one of claims 9-13, wherein when determining whether the first supplementary sentence includes an unclear word, the following steps are specifically executed:

Acquiring a detection result of a first supplementary sentence, where the detection result of the first supplementary sentence is a combination of the second sentence detection result and the first sentence detection result;

It is determined whether the detection result of the first supplementary sentence includes pronouns, quantifiers and articles.
14. The device according to claim 14, wherein if the first supplementary sentence includes an unknown term, the first supplementary sentence refers to the unknown term according to the first sentence detection result For word replacement, when the second supplementary sentence is obtained, the following steps are specifically performed:

If the detection result of the first supplementary sentence includes an unclear word, extract the unclear word, and use the unclear word as the third target word;

Screening words with the same grammatical structure as the third target word from the first sentence detection result to obtain the fourth target word;

In the second sentence, the fourth target word is used to replace the third target word to obtain a second supplementary sentence.
A computer-readable storage medium, which includes instructions, when the instructions run on a computer, cause the computer to perform the following steps:

Use the preset corpus sentence segmentation function and the preset analysis function to perform grammatical detection on the acquired first sentence input by the user to obtain the first sentence detection result, and the first sentence detection result is each word in the first sentence The goal dependence relationship between;

Use the preset corpus sentence segmentation function and the preset analysis function to perform grammatical detection on the acquired second sentence input by the user in a new round to obtain a second sentence detection result, and the second sentence detection result is the The target dependence relationship between each word in the second sentence;

Judging whether the detection result of the second sentence includes only a single entity and the second sentence is a question sentence;

When the second sentence detection result includes only a single entity and the second sentence is an interrogative sentence, fill in the semantic missing part of the second sentence according to the first sentence detection result to obtain a first supplementary sentence;

Judging whether the first supplementary sentence includes unidentified words, the unidentified words including pronouns, quantifiers, and articles;

If the first supplementary sentence includes an unidentified word, the unidentified word in the first supplementary sentence is replaced according to the detection result of the first sentence to obtain a second supplementary sentence.
The computer-readable storage medium according to claim 16, wherein the grammatical check is performed on the obtained first sentence input by the user by using the preset corpus sentence segmentation function and the preset analysis function, and when the first sentence detection result is obtained , Perform the following steps:

Get the first sentence entered by the user;

Use a preset corpus sentence segmentation function to segment the first sentence to obtain the first input sentence;

A preset analysis function is used to perform grammatical detection on the first input sentence to obtain a first sentence detection result. The first sentence detection result is a target dependency relationship between each word in the first input sentence.
18. The computer-readable storage medium according to claim 17, wherein when the first sentence is segmented by using a preset corpus sentence segmentation function to obtain the first input sentence, the following steps are specifically executed:

Segmenting the first sentence to obtain a segmented sentence;

Matching the segmented corpus in the segmented sentence with a preset corpus, the preset corpus is a corpus established in a preset intent rule base based on business data;

If the segmented corpus matches the preset corpus, segment the segmented sentence at the front and back positions of the segmented corpus to obtain a segmented sentence, and use the segmented sentence as the first input sentence;

If the segmented corpus does not match the preset corpus, the segmented corpus is directly used as the first input sentence.
18. The computer-readable storage medium according to claim 17, wherein the grammatical detection of the first input sentence is performed by using a preset analysis function, and when the first sentence detection result is obtained, the following steps are specifically executed:

Performing part-of-speech tagging and entity extraction on words in the first input sentence to obtain a first sentence tagging result;

Calculating the probability of dependence between each of the words, where the word is a word in the first sentence labeling result, and the dependence probability is the frequency of occurrence of a preset dependence relationship;

Determining the target dependency relationship between each of the words, and the preset dependency relationship corresponding to the probability with the largest dependency probability weight is the target dependency relationship between the words;

Obtain a first sentence detection result, where the first sentence detection result is the target dependency relationship between each of the words in the first input sentence.
The computer-readable storage medium according to claim 16, wherein when the second sentence detection result includes only a single entity and the second sentence is a question sentence, the first sentence detection result is Complete the missing part of the semantics of the second sentence. When the first supplementary sentence is obtained, the following steps are specifically performed:

When the second sentence detection result includes only a single entity and the second sentence is an interrogative sentence, extract the single entity in the second sentence detection result, and use the single entity as the first target word;

Screening the single entity with the same grammatical structure as the first target word from the first sentence detection result to obtain a second target word;

In the first sentence, the first target word is used to replace the second target word to obtain a first supplementary sentence.