Dialogue sentence processing method and device, storage medium and electronic device

Abstract

The application discloses a dialogue sentence processing method and device, a storage medium and an electronic device, and relates to the technical field of smart home / smart home, and the method comprises the steps of obtaining a current dialogue sentence, wherein the current dialogue sentence is a dialogue sentence input by a target object in a current round dialogue; in the case that it is determined that the current dialogue sentence is missing state information through slot extraction on the current dialogue sentence, determining target state information matched with the missing state information in state information of a group of dialogue sentences, wherein the state information comprises at least one of the following: a sentence field, a sentence intention, and dialogue sentences input by the target object in at least one round dialogue before the current round dialogue; and using the target state information to complete the missing state information in the current dialogue sentence, to obtain the current dialogue sentence after completion.

Description

Technical Field

[0001] This application relates to the field of smart home / intelligent home technology, and more specifically, to a method and apparatus for processing dialogue statements, a storage medium, and an electronic device. Background Technology

[0002] Currently, human-computer dialogues frequently involve continuous multi-turn conversations, during which users often omit information. Because these omitted sentences lack complete information, they need to be rewritten or supplemented to improve the machine's understanding of the dialogue. This can be achieved by parsing the user's current query (the user's chat input in the human-computer dialogue) to determine the missing sentence components, and then supplementing them based on the missing components and the syntactic analysis results above.

[0003] However, because the content of a user's dialogue is diverse in multi-turn queries—for example, controlling one smart device with one query, then controlling another smart device with another query, and then controlling the previous smart device with one query—the accuracy of supplementing missing sentence components based solely on syntactic analysis is low.

[0004] Therefore, it is evident that the processing methods for dialogue statements in related technologies suffer from low accuracy in supplementary information due to the diversity of dialogue content. Summary of the Invention

[0005] This application provides a method and apparatus for processing dialogue statements, a storage medium, and an electronic device to at least solve the problem of low accuracy in multi-turn dialogue statement completion methods in related technologies due to ignoring multi-turn state information.

[0006] According to one aspect of the embodiments of this application, a method for processing dialogue statements is provided, comprising: obtaining a current dialogue statement, wherein the current dialogue statement is a dialogue statement input by a target object in a current round of dialogue; determining, by performing slot extraction on the current dialogue statement and determining missing state information of the current dialogue statement, a target state information that matches the missing state information in a set of dialogue statements, wherein the state information includes at least one of the following: statement domain, statement intent, the set of dialogue statements being dialogue statements input by the target object in at least one round of dialogue prior to the current round of dialogue; and using the target state information to complete the missing state information in the current dialogue statement to obtain the completed current dialogue statement.

[0007] In an exemplary embodiment, determining the target state information that matches the missing state information in the state information of a set of dialogue statements includes: when the missing state information is a statement domain and the statement intent of the current dialogue statement is determined, determining the statement domain in the state information of the set of dialogue statements as a set of candidate statement domains; and determining the candidate statement domain that matches the statement intent of the current dialogue statement in the set of candidate statement domains according to a preset correspondence between statement domains and statement intents, thereby obtaining the target statement domain, wherein the target state information is the target statement domain.

[0008] In an exemplary embodiment, determining the target state information that matches the missing state information in the state information of a set of dialogue statements includes: when the missing state information is a statement intent and the statement domain of the current dialogue statement is determined, determining the statement intent in the state information of the set of dialogue statements as a set of candidate statement intents; and determining the candidate statement intent that matches the statement domain of the current dialogue statement in the set of candidate statement intents according to a preset correspondence between statement domains and statement intents, thereby obtaining the target statement intent, wherein the target state information is the target statement intent.

[0009] In an exemplary embodiment, determining the target state information that matches the missing state information among the state information of a set of dialogue statements includes: determining the state information that matches the missing state information among the state information of the set of dialogue statements to obtain multiple candidate state information; and determining the candidate state information whose dialogue statement has the shortest statement interval time with the current dialogue statement among the multiple candidate state information as the target state information.

[0010] In an exemplary embodiment, the step of using the target state information to complete the missing state information in the current dialogue statement to obtain the completed current dialogue statement includes: performing syntactic analysis on the current dialogue statement to determine the missing syntactic relationship corresponding to the missing state information; determining the target statement position in the current dialogue statement to which the target state information is to be inserted based on the missing syntactic relationship; and inserting the target state information at the target statement position in the current dialogue statement to obtain the completed current dialogue statement.

[0011] In an exemplary embodiment, determining the target statement position in the current dialogue statement where the target state information is to be inserted based on the missing syntactic relation includes: when the missing syntactic relation is a subject-predicate relation and a pronoun exists in the current dialogue statement, determining the position after the pronoun as the target statement position in the current dialogue statement where the target state information is to be inserted; when the missing syntactic relation is a subject-predicate relation and no pronoun exists in the current dialogue statement, determining the beginning position of the current dialogue statement as the target statement position in the current dialogue statement where the target state information is to be inserted.

[0012] In an exemplary embodiment, after obtaining the current dialogue statement, the method further includes: if the statement domain and statement intent of the current dialogue statement are determined by slot extraction of the current dialogue statement, searching for a preset correspondence between the statement domain and statement intent according to the statement domain and statement intent of the current dialogue statement; and if a correspondence matching the statement domain and statement intent of the current dialogue statement is found, performing a target device operation corresponding to the current dialogue statement.

[0013] According to another aspect of the embodiments of this application, a processing apparatus for dialogue statements is also provided, comprising: an acquisition unit, configured to acquire a current dialogue statement, wherein the current dialogue statement is a dialogue statement input by a target object in a current round of dialogue; a first determination unit, configured to, when determining missing state information of the current dialogue statement by performing slot extraction on the current dialogue statement, determine a target state information that matches the missing state information in a set of dialogue statements, wherein the state information includes at least one of the following: statement domain, statement intent, the set of dialogue statements being dialogue statements input by the target object in at least one round of dialogue prior to the current round of dialogue; and a first execution unit, configured to use the target state information to complete the missing state information in the current dialogue statement to obtain the completed current dialogue statement.

[0014] In an exemplary embodiment, the first determining unit includes: a first determining module, configured to determine the statement domain in the state information of the set of dialogue statements as a set of candidate statement domains when the missing state information is a statement domain and the statement intent of the current dialogue statement is determined; and a second determining module, configured to determine, according to a preset correspondence between statement domains and statement intents, the candidate statement domain that matches the statement intent of the current dialogue statement in the set of candidate statement domains, thereby obtaining a target statement domain, wherein the target state information is the target statement domain.

[0015] In an exemplary embodiment, the first determining unit includes: a third determining module, configured to determine the statement intent in the state information of the set of dialogue statements as a set of candidate statement intents when the missing state information is a statement intent and the statement domain of the current dialogue statement is determined; and a fourth determining module, configured to determine, according to a preset correspondence between statement domains and statement intents, the candidate statement intent that matches the statement domain of the current dialogue statement in the set of candidate statement intents, thereby obtaining a target statement intent, wherein the target state information is the target statement intent.

[0016] In an exemplary embodiment, the first determining unit includes: a fifth determining module, configured to determine, among the state information of the set of dialogue statements, state information that matches the missing state information, to obtain multiple candidate state information; and a sixth determining module, configured to determine, among the multiple candidate state information, the candidate state information whose dialogue statement has the shortest statement interval time with the current dialogue statement as the target state information.

[0017] In an exemplary embodiment, the first execution unit includes: a seventh determining module, configured to perform syntactic analysis on the current dialogue statement to determine the missing syntactic relation corresponding to the missing state information; an eighth determining module, configured to determine the target statement position in the current dialogue statement to be inserted based on the missing syntactic relation; and an insertion module, configured to insert the target state information at the target statement position in the current dialogue statement to obtain the completed current dialogue statement.

[0018] In an exemplary embodiment, the eighth determining module includes: a first determining submodule, configured to determine the position after the pronoun as the target statement position in the current dialogue statement where the target state information is to be inserted when the missing syntactic relation is a subject-predicate relation and there is a pronoun in the current dialogue statement; and a second determining submodule, configured to determine the beginning position of the current dialogue statement as the target statement position in the current dialogue statement where the target state information is to be inserted when the missing syntactic relation is a subject-predicate relation and there is no pronoun in the current dialogue statement.

[0019] In one exemplary embodiment, the apparatus further includes: a second determining unit, configured to, after acquiring the current dialogue statement, and after determining the statement domain and statement intent of the current dialogue statement by performing slot extraction on the current dialogue statement, search for a preset correspondence between the statement domain and statement intent of the current dialogue statement; and a second executing unit, configured to, if a correspondence matching the statement domain and statement intent of the current dialogue statement is found, execute a target device operation corresponding to the current dialogue statement.

[0020] According to another aspect of the embodiments of this application, a computer-readable storage medium is also provided, wherein a computer program is stored in the computer program, and the computer program is configured to execute the above-described processing method for dialogue statements when it is run.

[0021] According to another aspect of the embodiments of this application, an electronic device is also provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the above-described dialogue statement processing method through the computer program.

[0022] In this embodiment, a method is adopted that determines the state information of the currently input statement that needs to be supplemented, and supplements the missing state information based on the state information of previous rounds. This is achieved by obtaining the current dialogue statement, which is the dialogue statement input by the target object in the current round of dialogue; after extracting slots from the current dialogue statement to determine the missing state information, a target state information matching the missing state information is determined from a set of dialogue statement state information. This state information includes at least one of the following: statement domain, statement intent, and a set of dialogue statements included in at least one round of dialogue prior to the current round, and the target object inputting the dialogue statement. The target state information is then used to supplement the missing state information in the current dialogue statement, resulting in the supplemented current dialogue statement. By determining the state information of the current statement and previous rounds of statements and supplementing the statement based on the state information, the aim of accurately supplementing dialogues with diverse content can be achieved, thus improving the accuracy of the supplemented content. This solves the technical problem in related technologies where the accuracy of supplemented content is low due to the diversity of dialogue content. Attached Figure Description

[0023] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0024] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0025] Figure 1 This is a schematic diagram of the hardware environment for a method of processing dialogue statements according to an embodiment of this application.

[0026] Figure 2 This is a flowchart illustrating an optional method for processing dialogue statements according to an embodiment of this application.

[0027] Figure 3 This is a flowchart illustrating another optional method for processing dialogue statements according to an embodiment of this application;

[0028] Figure 4 This is a structural block diagram of an optional dialogue statement processing device according to an embodiment of this application;

[0029] Figure 5 This is a structural block diagram of an optional electronic device according to an embodiment of this application. Detailed Implementation

[0030] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0031] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0032] According to one aspect of the embodiments of this application, a method for processing dialogue statements is provided. This method is widely applicable to whole-house intelligent digital control application scenarios such as smart homes, smart home ecosystems, and intelligence house ecosystems. Optionally, in this embodiment, the above-mentioned dialogue statement processing method can be applied to, for example... Figure 1 The hardware environment shown consists of terminal device 102 and server 104. For example... Figure 1 As shown, server 104 is connected to terminal device 102 via a network and can be used to provide services (such as application services) to the terminal or clients installed on the terminal. A database can be set up on the server or independently of the server to provide data storage services for server 104. Cloud computing and / or edge computing services can be configured on the server or independently of the server to provide data processing services for server 104.

[0033] The aforementioned network may include, but is not limited to, at least one of the following: wired network, wireless network. The aforementioned wired network may include, but is not limited to, at least one of the following: wide area network, metropolitan area network, local area network. The aforementioned wireless network may include, but is not limited to, at least one of the following: Wi-Fi (Wireless Fidelity), Bluetooth. Terminal device 102 may not be limited to PCs, mobile phones, tablets, smart air conditioners, smart range hoods, smart refrigerators, smart ovens, smart stoves, smart washing machines, smart water heaters, smart washing equipment, smart dishwashers, smart projectors, smart TVs, smart clothes racks, smart curtains, smart audio-visual equipment, smart sockets, smart speakers, smart speakers, smart fresh air systems, smart kitchen and bathroom equipment, smart bathroom equipment, smart robot vacuum cleaners, smart window cleaning robots, smart mopping robots, smart air purifiers, smart steam ovens, smart microwave ovens, smart water heaters, smart air purifiers, smart water dispensers, smart door locks, and other voice-interactive devices.

[0034] The dialogue statement processing method of this application embodiment can be executed by server 104, terminal device 102, or jointly by server 104 and terminal device 102. Alternatively, the dialogue statement processing method of this application embodiment can be executed by a client installed on terminal device 102.

[0035] Taking the method of processing dialogue statements in this embodiment executed by server 104 as an example, Figure 2 This is a flowchart illustrating an optional method for processing dialogue statements according to an embodiment of this application, as shown below. Figure 2 As shown, the process of this method may include the following steps:

[0036] Step S202: Obtain the current dialogue statement, where the current dialogue statement is the dialogue statement entered by the target object in the current round of dialogue.

[0037] The dialogue processing method in this embodiment can be applied to scenarios where users engage in intelligent dialogue with terminal devices. The terminal device can be an intelligent device capable of recognizing user speech and providing voice feedback or performing related operations based on the recognition results. Examples include question-answering robots, task robots, and chatbots. These devices can be equipped with intelligent dialogue systems to support dialogue with users; these systems can be multi-turn dialogue systems. The question-answering robot relies primarily on a robust knowledge base and can provide specific responses to user questions. The task robot can fulfill a user's specific task needs through multi-turn dialogue interaction. It can complete user task requirements by responding after understanding the user's intent or by calling APIs (Application Programming Interfaces), such as booking tickets or ordering food. The chatbot can interact with users, satisfying their emotional needs through engaging and personalized responses. Furthermore, the terminal device can also be a comprehensive type of dialogue robot, a combination of the above dialogue types, capable of simultaneously satisfying multiple user needs such as question answering, task completion, or casual conversation.

[0038] In the aforementioned intelligent dialogue scenarios, multi-turn dialogue tasks are predominant. Compared to single-turn dialogues, users frequently use ellipsis in multi-turn dialogues. For example, the dialogue information in subsequent turns of a multi-turn dialogue may be incomplete, omitting identical information from previous turns. Directly analyzing and understanding incomplete elliptical dialogue statements due to the lack of information can lead to errors in the terminal device's understanding of the multiple turns. Therefore, for elliptical dialogue statements in multi-turn dialogues, it is necessary to first complete the statements before interpreting them.

[0039] In related technologies, multi-turn state tracking is generally used, or simple rules are used for rewriting, and deep learning methods are also used for direct rewriting. However, the above-mentioned multi-turn state tracking solutions require parsing information such as the domain and intent of the current user query before updating it to the multi-turn state and making a decision. This is lagging in the entire dialogue system processing flow, and ellipsis itself will reduce the accuracy of domain and intent understanding. Based on simple rules without combining grammatical knowledge for completion, the completion position will be incorrect or the completion content will be incomplete, thus introducing more errors and the accuracy of sentence completion is not high. Deep learning end-to-end solutions require a large amount of domain-related multi-turn data, which requires a large amount of multi-turn data for model training. The training cycle is long, and a lot of manpower is required for multi-turn data collection and processing, which is very costly and has a very high domain coupling.

[0040] Alternatively, a scheme entirely based on syntactic parsing can be used for sentence completion. This scheme typically uses an open-source parser to analyze the syntax of the current query, and then completes the sentence based on the missing sentence components and the syntactic analysis results from the preceding context. However, the problem with a scheme entirely based on open-source parsing is that, on the one hand, the accuracy of open-source parsers decreases in vertical domains and colloquial dialogues; on the other hand, it does not make good use of the multi-turn state understanding information from the preceding context.

[0041] To address the aforementioned issues, this embodiment provides a domain-independent completion scheme. Combining the understanding results from the preceding text, it utilizes the sentence components and the domains and intentions from multiple turn states to identify and complete the domains and intentions in the current dialogue statement. This maximizes efficiency in terms of time and resources while achieving the same accuracy. For each turn's dialogue statement input by the target object in a multi-turn dialogue system, analysis can be performed to determine if any parts require completion. The target object can identify the user bound to the terminal device executing the dialogue, or other users who need to interact with the terminal device. For statements determined not to require completion, corresponding operations can be performed directly based on the understanding of the statement. These operations can be providing a corresponding voice response to the target object's input, or performing relevant task operations directly based on the content of the target object's input. For statements determined to require completion, the relevant statement completion can be performed first, followed by statement understanding and subsequent operations.

[0042] In this embodiment, the current dialogue statement can be obtained through the voice acquisition component on the terminal device. The current dialogue statement can be the dialogue statement entered by the target object in the current round of dialogue, i.e., the current query. The current dialogue statement can be obtained by performing voice parsing on the voice data collected by the voice acquisition component to extract the current dialogue statement. The parsing operation can be performed by the terminal device or by the server; this embodiment does not limit this.

[0043] Step S204: If the missing state information of the current dialogue statement is determined by slot extraction of the current dialogue statement, the target state information that matches the missing state information in a set of dialogue statements is determined. The state information includes at least one of the following: statement domain, statement intent, a set of dialogue statements contained in at least one round of dialogue before the current round of dialogue, and dialogue statements input by the target object.

[0044] In multi-turn dialogues, the content of each turn's dialogue statement from the target user can be diverse. For example, in a home setting, the target user might control a smart device with one query, then another smart device with a different query, and so on. In such cases, state information such as domain recognition and intent recognition can be performed on each turn's dialogue statement. Based on the recognized domain and intent, the task specified by the target user can be determined. Domain recognition identifies the domain expressed by the user, and intent recognition identifies the user's expressed intent.

[0045] In this embodiment, the understanding of the target object's statement is primarily determined based on the recognition results of state information such as domain and intent. For the current dialogue statement, corresponding analysis can be performed to determine whether there are any missing parts that need to be supplemented. For example, slot extraction can be performed on the current dialogue statement to determine whether it contains complete domain and intent information, thereby determining whether the current dialogue statement is missing state information, and further determining whether the current dialogue statement needs to be supplemented. The state information here may include, but is not limited to, at least one of the following: statement domain, statement intent, and may also include at least one of the following: system information, target object profile, task status, etc. Here, the statement domain can be the domain in the current dialogue statement, and the statement intent can be the intent of the target object in the current dialogue statement. For example, if the current dialogue statement is: "The air conditioner temperature increases by 2 degrees," the statement domain can be "air conditioner," and the statement intent can be "the temperature increases by 2 degrees." Another example is: "Recommend several top-ranked restaurants nearby," where the statement domain can be "restaurants," and the statement intent can be "recommend top-ranked restaurants."

[0046] The state information of the current dialogue statement can be missing. This missing state information can be the statement domain, the statement intent, or both. Once the missing state information of the current dialogue statement is determined, a target state information matching the missing state information can be identified from a set of dialogue statement state information. This set of dialogue statements can include dialogue statements input by the target object in at least one previous dialogue round. The target state information can be state information used to complete the current dialogue statement by matching the missing state information of the current dialogue statement.

[0047] Step S206: Use the target state information to complete the missing state information in the current dialogue statement to obtain the completed current dialogue statement.

[0048] In this embodiment, the determined target state information can be used as the missing state information in the current dialogue statement, and the target state information can be placed at the completion position corresponding to the missing state information in the current dialogue statement. By using the target state information to complete the missing state information in the current dialogue statement, the completed current dialogue statement can be obtained.

[0049] Optionally, the current dialogue statement can be parsed using a syntactic parser to determine the grammatical structure corresponding to the missing state information and its completion position within the current dialogue statement. Based on the determined completion position, the target state information is directly added to the completion position in the current dialogue statement.

[0050] For example, for the current query input by the user, syntactic parsing can be performed to determine missing sentence components based on the sentence's linguistic structure. If missing components are found, slot extraction is performed on the current query, analyzing the intent and domain within the current state information. Queries lacking intent or domain are then supplemented. When supplementing missing content, the content of the user's previous queries can be analyzed and extracted to obtain multi-turn state information. Combining this information with the previous queries' content and multi-turn state information, the supplementary content for the current query is determined, and the corresponding sentence components from the preceding context are added to the current query. By combining contextual information, multi-turn states, and grammatical knowledge, user query completion can be achieved relatively accurately, without relying on large amounts of data to train the model or manual annotation, offering significant advantages in terms of time, cost, and effectiveness.

[0051] Through steps S202 to S206, the current dialogue statement is obtained, wherein the current dialogue statement is the dialogue statement input by the target object in the current round of dialogue; after determining the missing state information of the current dialogue statement by extracting slots from the current dialogue statement, the target state information that matches the missing state information in a set of dialogue statements is determined, wherein the state information includes at least one of the following: statement domain, statement intent, a set of dialogue statements contained in at least one round of dialogue before the current round of dialogue, and dialogue statements input by the target object; the target state information is used to complete the missing state information in the current dialogue statement to obtain the completed current dialogue statement. This solves the technical problem of low accuracy of supplementary content due to the diversity of dialogue content in the processing methods of dialogue statements in related technologies, and improves the accuracy of supplementary content.

[0052] In one exemplary embodiment, determining the target state information that matches the missing state information from a set of dialogue statements' state information includes:

[0053] S11, when the missing state information is a statement domain and the statement intent of the current dialogue statement is determined, the statement domain in the state information of a set of dialogue statements is determined as a set of candidate statement domains.

[0054] S12, based on the preset correspondence between statement domains and statement intents, determine the candidate statement domain that matches the statement intent of the current dialogue statement from a set of candidate statement domains, and obtain the target statement domain, wherein the target state information is the target statement domain.

[0055] In this embodiment, when the missing state information is a statement domain and the statement intent of the current dialogue statement is determined, the statement domains in the state information of a set of dialogue statements can be identified as a set of candidate statement domains. This set of candidate statement domains can be statement domains that have appeared in the state information of a set of dialogue statements, and the number of such candidate statement domains can be one or more.

[0056] Each statement domain can have multiple corresponding statement intents, and each statement intent can also correspond to multiple statement domains, but not every statement domain and statement intent can correspond to each other. For example, when the user's current dialogue statement is: "The temperature will increase by 2 degrees," the intent in the current dialogue statement is to increase the temperature. The missing domain in the current dialogue statement could be a device that can regulate temperature, such as an air conditioner or an oven, but it cannot be a smart device such as a television that is unrelated to the above intent.

[0057] In this embodiment, a correspondence between statement domains and statement intents can be pre-defined, for example, a device intent map. Based on the pre-defined correspondence between statement domains and statement intents, a candidate statement domain that matches the statement intent of the current dialogue statement can be determined from a set of candidate statement domains, and the target statement domain can be obtained. Here, the target statement domain can be the target state information in the aforementioned current dialogue statement.

[0058] For example, for a current query with a missing domain, the domain that matches the intent of the current query is determined by combining the content of the user's previous queries and the status information from multiple rounds, and the current query is then completed.

[0059] In this embodiment, by supplementing the missing domain information of the current dialogue statement based on the domains corresponding to the intent of the current dialogue statement in the previous rounds of state information, the accuracy of dialogue completion can be improved.

[0060] In one exemplary embodiment, determining the target state information that matches the missing state information from a set of dialogue statements' state information includes:

[0061] S21, if the missing state information is the statement intent and the statement domain of the current dialogue statement is determined, the statement intent in the state information of a set of dialogue statements is determined as a set of candidate statement intents.

[0062] S22, based on the preset correspondence between statement domains and statement intents, determine the candidate statement intent that matches the statement domain of the current dialogue statement from a set of candidate statement intents, and obtain the target statement intent, wherein the target state information is the target statement intent.

[0063] In this embodiment, when the missing state information is a statement intent and the statement domain of the current dialogue statement is determined, the statement intents in the state information of a set of dialogue statements can be identified as a set of candidate statement intents. Here, a set of candidate statement intents can be statement intents that have appeared in the state information of a set of dialogue statements, and the number of such intents can be one or more.

[0064] Based on the predefined correspondence between statement domains and statement intents, the candidate statement intent that matches the statement domain of the current dialogue statement can be determined from a set of candidate statement intents, and the target statement intent can be obtained. Here, the target statement intent is the target state information in the aforementioned current dialogue statement.

[0065] When the current dialogue statement lacks a statement intent and only has a statement domain, there may be multiple target statement intents among a set of candidate statement intents that match the statement domain of the current dialogue statement. Optionally, to determine the statement intent among multiple target statement intents that truly matches the target object's intent, the terminal device can interact with the target object for confirmation by sending a query statement to the target object via voice, and determining the supplementary target state information based on the target object's feedback. Optionally, the statement intent among multiple target statement intents that is most recent in time to the current statement can also be selected as the supplementary target state information for the current dialogue statement.

[0066] For example, for a current query lacking intent, by combining the content of the user's previous queries and multi-turn state information, the turn within the same domain as the current query can be determined. Based on the corresponding intent, the current query can be completed. For instance, if the user's current dialogue only contains the domain information "air conditioner," and the intent information corresponding to "air conditioner" determined from previous turns includes "raise by 2 degrees" and "turn on cooling," a question can be posed to the user, such as, "Do you need the air conditioner temperature raised by 2 degrees?" After receiving a positive response from the user, the current statement can be completed, and the operation of raising the air conditioner temperature can be executed.

[0067] In this embodiment, by supplementing the missing intent information of the current dialogue statement based on the intent corresponding to the domain of the current dialogue statement in the previous rounds of state information, the accuracy of dialogue completion can be improved.

[0068] In one exemplary embodiment, determining the target state information that matches the missing state information from a set of dialogue statements' state information includes:

[0069] S31, determine the state information that matches the missing state information in a set of dialogue statements, and obtain multiple candidate state information;

[0070] S32, among multiple candidate state information, the candidate state information with the shortest interval between the dialogue statement and the current dialogue statement is determined as the target state information.

[0071] In multi-turn dialogues, the correlation between the current dialogue statement and previous dialogues can be negatively correlated with the time interval; the earlier the turn, the weaker the correlation between the current dialogue statement and the previous turn, and vice versa. However, the previous turn of the turn containing the statement with the strongest correlation to the current dialogue statement may lack state information related to the current dialogue statement. Therefore, we can identify the state information in a set of dialogue statements that matches the missing state information, obtaining multiple candidate state information. This set of dialogue statements can include dialogue statements from at least one previous turn of dialogue. It can include dialogue statements from up to five turns of dialogue.

[0072] In this embodiment, determining the state information that matches the missing state information from a set of dialogue statements may yield multiple candidate state information. For these multiple candidate state information, the candidate state information whose dialogue statement has the shortest interval with the current dialogue statement can be selected as the target state information based on the round information of these candidate state information.

[0073] For example, when the state information in previous queries is not completely the same, there may be multiple instances where multiple content can be determined to be complete. When such ambiguity occurs, the content determined in the most recent query cycle takes precedence (the more recent the query cycle, the higher the weight).

[0074] In this embodiment, the candidate state information with the shortest interval between the current dialogue statement and the current dialogue statement is used as information to supplement the current dialogue statement, which can improve the accuracy of dialogue completion.

[0075] In one exemplary embodiment, target state information is used to complete the missing state information in the current dialogue statement to obtain the completed current dialogue statement, including:

[0076] S41, Perform syntactic analysis on the current dialogue statement to determine the missing syntactic relationship corresponding to the missing state information;

[0077] S42, Based on the missing syntactic relations, determine the position of the target statement in the current dialogue statement where the target state information is to be inserted;

[0078] S43, insert the target state information at the target statement position of the current dialogue statement to obtain the completed current dialogue statement.

[0079] Because the target state information for supplementing the current dialogue statement, determined based on the state information of the current dialogue statement and a set of dialogue statements, only involves state information and does not include the completion position within the current dialogue statement. The position of the missing state information within the current dialogue statement can be determined based on its syntactic structure.

[0080] In this embodiment, when it is determined that missing state information is present in the current dialogue statement, syntactic analysis can be performed on the current dialogue statement. Based on the grammatical structure information, the missing sentence components in the current dialogue statement can be identified. Combining the missing state information and the missing sentence components in the current dialogue statement, the missing syntactic relations corresponding to the missing state information can be determined. Based on the missing syntactic relations, the target statement position where the target state information should be inserted in the current dialogue statement can be determined. This syntactic analysis can be performed using a syntactic parser. The target statement position can be the location of the target state information within the current dialogue statement. Inserting the target state information at the target statement position in the current dialogue statement yields the completed current dialogue statement.

[0081] This embodiment determines the completion position of the current dialogue statement based on the grammatical structure information, which can avoid the situation where the logic is not logical after the state information is supplemented, leading to misunderstanding of the statement, thereby improving the efficiency of statement recognition in multi-turn dialogues.

[0082] In an exemplary embodiment, determining the target statement position where the target state information in the current dialogue statement should be inserted, based on missing syntactic relations, includes:

[0083] S51, In the case where the missing syntactic relation is a subject-predicate relation and there is a pronoun in the current dialogue statement, the position after the pronoun is determined as the target statement position in the current dialogue statement where the target state information is to be inserted.

[0084] S52, in the case where the syntactic relation is missing (subject-predicate relation) and there are no pronouns in the current dialogue statement, the beginning position of the current dialogue statement is determined as the target statement position where the target state information is to be inserted.

[0085] In this embodiment, the location for completing the target state information can be determined by combining common sentence components that need to be rewritten. The target sentence position where the target state information should be inserted can be determined based on the missing sentence components and the sentence component information of the current dialogue statement. Since the missing component information in a multi-turn dialogue is often a subject-predicate relationship, the position where the target state information should be inserted can be determined based on the presence or absence of pronouns in the current dialogue statement.

[0086] Optionally, if the missing syntactic relation is a subject-predicate relation and there is a pronoun in the current dialogue statement, the position after the pronoun can be determined as the target statement position where the target state information is to be inserted in the current dialogue statement; if the missing syntactic relation is a subject-predicate relation and there is no pronoun in the current dialogue statement, the beginning position of the current dialogue statement can be determined as the target statement position where the target state information is to be inserted in the current dialogue statement.

[0087] For example, the common rules for sentence composition that needs to be rewritten are as follows: When the sentence components in the current dialogue only have attributive-head and core relationships, or attributive-head, attributive-head, and core relationships, but no subject-predicate relationship, the domain or subject plus predicate from the previous round's state can be placed at the beginning of the sentence. If the current dialogue has pronouns, they should be placed after the pronouns. When the sentence components in the current dialogue only have core relationships and verb-object relationships, or core relationships, attributive-head, and verb-object relationships, but no subject-predicate relationship, the subject or domain from the previous round's dialogue can be placed at the beginning of the sentence. If the current dialogue has pronouns, they should be placed after the pronouns.

[0088] In this embodiment, when the syntactic relation is missing (subject-predicate relation), the system can fill in the missing pronoun target state information based on whether there is a position for the supplementation. This avoids the situation where the logic is not logical after the state information is supplemented, which leads to misunderstanding of the statement, thereby improving the efficiency of statement recognition in multi-turn dialogues.

[0089] In one exemplary embodiment, after obtaining the current dialogue statement, the above method further includes:

[0090] S61, after determining the statement domain and statement intent of the current dialogue statement by extracting the slots of the current dialogue statement, find the preset correspondence between the statement domain and statement intent according to the statement domain and statement intent of the current dialogue statement.

[0091] S62, if a correspondence is found that matches the statement domain and statement intent of the current dialogue statement, the target device operation corresponding to the current dialogue statement is executed.

[0092] In this embodiment, grammatical knowledge can be used to verify some common errors in syntactic parsing. Based on the obtained current dialogue statement, if it is determined that there are no missing statement domains or statement intentions, the statement domains and statement intentions of the current dialogue statement can be verified. If it is determined that the statement domains and statement intentions of the current dialogue statement match, the corresponding operation can be performed directly based on the understanding of the statement. The operation description here can be similar to that described above, and will not be repeated here.

[0093] In this embodiment, after determining the statement domain and statement intent of the current dialogue statement by extracting its slots, a preset correspondence between statement domains and statement intents can be found based on these correspondences. The found correspondence between the statement domains and statement intents of the current dialogue statement may or may not be a match.

[0094] If a correspondence is found that matches the statement domain and statement intent of the current dialogue statement, a target device operation corresponding to the current dialogue statement can be executed. This target device operation can be an operation performed by the terminal device, it can be controlling other devices to perform related operations (e.g., controlling an air conditioner to adjust the temperature), or it can be a specific response based on the user's question.

[0095] Optionally, if no correspondence is found between the statement domain and the statement intent of the current dialogue statement, the target can be asked via voice interaction through the terminal device whether the target needs to change the current dialogue statement.

[0096] Furthermore, the completed current dialogue statement can be validated using grammatical structure knowledge, including but not limited to checking the fluency of the complete statement. Confirmed correct completion of the current dialogue statement, along with corresponding multi-turn dialogue information, can be saved for experience accumulation. When encountering similar multi-turn dialogues, the saved results can be used directly for corresponding operations. Additionally, some common errors may occur during syntactic parsing; these can be verified by combining grammatical structure knowledge with the syntactic parsing results.

[0097] This embodiment uses the state information of the current dialogue statement to understand the statement and determine the semantic rationality of the current statement, thereby improving the accuracy of device control.

[0098] The following explanation, using optional examples, illustrates the method for processing dialogue statements in this application. In this optional example, the dialogue information is a query, and the current dialogue statement is the current query.

[0099] This optional example provides a solution for query completion in multi-turn dialogues. Based on the current query state information, combined with the syntactic parsing results, the content of the user's previous queries, and multi-turn state information, it achieves query completion for the user. For example... Figure 3 As shown, the process of processing dialogue statements in this optional example may include the following steps:

[0100] Step S302: The current query is parsed using a syntactic parser to analyze missing components.

[0101] Step S304: Analyze and extract the content of the user's previous queries (e.g., extract intent and domain) to obtain multi-round state information.

[0102] Step S306: Based on the content of the user's previous queries and the status information from multiple rounds, determine the supplementary content for the current query and complete the current query.

[0103] This optional example demonstrates how to use the sentence components mentioned above, as well as domain and intent recognition in multiple rounds of states, and grammatical knowledge, to achieve user query completion without relying on large amounts of data to train the model or manual annotation, thus saving time and costs associated with sentence completion.

[0104] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0105] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods according to the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM (Read-Only Memory) / RAM (Random Access Memory), magnetic disk, optical disk), and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods of the various embodiments of this application.

[0106] According to another aspect of the embodiments of this application, a dialog statement processing apparatus for implementing the above-described dialog statement processing method is also provided. Figure 4 This is a structural block diagram of an optional dialogue statement processing apparatus according to an embodiment of this application, such as... Figure 4 As shown, the device may include:

[0107] The acquisition unit 402 is used to acquire the current dialogue statement, wherein the current dialogue statement is the dialogue statement input by the target object in the current round of dialogue;

[0108] The first determining unit 404, connected to the obtaining unit 402, is used to determine the target state information that matches the missing state information in a set of state information of dialogue statements when the missing state information of the current dialogue statement is determined by slot extraction of the current dialogue statement. The state information includes at least one of the following: statement domain, statement intent, a set of dialogue statements contained in at least one round of dialogue before the current round of dialogue, and dialogue statements input by the target object.

[0109] The first execution unit 406 is connected to the first determination unit 404 and is used to complete the missing state information in the current dialogue statement using the target state information to obtain the completed current dialogue statement.

[0110] It should be noted that the acquisition unit 402 in this embodiment can be used to execute the above step S202, the first determination unit 404 in this embodiment can be used to execute the above step S204, and the first execution unit 406 in this embodiment can be used to execute the above step S206.

[0111] The above modules are used to obtain the current dialogue statement, which is the dialogue statement input by the target object in the current round of dialogue. After extracting slots from the current dialogue statement to determine the missing state information, target state information matching the missing state information is determined from a set of dialogue statement state information. This state information includes at least one of the following: statement domain, statement intent, a set of dialogue statements contained in at least one round of dialogue prior to the current round, and dialogue statements input by the target object. The target state information is used to complete the missing state information in the current dialogue statement, resulting in the completed current dialogue statement. This solves the problem of low accuracy in supplementing dialogue statements due to the diversity of dialogue content in related technologies, and improves the accuracy of the supplemented content.

[0112] In one exemplary embodiment, the first determining unit includes:

[0113] The first determining module is used to determine the statement domain in the state information of a set of dialogue statements as a set of candidate statement domains when the missing state information is a statement domain and the statement intent of the current dialogue statement is determined.

[0114] The second determining module is used to determine, based on the preset correspondence between statement domains and statement intents, a set of candidate statement domains that match the statement intent of the current dialogue statement, thereby obtaining the target statement domain, wherein the target state information is the target statement domain.

[0115] In one exemplary embodiment, the first determining unit includes:

[0116] The third determination module is used to determine the statement intent in the state information of a set of dialogue statements as a set of candidate statement intents when the missing state information is statement intent and the statement domain of the current dialogue statement is determined.

[0117] The fourth determination module is used to determine the candidate statement intent that matches the statement domain of the current dialogue statement from a set of candidate statement intents based on the preset correspondence between statement domains and statement intents, thereby obtaining the target statement intent, wherein the target state information is the target statement intent.

[0118] In one exemplary embodiment, the first determining unit includes:

[0119] The fifth determination module is used to determine the state information that matches the missing state information in a set of dialogue statements, and obtain multiple candidate state information.

[0120] The sixth determination module is used to determine the candidate state information that has the shortest interval between the dialogue statement and the current dialogue statement among multiple candidate state information as the target state information.

[0121] In one exemplary embodiment, the first execution unit includes:

[0122] The seventh determination module is used to perform syntactic analysis on the current dialogue statement and determine the missing syntactic relationship corresponding to the missing state information;

[0123] The eighth determination module is used to determine the position of the target statement in the current dialogue statement where the target state information is to be inserted, based on the missing syntactic relations.

[0124] The insertion module is used to insert target state information at the target statement position in the current dialogue statement to obtain the completed current dialogue statement.

[0125] In one exemplary embodiment, the eighth determining module includes:

[0126] The first determination submodule is used to determine the position after the pronoun as the target statement position where the target state information is to be inserted in the current dialogue statement when the missing syntactic relation is a subject-predicate relation and there is a pronoun in the current dialogue statement.

[0127] The second determination submodule is used to determine the position of the beginning of the current dialogue statement as the target statement position where the target state information is to be inserted, when there is a missing syntactic relation (subject-predicate relation) and no pronouns in the current dialogue statement.

[0128] In one exemplary embodiment, the above-described apparatus further includes:

[0129] The second determining unit is used to, after obtaining the current dialogue statement, and after determining the statement domain and statement intent of the current dialogue statement by extracting slots from the current dialogue statement, search for a preset correspondence between the statement domain and statement intent according to the statement domain and statement intent of the current dialogue statement.

[0130] The second execution unit is used to execute the target device operation corresponding to the current dialogue statement when a correspondence is found that matches the statement domain and statement intent of the current dialogue statement.

[0131] It should be noted that the examples and application scenarios implemented by the above modules and corresponding steps are the same, but are not limited to the content disclosed in the above embodiments. It should also be noted that the above modules, as part of a device, can operate in environments such as... Figure 1The hardware environment shown can be implemented through software or hardware, and the hardware environment includes the network environment.

[0132] According to another aspect of the embodiments of this application, a storage medium is also provided. Optionally, in this embodiment, the storage medium can be used to execute program code for processing any of the dialogue statements described above in the embodiments of this application.

[0133] Optionally, in this embodiment, the storage medium may be located on at least one of the network devices in the network shown in the above embodiment.

[0134] Optionally, in this embodiment, the storage medium is configured to store program code for performing the following steps:

[0135] S1, Get the current dialogue statement, where the current dialogue statement is the dialogue statement entered by the target object in the current round of dialogue;

[0136] S2, when the missing state information of the current dialogue statement is determined by slot extraction of the current dialogue statement, the target state information that matches the missing state information in a set of dialogue statements is determined, wherein the state information includes at least one of the following: statement domain, statement intent, a set of dialogue statements contained in at least one round of dialogue before the current round of dialogue, and dialogue statements input by the target object.

[0137] S3, use the target state information to complete the missing state information in the current dialogue statement, and obtain the completed current dialogue statement.

[0138] Optionally, specific examples in this embodiment can refer to the examples described in the above embodiments, and will not be repeated in this embodiment.

[0139] Optionally, in this embodiment, the storage medium may include, but is not limited to, various media capable of storing program code, such as USB flash drives, ROMs, RAMs, portable hard drives, magnetic disks, or optical disks.

[0140] According to another aspect of the embodiments of this application, an electronic device for implementing the above-described method for processing dialogue statements is also provided. The electronic device may be a server, a terminal, or a combination thereof.

[0141] Figure 5 This is a structural block diagram of an optional electronic device according to an embodiment of this application, such as... Figure 5 As shown, it includes a processor 502, a communication interface 504, a memory 506, and a communication bus 508. The processor 502, communication interface 504, and memory 506 communicate with each other via the communication bus 508.

[0142] Memory 506 is used to store computer programs;

[0143] When processor 502 executes a computer program stored in memory 506, it performs the following steps:

[0144] S1, Get the current dialogue statement, where the current dialogue statement is the dialogue statement entered by the target object in the current round of dialogue;

[0145] S2, when the missing state information of the current dialogue statement is determined by slot extraction of the current dialogue statement, the target state information that matches the missing state information in a set of dialogue statements is determined, wherein the state information includes at least one of the following: statement domain, statement intent, a set of dialogue statements contained in at least one round of dialogue before the current round of dialogue, and dialogue statements input by the target object.

[0146] S3, use the target state information to complete the missing state information in the current dialogue statement, and obtain the completed current dialogue statement.

[0147] Optionally, the communication bus can be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus, etc. This communication bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 5 The symbol is represented by a single thick line, but this does not indicate that there is only one bus or one type of bus. The communication interface is used for communication between the aforementioned electronic device and other devices.

[0148] The memory may include RAM, or non-volatile memory, such as at least one disk storage device. Optionally, the memory may also be at least one storage device located remotely from the aforementioned processor.

[0149] As an example, the memory 506 described above may include, but is not limited to, the acquisition unit 402, the first determination unit 404, and the first execution unit 406 in the dialogue statement processing device described above. Furthermore, it may include, but is not limited to, other module units in the dialogue statement processing device described above, which will not be elaborated upon in this example.

[0150] The processors mentioned above can be general-purpose processors, including but not limited to: CPU (Central Processing Unit), NP (Network Processor), etc.; they can also be DSP (Digital Signal Processor), ASIC (Application Specific Integrated Circuit), FPGA (Field-Programmable Gate Array), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0151] Optionally, specific examples in this embodiment can refer to the examples described in the above embodiments, and will not be repeated here.

[0152] Those skilled in the art will understand that Figure 5 The structure shown is for illustrative purposes only. The device that implements the above-described dialogue processing method can be a terminal device, such as a smartphone (e.g., an Android phone, an iOS phone), a tablet computer, a PDA, a mobile internet device (MID), a PAD, or other terminal devices. Figure 5 This does not limit the structure of the aforementioned electronic device. For example, the electronic device may also include components that are more... Figure 5 The more or fewer components shown (such as network interfaces, display devices, etc.), or having the same Figure 5 The different configurations shown.

[0153] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing the hardware related to the terminal device. The program can be stored in a computer-readable storage medium, which may include: flash drive, ROM, RAM, disk or optical disk, etc.

[0154] The sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0155] If the integrated units in the above embodiments are implemented as software functional units and sold or used as independent products, they can be stored in the aforementioned computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause one or more computer devices (which may be personal computers, servers, or network devices, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application.

[0156] In the above embodiments of this application, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0157] In the several embodiments provided in this application, it should be understood that the disclosed client can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces, indirect coupling or communication connection between units or modules, and may be electrical or other forms.

[0158] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of the solution provided in this embodiment, depending on actual needs.

[0159] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or at least two units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0160] The above description is only a preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A method for processing dialogue statements, characterized in that, include: Obtain the current dialogue statement, wherein the current dialogue statement is the dialogue statement entered by the target object in the current round of dialogue; When the missing state information of the current dialogue statement is determined by slot extraction of the current dialogue statement, the target state information that matches the missing state information in a set of dialogue statements is determined, wherein the state information includes at least one of the following: statement domain, statement intent, the set of dialogue statements being included in at least one round of dialogue before the current round of dialogue, and the dialogue statement input by the target object. The step of determining the target state information that matches the missing state information in the state information of a set of dialogue statements includes: when the missing state information is a statement intent and the statement domain of the current dialogue statement is determined, the statement intent in the state information of the set of dialogue statements is determined as a set of candidate statement intents; there are multiple target statement intents in the set of candidate statement intents that match the statement domain of the current dialogue statement; interactive confirmation is performed between the terminal device and the target object, an inquiry statement is sent to the target object by voice, and supplementary target state information is determined based on the target object's feedback; when the missing state information is a statement domain and the statement intent of the current dialogue statement is determined, the statement domain in the state information of the set of dialogue statements is determined as a set of candidate statement domains; according to a preset correspondence between statement domains and statement intents, the candidate statement domain that matches the statement intent of the current dialogue statement in the set of candidate statement domains is determined to obtain the target statement domain, wherein the target state information is the target statement domain; the correspondence between the statement domain and the statement intent is a device intent map; Based on the obtained current dialogue statement, if it is determined that there are no missing statement domains and statement intentions, the statement domains and statement intentions of the current dialogue statement are verified; if a correspondence matching the statement domains and statement intentions of the current dialogue statement is found, the target device operation corresponding to the current dialogue statement is executed; if no correspondence matching the statement domains and statement intentions of the current dialogue statement is found, voice interaction is conducted between the terminal device and the target object to inquire whether the target object needs to modify the current dialogue statement. The target state information is used to complete the missing state information in the current dialogue statement, resulting in the completed current dialogue statement.

2. The method according to claim 1, characterized in that, When the missing state information is a statement intent and the statement domain of the current dialogue statement is determined, after determining the statement intent in the state information of the group of dialogue statements as a group of candidate statement intents, the method further includes: Based on the preset correspondence between statement domains and statement intents, the candidate statement intent that matches the statement domain of the current dialogue statement in the set of candidate statement intents is determined to obtain the target statement intent, wherein the target state information is the target statement intent.

3. The method according to any one of claims 1 to 2, characterized in that, The target state information that matches the missing state information in the state information of a set of dialogue statements includes: Determine the state information that matches the missing state information in the state information of the set of dialogue statements to obtain multiple candidate state information; The candidate state information with the shortest interval between the dialogue statement and the current dialogue statement among the multiple candidate state information is determined as the target state information.

4. The method according to claim 3, characterized in that, The step of using the target state information to complete the missing state information in the current dialogue statement to obtain the completed current dialogue statement includes: Perform syntactic analysis on the current dialogue statement to determine the missing syntactic relationship with the missing state information; Based on the missing syntactic relations, determine the position of the target statement in the current dialogue statement where the target state information is to be inserted; The target state information is inserted at the target statement position in the current dialogue statement to obtain the completed current dialogue statement.

5. The method according to claim 4, characterized in that, The step of determining the target statement position in the current dialogue statement where the target state information is to be inserted, based on the missing syntactic relation, includes: In the case where the missing syntactic relation is a subject-predicate relation and there is a pronoun in the current dialogue statement, the position after the pronoun is determined as the position in the current dialogue statement where the target state information is to be inserted. In the case where the missing syntactic relation is a subject-predicate relation and there are no pronouns in the current dialogue statement, the beginning position of the current dialogue statement is determined as the position in the target statement where the target state information is to be inserted.

6. The method according to claim 1, characterized in that, After obtaining the current dialogue statement, the method further includes: By extracting slots from the current dialogue statement and determining its statement domain and statement intent, a preset correspondence between statement domains and statement intents is found according to the statement domain and statement intent of the current dialogue statement.

7. A device for processing dialogue statements, characterized in that, include: The acquisition unit is used to acquire the current dialogue statement, wherein the current dialogue statement is the dialogue statement input by the target object in the current round of dialogue; The first determining unit is configured to, when determining the missing state information of a set of dialogue statements by slot extraction of the current dialogue statement, determine target state information that matches the missing state information from the state information of the current dialogue statement, wherein the state information includes at least one of the following: statement domain, statement intent, and the set of dialogue statements is included in at least one round of dialogue before the current round of dialogue, and is input by the target object; when the missing state information is a statement intent and the statement domain of the current dialogue statement is determined, the statement intent in the state information of the set of dialogue statements is determined as a set of candidate statement intents; there are multiple target statement intents in the set of candidate statement intents that match the statement domain of the current dialogue statement; interactive confirmation is performed between the terminal device and the target object, an inquiry statement is sent to the target object by voice, and supplementary target state information is determined based on the feedback of the target object; when the missing state information is a statement domain and the statement intent of the current dialogue statement is determined, the set of dialogue statements is determined as a set of target state information that matches the missing state information from the current dialogue statement, and is input by the target object by slot extraction of ... The statement domain in the state information of the statement is determined as a set of candidate statement domains; according to the preset correspondence between statement domains and statement intents, the candidate statement domain that matches the statement intent of the current dialogue statement is determined from the set of candidate statement domains to obtain the target statement domain, wherein the target state information is the target statement domain; the correspondence between the statement domain and the statement intent is a device intent map; based on the obtained current dialogue statement, if it is determined that there are no missing statement domains and statement intents, the statement domains and statement intents of the current dialogue statement are verified; if a correspondence matching the statement domains and statement intents of the current dialogue statement is found, the target device operation corresponding to the current dialogue statement is executed; if no correspondence matching the statement domains and statement intents of the current dialogue statement is found, voice interaction is performed between the terminal device and the target object to ask the target object whether the current dialogue statement needs to be changed; The first execution unit is used to use the target state information to complete the missing state information in the current dialogue statement, so as to obtain the completed current dialogue statement.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored program, wherein the program, when executed, performs the method of any one of claims 1 to 6.

9. An electronic device comprising a memory and a processor, characterized in that, The memory stores a computer program, and the processor is configured to execute the method of any one of claims 1 to 6 through the computer program.

Images