Dialogue interaction method, device and electronic equipment

Abstract

This application provides a dialogue interaction method, apparatus, and electronic device, relating to the field of artificial intelligence technology. The method includes: responding to a dialogue request from a target user, using a knowledge graph to perform semantic recognition on the dialogue data carried in the dialogue request to obtain an initial intent, whereby the initial intent represents a preliminarily matched government service item. The knowledge graph includes entities and edge relationships, where entities represent government service elements, and edge relationships represent the association attribute relationships between entities; detecting the entity slots associated with the initial intent, and triggering an interactive information completion process for the missing entity slots if any are missing; and, in response to the completion of information completion for the missing entity slots, sending a government service path corresponding to the initial intent to the target user, whereby the government service path carries government service element information corresponding to the entity slots.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence technology, and more specifically, to a dialogue interaction method, apparatus, and electronic device. Background Technology

[0002] In government service Q&A systems, users typically use natural language dialogue to inquire about procedures, submit applications, or check processing progress. The system identifies the initial service request based on the user's dialogue data and then pushes relevant processing entry points and service guides.

[0003] In the process of realizing the concept of this application, the inventors discovered that the relevant technology has at least the following problems: the government affairs question and answer system relies on keyword matching to identify dialogue intent, and the accuracy of dialogue intent identification for colloquial expressions is low. Summary of the Invention

[0004] In view of this, this application provides a dialogue interaction method, apparatus, and electronic device.

[0005] One aspect of this application provides a dialogue interaction method that, in response to a dialogue request from a target user, utilizes a knowledge graph to perform semantic recognition on the dialogue data carried in the dialogue request to obtain an initial intent. The initial intent represents a preliminarily matched government service item. The knowledge graph includes entities and edge relationships, where entities represent government service elements and edge relationships represent the association attribute relationships between entities. The method detects the entity slots associated with the initial intent, and if a missing entity slot exists, triggers an interactive information completion process for the missing entity slot. In response to the completion of information completion for the missing entity slot, a government service path corresponding to the initial intent is sent to the target user. The government service path carries government service element information corresponding to the entity slot.

[0006] According to an embodiment of this application, the knowledge graph is constructed based on the following operations: extracting entities from government structured data, whereby entities include at least one of government service items, processing levels, compliance prerequisites, and processing locations; constructing edge relationships based on the jurisdictional relationship between government service items and processing locations, the dependency relationship between government service items and compliance prerequisites, and the affiliation relationship between government service items and processing levels; and constructing the knowledge graph based on the entities and edge relationships.

[0007] According to embodiments of this application, the dialogue interaction method further includes: responding to the target user's supplementary request for a desired processing location, retrieving valid government service items with jurisdictional relationship to the desired processing location from a knowledge graph; and, if the initial intent does not match the valid government service items, sending a first prompt message or a first adjustment of the government service path to the target user, wherein the first adjustment of the government service path represents the processing path associated with government service items with jurisdictional relationship to the desired processing location and semantically matching the initial intent, and the first prompt message represents the result of mismatch in jurisdictional authority.

[0008] According to embodiments of this application, the dialogue interaction method further includes: responding to the desired processing level supplemented by the target user, retrieving valid government service items with an affiliation relationship to the desired processing level in the knowledge graph; and, in the case where the initial intent does not match the valid government service items, sending a second prompt message or a second adjustment of the government service path to the target user, wherein the second adjustment of the government service path represents the processing path associated with the government service items that have an affiliation relationship to the desired processing level and whose semantics match the initial intent, and the second prompt message represents the result of the processing level permission mismatch.

[0009] According to embodiments of this application, the compliance prerequisites include qualification certification information. The dialogue interaction method further includes: responding to the qualification certification information supplemented by the target user, retrieving valid government service items with an attribution relationship to the qualification certification information in the knowledge graph; and, in the case that the initial intent does not match the valid government service items, sending a third prompt message or a third adjustment of the government service path to the target user. The third adjustment of the government service path represents the processing path of the qualification certification information that is dependent on the initial intent, and the third prompt message represents the result of the compliance prerequisites not matching.

[0010] According to an embodiment of this application, the entity slots associated with the initial intent are detected. If there are missing entity slots, an interactive information completion process for the missing entity slots is triggered, including: determining a preset set of entity slots associated with the initial intent in the knowledge graph; detecting the filling status of each of the multiple entity slots in the preset entity slot set according to the dialogue data; and triggering an interactive information completion process for the missing entity slots if there are missing entity slots with an empty filling status.

[0011] According to an embodiment of this application, triggering an interactive information completion process for a missing entity slot includes: determining a guiding query template associated with the missing entity slot from a preset template library; sending the guiding query template to a target user; and in response to the target user's completion data for the guiding query template, extracting the missing entity information associated with the missing entity slot from the completion data and filling the missing entity information into the missing entity slot.

[0012] According to embodiments of this application, semantic recognition of dialogue data carried in a dialogue request using a knowledge graph to obtain an initial intent includes: mapping the dialogue data and entities representing government service items in the knowledge graph to a target semantic embedding space to obtain a first semantic vector and a second semantic vector, wherein the first semantic vector represents the vector representation of the dialogue data in the target semantic embedding space, and the second semantic vector represents the vector representation of the government service item in the target semantic embedding space; performing semantic matching on the first semantic vector and the second semantic vector to obtain a semantic similarity; and determining entities with a semantic similarity greater than a preset threshold as the initial intent.

[0013] Another aspect of this application provides a dialogue interaction device, comprising: an identification module, configured to, in response to a dialogue request from a target user, perform semantic recognition on the dialogue data carried in the dialogue request using a knowledge graph to obtain an initial intent, wherein the initial intent represents a preliminarily matched government service item, and the knowledge graph includes entities and edge relationships, wherein entities represent government service elements, and edge relationships represent the association attribute relationships between entities; a detection module, configured to detect entity slots associated with the initial intent, and, in the case of missing entity slots, trigger an interactive information completion process for the missing entity slots; and a sending module, configured to, in response to the completion of information completion for the missing entity slots, send a government service path corresponding to the initial intent to the target user, wherein the government service path carries government service element information corresponding to the entity slots.

[0014] Another aspect of this application provides an electronic device comprising:

[0015] One or more processors;

[0016] Memory, used to store one or more programs.

[0017] Specifically, when one or more programs are executed by one or more processors, the one or more processors implement the above method.

[0018] Another aspect of this application provides a computer-readable storage medium storing computer-executable instructions that, when executed, are used to implement the method described above.

[0019] Another aspect of this application provides a computer program product including computer-executable instructions that, when executed, are used to implement the methods described above.

[0020] According to the embodiments of this application, a semantic network knowledge graph with integrated permission constraints is constructed based on the association attribute relationships between government service elements. Then, the entities of government service items in the knowledge graph and the dialogue data are matched, which reduces the semantic difference between colloquial expressions and government standard terminology, and achieves accurate identification of the initial intent. At the same time, the intent identification process has the ability to verify business norms and permissions. By utilizing the topological structured features and node adjacency relationships of the knowledge graph, entity slots are dynamically detected in real time during multi-round dialogue interactions. The filling status of multiple entity slots associated with the government service item entity corresponding to the initial intent is traversed. When a missing entity slot is detected, an interactive information completion process for the missing entity slot is triggered, thereby accurately capturing the processing scenario elements hidden in the dialogue data. This achieves a closed-loop fusion from intent identification to government service element completeness verification, and solves the technical problem of ineffective guidance caused by the lack of key information in multi-round dialogues. Attached Figure Description

[0021] The above and other objects, features and advantages of this application will become clearer from the following description of embodiments with reference to the accompanying drawings, in which:

[0022] Figure 1 The illustration shows application scenario diagrams of the dialogue interaction method, apparatus, and electronic device according to embodiments of this application;

[0023] Figure 2 A flowchart of a dialogue interaction method according to an embodiment of this application is shown;

[0024] Figure 3 A schematic diagram illustrating the construction of a knowledge graph according to an embodiment of this application is shown;

[0025] Figure 4 A block diagram of a dialogue interaction device according to an embodiment of this application is shown; and

[0026] Figure 5 A block diagram of an electronic device suitable for implementing a dialogue interaction method according to an embodiment of this application is shown. Detailed Implementation

[0027] The embodiments of this application will now be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of this application. In the following detailed description, numerous specific details are set forth to provide a thorough understanding of the embodiments of this application for ease of explanation. However, it will be apparent that one or more embodiments may be implemented without these specific details. Furthermore, descriptions of well-known structures and technologies are omitted in the following description to avoid unnecessarily obscuring the concepts of this application.

[0028] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of this application. The terms “comprising,” “including,” etc., as used herein indicate the presence of features, steps, operations, and / or components, but do not exclude the presence or addition of one or more other features, steps, operations, or components.

[0029] All terms used herein (including technical and scientific terms) have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein are to be interpreted in a manner consistent with the context of this specification, and not in an idealized or overly rigid way.

[0030] When using expressions such as "at least one of A, B and C", they should generally be interpreted in accordance with the meaning that is commonly understood by those skilled in the art (e.g., "a system having at least one of A, B and C" should include, but is not limited to, a system having A alone, a system having B alone, a system having C alone, a system having A and B, a system having A and C, a system having B and C, and / or a system having A, B and C, etc.).

[0031] In the embodiments of this application, the collection, updating, analysis, processing, use, transmission, provision, disclosure, and storage of data (e.g., including but not limited to user personal information) comply with relevant laws and regulations, are used for legitimate purposes, and do not violate public order and good morals. In particular, necessary measures have been taken to prevent unauthorized access to user personal information data and to safeguard user personal information security and network security.

[0032] In the embodiments of this application, the user's authorization or consent was obtained before obtaining or collecting the user's personal information.

[0033] The following problems exist in the dialogue interaction process of the existing system:

[0034] First, semantic differences lead to matching errors. While general models possess strong generalization capabilities, they lack a deep understanding of government service elements such as "standard terminology for government affairs," "processing levels," and "local authority." Significant semantic differences exist between colloquial user input and standard government terminology, resulting in low accuracy in intent recognition.

[0035] Second, there is a lack of permission filtering mechanisms. The relevant technologies cannot dynamically perceive the user's desired processing location, subject type, etc., during multi-turn dialogues, easily leading users to process matters that cannot be handled in their region or that do not conform to their subject type, resulting in ineffective guidance and user confusion.

[0036] Thirdly, there is the issue of intent drift in multi-turn dialogues. In complex multi-turn interactions, the relevant technologies struggle to dynamically supplement the user's missing key information, causing the system's understanding of the user's intent to drift as the number of dialogue rounds increases, ultimately leading to incorrect answers or service guidance.

[0037] In view of this, embodiments of this application provide a dialogue interaction method, including: responding to a dialogue request from a target user, using a knowledge graph to perform semantic recognition on the dialogue data carried in the dialogue request to obtain an initial intent, the initial intent representing a preliminarily matched government service item, the knowledge graph including entities and edge relationships, entities representing government service elements, and edge relationships representing the association attribute relationships between entities; detecting the entity slots associated with the initial intent, and triggering an interactive information completion process for the missing entity slots if there are missing entity slots; responding to the completion of information completion for the missing entity slots, sending a government service path corresponding to the initial intent to the target user, the government service path carrying government service element information corresponding to the entity slots.

[0038] Figure 1 The diagram illustrates application scenarios of the dialogue interaction method, apparatus, and electronic device according to embodiments of this application.

[0039] like Figure 1 As shown, application scenario 100 according to an embodiment of this application may include a first terminal device 101, a second terminal device 102, a third terminal device 103, a network 104, and a server 105. The network 104 serves as a medium for providing a communication link between the first terminal device 101, the second terminal device 102, the third terminal device 103, and the server 105. The network 104 may include various connection types, such as wired or wireless communication links or fiber optic cables. For example, a user can use the first terminal device 101, the second terminal device 102, and the third terminal device 103 to interact with the server 105 through the network 104 to receive or send information, etc.

[0040] The first terminal device 101, the second terminal device 102, and the third terminal device 103 can be electronic devices such as smartphones, wearable devices, personal computers, intelligent voice interaction devices, smart home appliances, intelligent vehicles, in-vehicle terminals, aircraft, unmanned vending terminals, and extended reality devices. Extended reality devices can include virtual reality devices, augmented reality devices, and mixed reality devices. A client application for the target application can be installed and run on the terminal devices. This target application can include, but is not limited to, financial transaction applications, payment applications, shopping applications, web browser applications, search applications, instant messaging tools, email clients, and social media platform software (these are just examples). Furthermore, this application embodiment does not limit the form of the target application, and it can include, but is not limited to, applications, mini-programs, etc., installed on the terminal devices, and can also be in the form of web pages.

[0041] Server 105 can be a server providing various services, such as a backend management server supporting websites browsed by users using the first terminal device 101, the second terminal device 102, and the third terminal device 103 (this is just an example). The backend management server can analyze and process received user requests and other data, and feed back the processing results (such as web pages, information, or data obtained or generated according to user requests) to the terminal devices. The server can be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing cloud computing services such as cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks, and basic cloud computing services such as big data. The server can be the backend server of the aforementioned target application, used to provide backend services to the clients of the target application.

[0042] It should be noted that the dialogue interaction method provided in the embodiments of this application can generally be executed by server 105 and / or terminal devices 101-103. Accordingly, the dialogue interaction device provided in the embodiments of this application can generally be set in server 105 and / or terminal devices 101-103.

[0043] It should be understood that Figure 1 The number of terminal devices, networks, and servers shown is merely illustrative. Depending on implementation needs, any number of terminal devices, networks, and servers can be included.

[0044] It should be noted that the dialogue interaction method provided in the embodiments of this application can generally be executed by server 105 and / or terminal devices 101-103. Accordingly, the dialogue interaction device provided in the embodiments of this application can generally be set in server 105 and / or terminal devices 101-103.

[0045] It should be understood that Figure 1 The number of terminal devices, networks, and servers shown is merely illustrative. Depending on implementation needs, any number of terminal devices, networks, and servers can be included.

[0046] Figure 2 A flowchart of a dialogue interaction method according to an embodiment of this application is shown.

[0047] like Figure 2 As shown, the dialogue interaction method includes operations S210~S230.

[0048] In operation S210, in response to the dialogue request from the target user, the knowledge graph is used to perform semantic recognition on the dialogue data carried in the dialogue request to obtain the initial intent.

[0049] In operation S220, the entity slots associated with the initial intent are detected. If there are missing entity slots, an interactive information completion process for the missing entity slots is triggered.

[0050] In operation S230, in response to the completion of information completion for missing entity slots, a government service path corresponding to the initial intent is sent to the target user.

[0051] According to an embodiment of this application, the dialogue interaction method can be applied to a government affairs question-and-answer system, where the target user can be the object currently interacting with the government affairs question-and-answer system.

[0052] A dialogue request represents the first dialogue interaction initiated by a target user with the government affairs Q&A system. The dialogue data carried in the dialogue request includes natural language text or voice data.

[0053] Extract entities related to government service elements and the relationships between entities from structured data in the government sector. For example, structured data may include laws and regulations, service guides, etc., and construct a knowledge graph based on the entities and the relationships between them.

[0054] A knowledge graph consists of entities and edge relationships. Entities are nodes in the knowledge graph, and they represent government service elements. For example, government service elements can include standardized government service items, processing departments, processing locations, applicable entities, and other service elements.

[0055] Edges in a knowledge graph are used to connect entities, and edge relationships represent the association attributes between entities. For example, association attribute relationships can be the affiliation relationship between government service items and the handling departments, or the jurisdiction relationship between government service items and the handling location.

[0056] In response to the government affairs question-and-answer system receiving a dialogue request from a target user, the system extracts the dialogue data carried in the dialogue request and calls a pre-built knowledge graph for semantic recognition. By calculating the semantic similarity between the dialogue data and the government affairs service entities stored in the knowledge graph, the system identifies the government affairs service that the target user is most likely to want to handle, thus obtaining the initial intent.

[0057] The initial intent represents the government service item that is initially matched. For example, the initial intent is "registration of the establishment of individual businesses".

[0058] Entity slots are variables used to populate entity information in the government affairs Q&A system. For example, in "Handling [Item] in [City]", both "Item" and "City" are entity slots. The "Item" entity slot corresponds to the government service item node in the knowledge graph, which is the government service item that the target user expects to handle. The "City" entity slot corresponds to the handling location node in the knowledge graph, which is the administrative region that the target user expects to handle.

[0059] Based on the information extracted from the dialogue data, the entity slots associated with the initial intent are detected. If an entity slot that has not been filled by the target user is detected, an interactive information completion process for the missing entity slot is triggered.

[0060] The interactive information completion process involves the government Q&A system guiding users to fill in the missing information for entity slots by sending inquiries to them and receiving their responses.

[0061] For example, if the missing entity slot is the location of the application, an inquiry template is sent to the target user: "Which city did you apply in?" In response to the target user's reply: "xx city", the location information "xx city" is filled into the missing entity slot.

[0062] The interactive information completion process includes multiple rounds of question-and-answer interaction, responding to the target user's completion of all missing entity slots associated with the initial intent, and automatically triggering the generation and push of government service paths.

[0063] The entity slot information extracted from dialogue data can be used to verify permissions based on the entity information associated with the initial intent stored in the knowledge graph. A standardized service process can be generated based on the entity slot information that passes the verification. The government service path can be presented in the form of a service guide based on the standardized service process, carrying the government service element information corresponding to all entity slots associated with the target user's initial intent.

[0064] According to the embodiments of this application, a knowledge graph of the government domain, which integrates permission constraints, is constructed based on the associated attribute relationships between government service elements. Then, the entities of government service items in the knowledge graph and the dialogue data are matched, which reduces the semantic difference between colloquial expressions and government standard terminology, and achieves accurate identification of the initial intent. At the same time, the intent identification process has the ability to verify business norms and permissions. By utilizing the topological structured features and node adjacency relationships of the knowledge graph, entity slots are dynamically detected in real time during multi-round dialogue interactions. The filling status of multiple entity slots associated with the government service item entity corresponding to the initial intent is traversed. When a missing entity slot is detected, an interactive information completion process for the missing entity slot is triggered, thereby accurately capturing the processing scenario elements hidden in the dialogue data. This achieves a closed-loop fusion from intent identification to government service element completeness verification, and solves the technical problem of invalid guidance caused by the lack of key information in multi-round dialogues.

[0065] According to an embodiment of this application, the knowledge graph is constructed based on the following operations: extracting entities from government structured data, whereby entities include at least one of government service items, processing levels, compliance prerequisites, and processing locations; constructing edge relationships based on the jurisdictional relationship between government service items and processing locations, the dependency relationship between government service items and compliance prerequisites, and the affiliation relationship between government service items and processing levels; and constructing the knowledge graph based on the entities and edge relationships.

[0066] Entities are extracted from structured data in the government sector. These entities include government service items, processing levels, applicable entities, compliance prerequisites, and processing locations.

[0067] The term "government service item" refers to the standard name of the business to be handled. The "handling level" refers to the level of the administrative agency responsible for approving the government service item. The "compliance prerequisites" refer to the qualifications or material requirements that need to be met in advance for handling the government service item. The "handling location" indicates the administrative region where the item is handled. The applicable subjects are individuals or legal persons.

[0068] For example, from the structured data of the government service network, “registration of establishment of individual business households” is extracted as the entity of government service items, “district and county level” as the entity of processing level, “proof of use of business premises” as the entity of compliance prerequisite, “XX city” as the entity of processing location, and “individual” as the entity of applicable subject.

[0069] Border relations include jurisdictional relations, dependency relations, and affiliation relations.

[0070] Jurisdiction is used to connect the entity handling the government service matter with the entity in the jurisdiction where it is handled, indicating whether the entity has the authority to handle the government service matter within the jurisdiction where it is handled.

[0071] Dependency relationships are used to connect entities related to government service items with entities related to compliance prerequisites, representing the legal and regulatory preconditions that need to be met to process government service items.

[0072] The attribution relationship is used to connect the entity of the government service item with the entity at the processing level, indicating the level of the administrative agency that manages the government service item.

[0073] A knowledge graph is constructed based on the extracted entities and edge relationships. The knowledge graph stores knowledge of government service elements in a graph structure, with entities as nodes and edge relationships as directed edges connecting the nodes, forming a knowledge network that can be understood and reasoned about by computers.

[0074] According to the embodiments of this application, by embedding permission information such as processing location, processing level, and compliance prerequisites as edge relationships into the knowledge graph structure, the intent recognition process can complete business permission verification simultaneously with semantic matching, reducing the computational overhead of the system, improving the accuracy and efficiency of intent recognition, and avoiding invalid intent output.

[0075] Figure 3 A schematic diagram illustrating the construction of a knowledge graph according to an embodiment of this application is shown.

[0076] like Figure 3 As shown, a data processing pipeline model is used to preprocess structured data in the government affairs field, including cleaning, deduplication, filling in missing values, correcting erroneous values, removing outliers, standardizing fields, formats, units, and normalization. A large-scale artificial intelligence model combined with a graph retrieval-augmented generation (GraphRAG) algorithm is used to extract entities and edge relationships from the preprocessed data, thereby constructing a knowledge graph. Entities in the knowledge graph include government service items, processing levels, processing locations, and compliance prerequisites, and the entity and entity relationship data are stored in a database. The knowledge graph is used to perform semantic recognition on the dialogue data carried in the dialogue request to obtain the initial intent; the entity slots associated with the initial intent are detected, and if a missing entity slot exists, an interactive information completion process for the missing entity slot is triggered.

[0077] According to embodiments of this application, the dialogue interaction method further includes: responding to the target user's supplementary request for a desired processing location, retrieving valid government service items with jurisdictional relationship to the desired processing location from a knowledge graph; and, if the initial intent does not match the valid government service items, sending a first prompt message or a first adjustment of the government service path to the target user, wherein the first adjustment of the government service path represents the processing path associated with government service items with jurisdictional relationship to the desired processing location and semantically matching the initial intent, and the first prompt message represents the result of mismatch in jurisdictional authority.

[0078] The desired processing location is the administrative region that the target user actively chooses to process the matter, which is generally determined based on the target user's place of household registration, place of residence, or place of business.

[0079] Jurisdiction refers to the mapping of legal administrative authority between government service items and the jurisdiction where they are handled. For example, the "registration of the establishment of individual businesses" can only be under the jurisdiction of the "District-level Market Supervision Bureau".

[0080] A valid government service item is a set of nodes in the knowledge graph that represent government service items and have a jurisdictional relationship with the node representing the expected location.

[0081] For example, if a target user's initial intent is to "apply for a passport," and they wish to apply in a location defined as "City X1, District X2," the knowledge graph is searched for valid government services related to the desired location. If "apply for a passport" is among the valid government services, it means there is a jurisdictional relationship between "applying for a passport" and "City X1, District X2," resulting in a successful match. The target user is then sent an application portal and application guide. Otherwise, the match is considered unsuccessful, and the target user is sent an initial notification or an option to adjust the government service path.

[0082] The first prompt message is an interactive message from the government affairs Q&A system to the target user, used to inform the target user that the selected desired jurisdiction does not have the authority to handle the initial intent. For example, the first prompt message is "Jurisdiction authority mismatch result".

[0083] The first adjustment of the government service path represents the alternative processing path automatically recommended by the system when the initial intention does not belong to the set of valid government service items but there are valid government service items with semantic matching.

[0084] In one embodiment, the government service item associated with the processing path semantically matches the initial intent and has a jurisdictional relationship with the desired processing location; in another embodiment, the processing path is an alternative processing location with a jurisdictional relationship with the initial intent, retrieved based on a knowledge graph.

[0085] For example, if the valid government service item that matches the initial intent semantics is "passport renewal", the target customer is sent the first adjusted government service path: "We have detected that you may need to renew your passport. This item can be done in "City X1, District X2". Do you want to redirect?" or "We have detected that you may need to apply for a passport. This item can be done in "City X3". Do you want to redirect?" At the same time, relevant application guidelines are sent.

[0086] According to the embodiments of this application, by verifying the semantic association and jurisdiction of the knowledge graph, when the user selects a jurisdiction that does not have the corresponding authority, the system can provide real-time feedback on the authority mismatch result, avoiding the target user from making ineffective transfers on the wrong path and reducing the cost of trial and error in handling affairs.

[0087] According to embodiments of this application, the dialogue interaction method further includes: responding to the desired processing level supplemented by the target user, retrieving valid government service items with an affiliation relationship to the desired processing level in the knowledge graph; and, in the case where the initial intent does not match the valid government service items, sending a second prompt message or a second adjustment of the government service path to the target user, wherein the second adjustment of the government service path represents the processing path associated with the government service items that have an affiliation relationship to the desired processing level and whose semantics match the initial intent, and the second prompt message represents the result of the processing level permission mismatch.

[0088] The desired processing level is the administrative level at which the target user actively selects the matter to be processed.

[0089] The attribution relationship refers to the legally mandated hierarchical authority mapping between government service items and the processing levels. For example, the "environmental project approval" item can only be handled by the "provincial XX department" level.

[0090] A valid government service item that has an affiliation relationship with the desired processing level is the set of nodes representing government service items in the knowledge graph that have an affiliation relationship with the node representing the desired processing level.

[0091] For example, the target user's initial intent is to "apply for a construction permit," and they expect the service level to be "county-level." The knowledge graph is searched for valid government service items related to the desired "county-level" service. If a valid government service item includes "apply for a construction permit," it means there is a relationship between "applying for a construction permit" and "county-level," resulting in a successful match. The target user is then sent an access path and a service guide. Otherwise, the match is considered unsuccessful, and a second prompt or a second adjustment to the government service path is sent to the target user.

[0092] The second prompt message is an interactive message from the government affairs Q&A system to the target user, used to inform the target user that the selected desired processing level does not have the authority to process the initial intention. For example, the second prompt message is "Processing level authority mismatch result".

[0093] The second adjustment of the government service path represents the alternative processing path automatically recommended by the system when the initial intention does not belong to the set of valid government service items but there are valid government service items with semantic matching.

[0094] In one embodiment, the government service item associated with the processing path semantically matches the initial intent and has a hierarchical relationship with the desired processing level; in another embodiment, the processing path is an alternative processing level that has a hierarchical relationship with the initial intent, retrieved based on a knowledge graph.

[0095] For example, if the valid government service item that matches the initial intent semantics is "construction project filing", the target user is sent a second adjusted government service path: "It has been detected that you may need to apply for 'construction project filing'. This item can be processed at the 'county level'. Do you want to jump to the next level?" or "It has been detected that you may need to apply for 'construction permit'. This item can be processed at the 'city level'. Do you want to jump to the next level?" At the same time, relevant processing guidelines are sent.

[0096] According to the embodiments of this application, by verifying the semantic association and attribution relationship of the knowledge graph, when the user selects a processing level that does not have the corresponding authority, the system can provide real-time feedback on the mismatch of level authority or send an alternative second adjustment of the government service path, thereby avoiding the target user from being ineffectively transferred at the wrong level and reducing the trial and error cost of handling affairs caused by the misselection of administrative level.

[0097] According to embodiments of this application, the compliance prerequisites include qualification certification information. The dialogue interaction method further includes: responding to the qualification certification information supplemented by the target user, retrieving valid government service items with an attribution relationship to the qualification certification information in the knowledge graph; and, in the case that the initial intent does not match the valid government service items, sending a third prompt message or a third adjustment of the government service path to the target user. The third adjustment of the government service path represents the processing path of the qualification certification information that is dependent on the initial intent, and the third prompt message represents the result of the compliance prerequisites not matching.

[0098] The qualification certification information is supplementary supporting documentation for the target user to meet the prerequisites for compliance in government services. It is generally determined based on the qualification certificates and licenses currently held by the target user, such as business licenses and health permits.

[0099] The attribution relationship refers to the mapping between government service items and qualification certification information as legal prerequisites. For example, the "issuance of business license" item and the "business license" qualification certification information have an attribution relationship, meaning that a valid business license is a prerequisite for compliance in handling this item.

[0100] A valid government service item that has an attribution relationship with qualification certification information is a set of nodes representing government service items in the knowledge graph that have an attribution relationship with the nodes representing qualification certification information.

[0101] For example, the target user's initial intent is to "apply for a transportation business license," and the supplementary qualification certificate information is a "motor vehicle driver's license." The knowledge graph is searched for valid government service items related to the qualification certificate information "motor vehicle driver's license." If a valid government service item includes "apply for a transportation business license," it means there is a relationship between "apply for a transportation business license" and "motor vehicle driver's license," resulting in a successful match. The target user is then sent an application entry path and application guide. Otherwise, the match is considered unsuccessful, and the target user is sent a third-party prompt or a third-party adjustment to the government service path.

[0102] The third prompt message is an interactive message from the government affairs Q&A system to the target user, used to inform the target user that the supplementary qualification certificate information does not meet the prerequisites for the initial intention to apply. For example, the third prompt message is "Compliance prerequisites do not match result".

[0103] Third, when adjusting the government service path representation, if the initial intention does not belong to the set of valid government service items, the system will automatically recommend a processing path that is dependent on the qualification certificate information of the initial intention.

[0104] For example, if the knowledge graph retrieves the qualification certificate information "business license" that is dependent on the initial intent "to apply for a transportation operation license", the system will send the target user the application path for "business license".

[0105] According to the embodiments of this application, by verifying the semantic association and attribution relationship of the knowledge graph, when the qualification certificate information supplemented by the user does not meet the prerequisites for the corresponding matter, the system can provide real-time feedback on the non-matching result of the compliance prerequisites and push the processing path of qualification certificate information that is dependent on the initial intention, thereby avoiding invalid declarations by the target user when the prerequisite qualifications are missing, and improving the efficiency and accuracy of handling affairs.

[0106] According to an embodiment of this application, the entity slots associated with the initial intent are detected. If there are missing entity slots, an interactive information completion process for the missing entity slots is triggered, including: determining a preset set of entity slots associated with the initial intent in the knowledge graph; detecting the filling status of each of the multiple entity slots in the preset entity slot set according to the dialogue data; and triggering an interactive information completion process for the missing entity slots if there are missing entity slots with an empty filling status.

[0107] The preset entity slot set is the set of other entities in the knowledge graph that are connected to the government service item entity through edge relationships. The other entity set is the set of parameter information necessary for handling the government service item, such as the place of handling, the level of handling, the scope of business, the type of qualification certificate, etc.

[0108] The fill status is a boolean flag indicating whether each entity slot has been assigned a value, as detected by the system based on the dialogue data. If the target user has explicitly provided the information corresponding to the entity slot in the dialogue, the entity slot fill status is filled; otherwise, it is empty.

[0109] Missing entity slots are parameter items whose entity slot filling status is empty after detection in the preset entity slot set.

[0110] The interactive information completion process is a multi-round dialogue and follow-up mechanism that is proactively triggered by the government affairs Q&A system after identifying missing entity slots. By sending inquiry messages to the target user regarding the missing entity slots, it guides the user to fill in the missing parameters until all entity slots are filled or the user explicitly gives up the process.

[0111] For example, when a target user inputs the dialogue data "How to open a store" in the first round, semantic matching using a knowledge graph initially identifies the initial intent as "registration of an individual business". The knowledge graph retrieves a set of pre-defined entity slots associated with the initial intent, detecting slots that need to be filled, including "Registration Location", "Operator Identity Type", and "Proof of Business Premises". Based on the dialogue data, the filling status of each of these entity slots is checked. If the "Registration Location" slot is found to be empty, the system automatically triggers an interactive information completion process for the missing slot, sending a follow-up message to the target user: "In which city are you applying?" After the target user replies with "City X1, District X2", the system marks the "Registration Location" slot as filled. Through rounds of interactive information completion, until all necessary parameters in the pre-defined entity slot set are filled, the system pushes a precise application entry path and application guide to the target user.

[0112] According to the embodiments of this application, by using a set of preset entity slots in the knowledge graph and dynamic filling status detection, the system can accurately identify information gaps in multi-turn dialogues, proactively trigger interactive information completion processes, and avoid target users repeatedly submitting or misselecting processing paths due to missing key information, thereby improving the guidance accuracy and efficiency of government service processing processes.

[0113] According to an embodiment of this application, triggering an interactive information completion process for a missing entity slot includes: determining a guiding query template associated with the missing entity slot from a preset template library; sending the guiding query template to a target user; and in response to the target user's completion data for the guiding query template, extracting the missing entity information associated with the missing entity slot from the completion data and filling the missing entity information into the missing entity slot.

[0114] The guided inquiry template is a standardized follow-up message for missing entity slots, matched and called from a preset template library, used to guide the target user to supplement the missing parameter information in natural language.

[0115] The preset template library is a set of standardized query statements pre-configured in the system and corresponding to each entity slot. It stores corresponding guiding question templates and multi-turn dialogue branch logic for different missing parameters.

[0116] For example, a guiding question template for addressing the lack of a physical slot associated with "health certificate" would be: "Please provide valid health certificates for your store employees. Have you already obtained them?" This guiding question template would be sent to the target user.

[0117] The supplementary data refers to the response content containing the required information that the target user provides during the dialogue interaction after receiving the guided question template.

[0118] Missing entity information refers to the valid parameter values ​​corresponding to the missing entity slots extracted by the system after performing semantic parsing and entity recognition on the completed data.

[0119] For example, if the target user replies with the complete data: "All employees hold", the system will extract the missing entity information "hold" from the complete data and fill it into the entity slot of the compliance prerequisite.

[0120] According to the embodiments of this application, by mapping the preset template library with missing entity slots, the system can automatically generate and send standardized guided inquiry templates, thereby accurately extracting the data to be completed and automatically filling it into the corresponding entity slots. This achieves closed-loop management of information collection, effectively reduces repetitive input and ambiguity of expression by target users, and improves the standardization and efficiency of multi-round dialogues in government services.

[0121] According to embodiments of this application, semantic recognition of dialogue data carried in a dialogue request using a knowledge graph to obtain an initial intent includes: mapping the dialogue data and entities representing government service items in the knowledge graph to a target semantic embedding space to obtain a first semantic vector and a second semantic vector, wherein the first semantic vector represents the vector representation of the dialogue data in the target semantic embedding space, and the second semantic vector represents the vector representation of the government service item in the target semantic embedding space; performing semantic matching on the first semantic vector and the second semantic vector to obtain a semantic similarity; and determining entities with a semantic similarity greater than a preset threshold as the initial intent.

[0122] The target semantic embedding space is a high-dimensional vector space that has been pre-trained and fine-tuned by a large-scale government affairs corpus. It can map natural language text and structured entities into dense vector representations with semantic relevance.

[0123] In the target semantic embedding space, the dialogue data input by the target user is vectorized by a semantic encoder to obtain the first semantic vector. The first semantic vector represents the distributed semantic features of the dialogue data in the target semantic embedding space.

[0124] In the target semantic embedding space, each entity representing government service items in the knowledge graph is encoded to obtain a second semantic vector. The second semantic vector represents the structured semantic features of government service items in the target semantic embedding space.

[0125] Semantic similarity can be obtained by calculating the geometric distance or directional consistency between the first and second semantic vectors in the target semantic embedding space. Semantic similarity reflects the strength of the semantic association between the dialogue data and the government service items stored in the knowledge graph. Entities with semantic similarity greater than a preset threshold are identified as the initial intent.

[0126] A preset threshold is used as the decision boundary for intent determination. When the semantic similarity between the second semantic vector corresponding to the government service item and the first semantic vector of the user dialogue data is greater than the preset threshold, this standardized government service item is determined as the initial intent, thereby completing the accurate mapping of intent from unstructured dialogue data to standardized government service items.

[0127] According to embodiments of this application, by constructing a knowledge graph in the field of government affairs, the user's colloquial dialogue data and the entities of government service items in the knowledge graph are projected into the same semantic embedding space. The deep semantic similarity between the two is calculated to extract the initial intent, forming a non-linear mapping mechanism from the user's colloquial dialogue data to standardized government service items, bridging the semantic gap between colloquial expressions and government standard terminology, and achieving accurate identification of the initial intent.

[0128] Based on the above-described dialogue interaction method, embodiments of this application also provide a dialogue interaction device. The following will be combined with... Figure 4 The device is described in detail.

[0129] Figure 4 A structural block diagram of a dialogue interaction device according to an embodiment of this application is shown.

[0130] like Figure 4 As shown, the dialogue interaction device 400 in this embodiment includes an identification module 410, a detection module 420, and a sending module 430.

[0131] The identification module 410 is used to respond to the dialogue request of the target user, use the knowledge graph to perform semantic recognition on the dialogue data carried in the dialogue request, obtain the initial intent, the initial intent represents the government service item that is initially matched, the knowledge graph includes entities and edge relationships, entities represent government service elements, and edge relationships represent the association attribute relationships between entities.

[0132] The detection module 420 is used to detect the entity slots associated with the initial intent. If there are missing entity slots, it triggers an interactive information completion process for the missing entity slots.

[0133] The sending module 430 is used to send a government service path corresponding to the initial intent to the target user in response to the completion of information completion for the missing entity slot. The government service path carries government service element information corresponding to the entity slot.

[0134] According to embodiments of this application, the dialogue interaction device 400 further includes a first retrieval module and a first adjustment module.

[0135] The first search module is used to respond to the target user's requested jurisdiction and search the knowledge graph for valid government service items that have jurisdiction over the requested jurisdiction.

[0136] The first adjustment module is used to send a first prompt message or a first adjustment government service path to the target user when the initial intent does not match the valid government service item. The first adjustment government service path represents the processing path associated with the government service item that has jurisdiction over the expected processing location and semantically matches the initial intent. The first prompt message represents the result of mismatch between local authority.

[0137] According to embodiments of this application, the dialogue interaction device 400 further includes a second retrieval module and a second adjustment module.

[0138] The second search module is used to retrieve valid government service items that have a relationship with the desired processing level in the knowledge graph in response to the target user's supplementary information.

[0139] The second adjustment module is used to send a second prompt message or a second adjusted government service path to the target user when the initial intent does not match the valid government service item. The second adjusted government service path represents the processing path associated with the government service item that has a hierarchical relationship with the expected processing level and matches the semantics of the initial intent. The second prompt message represents the result of the mismatch between the processing level and the authority.

[0140] According to embodiments of this application, the dialogue interaction device 400 further includes a third retrieval module and a third adjustment module.

[0141] The third search module is used to respond to the qualification certificate information supplemented by the target user and search the knowledge graph for valid government service items that have an attribution relationship with the qualification certificate information.

[0142] The third adjustment module is used to send a third prompt message or a third adjustment government service path to the target user when the initial intent does not match the valid government service item. The third adjustment government service path represents the processing path of qualification certificate information that is dependent on the initial intent, and the third prompt message represents the result of mismatch of compliance prerequisites.

[0143] According to an embodiment of this application, the detection module 420 includes a first detection submodule, a second detection submodule, and a third detection submodule.

[0144] The first detection submodule is used to determine the set of preset entity slots associated with the initial intent in the knowledge graph.

[0145] The second detection submodule is used to detect the filling status of multiple entity slots in the preset entity slot set based on the dialogue data.

[0146] The third detection submodule is used to trigger an interactive information completion process for missing entity slots when there are missing entity slots with an empty filling status.

[0147] According to an embodiment of this application, the detection module 420 further includes a fourth detection submodule, a fifth detection submodule, and a sixth detection submodule.

[0148] The fourth detection submodule is used to determine the guiding query template associated with the missing entity slot from the preset template library.

[0149] The fifth detection submodule is used to send guiding question templates to target users.

[0150] The sixth detection submodule is used to respond to the target user's completion data for the guided query template, extract the missing entity information associated with the missing entity slot from the completion data, and fill the missing entity information into the missing entity slot.

[0151] According to an embodiment of this application, the identification module 410 includes a first identification submodule, a second identification submodule, and a third identification submodule.

[0152] The first identification submodule is used to map the dialogue data and entities representing government service items in the knowledge graph to the target semantic embedding space, respectively, to obtain a first semantic vector and a second semantic vector. The first semantic vector represents the vector representation of the dialogue data in the target semantic embedding space, and the second semantic vector represents the vector representation of the government service item in the target semantic embedding space.

[0153] The second identification submodule is used to perform semantic matching on the first semantic vector and the second semantic vector to obtain semantic similarity.

[0154] The third identification submodule is used to identify entities with semantic similarity greater than a preset threshold as the initial intent.

[0155] For example, any plurality of the identification module 410, detection module 420, and transmission module 430 can be implemented in a single module / submodule, or any one of these modules / submodules can be split into multiple modules / submodules. Alternatively, at least part of the functionality of one or more of these modules / submodules can be combined with at least part of the functionality of other modules / submodules and implemented in a single module / submodule. According to embodiments of this application, at least one of the identification module 410, detection module 420, and transmission module 430 can be at least partially implemented as hardware circuitry, such as a field-programmable gate array (FPGA), a programmable logic array (PLA), a system-on-a-chip, a system-on-a-substrate, a system-on-package, an application-specific integrated circuit (ASIC), or any other reasonable means of integrating or packaging the circuitry, or implemented in software, hardware, or firmware, or in any suitable combination of any of these three implementation methods. Alternatively, at least one of the identification module 410, detection module 420, and transmission module 430 can be at least partially implemented as a computer program module, which, when run, can perform corresponding functions.

[0156] It should be noted that the dialogue interaction device part in the embodiments of this application corresponds to the dialogue interaction method part in the embodiments of this application. The specific description of the dialogue interaction device part is referred to in the dialogue interaction method part, and will not be repeated here.

[0157] Figure 5 A block diagram of an electronic device suitable for implementing a dialogue interaction method according to an embodiment of this application is shown. Figure 5 The electronic device shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments of this application.

[0158] like Figure 5 As shown, an electronic device 500 according to an embodiment of this application includes a processor 501, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 502 or a program loaded from a storage portion 508 into a random access memory (RAM) 503. The processor 501 may include, for example, a general-purpose microprocessor (e.g., a CPU), an instruction set processor and / or an associated chipset and / or a special-purpose microprocessor (e.g., an application-specific integrated circuit (ASIC)), etc. The processor 501 may also include onboard memory for caching purposes. The processor 501 may include a single processing unit or multiple processing units for performing different actions of the method flow according to an embodiment of this application.

[0159] RAM 503 stores various programs and data required for the operation of electronic device 500. Processor 501, ROM 502, and RAM 503 are interconnected via bus 504. Processor 501 executes various operations of the method flow according to embodiments of this application by executing programs in ROM 502 and / or RAM 503. It should be noted that programs may also be stored in one or more memories other than ROM 502 and RAM 503. Processor 501 may also execute various operations of the method flow according to embodiments of this application by executing programs stored in one or more memories.

[0160] According to embodiments of this application, the electronic device 500 may further include an input / output (I / O) interface 505, which is also connected to a bus 504. The electronic device 500 may also include one or more of the following components connected to the input / output (I / O) interface 505: an input section 506 including a keyboard, mouse, etc.; an output section 507 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 508 including a hard disk, etc.; and a communication section 509 including a network interface card such as a LAN card, modem, etc. The communication section 509 performs communication processing via a network such as the Internet. A drive 510 is also connected to the input / output (I / O) interface 505 as needed. A removable medium 511, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on the drive 510 as needed so that computer programs read from it can be installed into the storage section 508 as needed.

[0161] According to embodiments of this application, the method flow according to embodiments of this application can be implemented as a computer software program. For example, embodiments of this application include a computer program product comprising a computer program carried on a computer-readable storage medium, the computer program containing program code for performing the methods shown in the flowchart. In such embodiments, the computer program can be downloaded and installed from a network via communication section 509, and / or installed from removable medium 511. When the computer program is executed by processor 501, it performs the functions defined in the system of embodiments of this application. According to embodiments of this application, the systems, devices, apparatuses, modules, units, etc., described above can be implemented by computer program modules.

[0162] This application also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiments; or it may exist independently and not assembled into the device / apparatus / system. The aforementioned computer-readable storage medium carries one or more programs, which, when executed, implement the dialogue interaction method according to the embodiments of this application.

[0163] According to embodiments of this application, the computer-readable storage medium can be a non-volatile computer-readable storage medium. Examples include, but are not limited to: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this application, the computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0164] For example, according to embodiments of this application, a computer-readable storage medium may include the ROM 502 and / or RAM 503 described above and / or one or more memories other than ROM 502 and RAM 503.

[0165] Embodiments of this application also include a computer program product comprising a computer program containing program code for performing the methods provided in the embodiments of this application. When the computer program product is run on an electronic device, the program code is used to enable the electronic device to implement the dialogue interaction method provided in the embodiments of this application.

[0166] When the computer program is executed by the processor 501, it performs the functions defined in the system / apparatus of this application embodiment. According to the embodiments of this application, the systems, apparatuses, modules, units, etc., described above can be implemented by computer program modules.

[0167] In one embodiment, the computer program may rely on a tangible storage medium such as an optical storage device or a magnetic storage device. In another embodiment, the computer program may also be transmitted and distributed in the form of signals over a network medium, and may be downloaded and installed via the communication section 509, and / or installed from a removable medium 511. The program code contained in the computer program can be transmitted using any suitable network medium, including but not limited to: wireless, wired, etc., or any suitable combination thereof.

[0168] According to embodiments of this application, program code for executing the computer programs provided in the embodiments of this application can be written in any combination of one or more programming languages. Specifically, these computational programs can be implemented using high-level procedural and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages ​​include, but are not limited to, languages ​​such as Java, C++, Python, "C", or similar programming languages. The program code can be executed entirely on the user's computing device, partially on the user's device, partially on a remote computing device, or entirely on a remote computing device or server. In cases involving remote computing devices, the remote computing device can be connected to the user's computing device via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computing device (e.g., via the Internet using an Internet service provider).

[0169] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions. Those skilled in the art will understand that the features described in the various embodiments of this application can be combined and / or combined in various ways, even if such combinations are not explicitly described in this application. In particular, without departing from the spirit and teachings of this application, the features described in the various embodiments of this application can be combined and / or combined in various ways. All such combinations and / or combinations fall within the scope of this application.

[0170] The embodiments of this application have been described above. However, these embodiments are merely illustrative and not intended to limit the scope of this application. Although various embodiments have been described above, this does not mean that the measures in the various embodiments cannot be used advantageously in combination. Without departing from the scope of this application, those skilled in the art can make various substitutions and modifications, all of which should fall within the scope of this application.

Claims

1. A dialogue interaction method, comprising: In response to a dialogue request from a target user, the knowledge graph is used to perform semantic recognition on the dialogue data carried in the dialogue request to obtain an initial intent. The initial intent represents a preliminarily matched government service item. The knowledge graph includes entities and edge relationships. The entities represent government service elements, and the edge relationships represent the association attribute relationships between entities. The entity slots associated with the initial intent are detected, and if there are missing entity slots, an interactive information completion process for the missing entity slots is triggered. In response to the completion of information completion for the missing entity slot, a government service path corresponding to the initial intent is sent to the target user, the government service path carrying government service element information corresponding to the entity slot.

2. The method according to claim 1, wherein, The knowledge graph is constructed based on the following operations: Entities are extracted from structured government data, and the entities include at least one of the following: government service items, processing levels, compliance prerequisites, and processing locations; Based on the jurisdictional relationship between the government service item and the processing location, the dependency relationship between the government service item and the compliance prerequisite, and the affiliation relationship between the government service item and the processing level, an edge relationship is constructed; The knowledge graph is constructed based on the entities and the edge relationships.

3. The method according to claim 2, further comprising: In response to the target user's supplementary request for a desired processing location, the system retrieves valid government service items that have jurisdictional relationships with the desired processing location from the knowledge graph. If the initial intent does not match the valid government service item, a first prompt message or a first adjustment of the government service path is sent to the target user. The first adjustment of the government service path represents the processing path associated with the government service item that has jurisdiction over the expected processing location and semantically matches the initial intent. The first prompt message represents the result of mismatch in jurisdictional authority.

4. The method according to claim 2, further comprising: In response to the target user's supplementary expected processing level, retrieve valid government service items that have an attribution relationship with the expected processing level in the knowledge graph; If the initial intent does not match the valid government service item, a second prompt message or a second adjusted government service path is sent to the target user. The second adjusted government service path represents the processing path associated with a government service item that has a hierarchical relationship with the expected processing level and semantically matches the initial intent. The second prompt message represents the result of mismatch between processing level permissions.

5. The method according to claim 2, wherein, The aforementioned compliance prerequisites include qualification certification information, and the method further includes: In response to the qualification certificate information supplemented by the target user, the system retrieves valid government service items that have an attribution relationship with the qualification certificate information from the knowledge graph. If the initial intent does not match the valid government service item, a third prompt message or a third adjustment of the government service path is sent to the target user. The third adjustment of the government service path represents the processing path of qualification certificate information that is dependent on the initial intent, and the third prompt message represents the result of mismatch of compliance prerequisites.

6. The method according to claim 1, wherein, The entity slots associated with the initial intent are detected. If a missing entity slot exists, an interactive information completion process for the missing entity slot is triggered, including: In the knowledge graph, determine a set of preset entity slots associated with the initial intent; Based on the dialogue data, detect the filling status of each of the multiple entity slots in the preset entity slot set; If there are missing entity slots that are not filled, an interactive information completion process for the missing entity slots is triggered.

7. The method according to claim 6, wherein, Trigger an interactive information completion process for the missing entity slot, including: Determine a guiding query template associated with the missing entity slot from the preset template library; Send the guided inquiry template to the target user; In response to the target user's completion data for the guided query template, the missing entity information associated with the missing entity slot is extracted from the completion data, and the missing entity information is filled into the missing entity slot.

8. The method according to claim 1, wherein, Using a knowledge graph, semantic recognition is performed on the dialogue data carried in the dialogue request to obtain the initial intent, including: The dialogue data and the entities representing government service items in the knowledge graph are respectively mapped to the target semantic embedding space to obtain a first semantic vector and a second semantic vector. The first semantic vector represents the vector representation of the dialogue data in the target semantic embedding space, and the second semantic vector represents the vector representation of the government service item in the target semantic embedding space. Semantic matching is performed on the first semantic vector and the second semantic vector to obtain semantic similarity; Entities with semantic similarity greater than a preset threshold are identified as the initial intent.

9. A dialogue interaction device, the device comprising: The identification module is used to respond to the dialogue request of the target user, and use a knowledge graph to perform semantic recognition on the dialogue data carried by the dialogue request to obtain an initial intent. The initial intent represents the government service item that is initially matched. The knowledge graph includes entities and edge relationships. The entities represent government service elements, and the edge relationships represent the association attribute relationships between entities. The detection module is used to detect the entity slots associated with the initial intent, and if there are missing entity slots, it triggers an interactive information completion process for the missing entity slots. The sending module is used to send a government service path corresponding to the initial intent to the target user in response to the completion of information completion for the missing entity slot. The government service path carries government service element information corresponding to the entity slot.

10. An electronic device, comprising: One or more processors; Memory, used to store one or more programs. Wherein, when the one or more programs are executed by the one or more processors, the one or more processors implement the method of any one of claims 1 to 8.

11. A computer-readable storage medium having stored thereon executable instructions that, when executed by a processor, cause the processor to perform the method of any one of claims 1 to 8.

12. A computer program product comprising a computer program that, when executed by a processor, implements the method according to any one of claims 1 to 8.

Images