Knowledge Query Enhancement Method and Electronic Devices Based on Process Semantics
By identifying target process nodes and generating retrieval constraints from the process semantic information set, the problem of mismatch between business logic and user intent in knowledge query enhancement in the knowledge base is solved, and higher quality knowledge query results are achieved.
Patent Information
- Application Number
- CN202611341999.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-08-31
- Publication Date
- 2026-11-17
AI Technical Summary
When existing technologies enhance knowledge queries in knowledge bases, they struggle to understand the business logic and process structure behind the knowledge, resulting in search results containing a large amount of irrelevant content and failing to accurately reflect the user's intent and attribute information. Consequently, the generated answers lack business logic and completeness.
By identifying target process nodes from the set of process semantic information, generating retrieval constraints using semantic mapping information and user profile information, and performing knowledge base retrieval, the retrieval results are ensured to match user intent and business logic.
It improves the accuracy and quality of knowledge retrieval, ensuring that search results match user intent and business logic, and generating more complete and actionable answers.
Smart Images

Figure CN123115384A_ABST
Abstract
Description
Technical Field
[0001] This disclosure relates to the field of computer information technology, and in particular to a knowledge query enhancement method and electronic device based on process semantics. Background Technology
[0002] In related technologies, retrieval enhancement operations in the knowledge question answering process mainly employ tag retrieval, keyword retrieval, full-text retrieval, and vector retrieval to obtain semantically relevant data from the knowledge base.
[0003] However, for knowledge bases containing knowledge content across multiple business processes, the retrieved knowledge content may contain interfering information. Furthermore, within an enterprise or in a specific business scenario, the focus on the same issue may differ, resulting in less than ideal performance of general knowledge base retrieval methods in the knowledge query enhancement process. Summary of the Invention
[0004] In view of this, this disclosure provides a knowledge query enhancement method and electronic device based on process semantics.
[0005] One aspect of this disclosure provides a knowledge query enhancement method based on process semantics, comprising: in response to receiving a query request, determining a target process node matching the query request in a process semantic information set, the process semantic information set including at least one process node included in at least one business domain and the association between process nodes, each process node having semantic mapping information, and the target process node being the process node whose semantic mapping information matches the query request; determining retrieval constraints for the query request based on the target process node; and retrieving the knowledge base using the retrieval constraints and the query request to obtain the retrieval results of the query request.
[0006] According to embodiments of this disclosure, determining a target process node that matches a query request within a process semantic information set includes: determining the semantic element features of the query request; determining candidate process nodes that match the semantic element features within the process semantic information set; and determining the target process node based on the candidate process nodes and the query request.
[0007] According to embodiments of this disclosure, determining a target process node based on candidate process nodes and a query request includes: obtaining the allowed access level of the object initiating the query request; determining a matching score between the query request and each candidate process node, wherein the matching score is determined based on at least one of the following: text matching score between the query request and each candidate node, vector similarity between the query request and each candidate node, matching degree of semantic mapping information between the query request and each candidate node, fit between the allowed access level and each candidate process node, and distance between the process node corresponding to the context information of the query request and each candidate process node; and selecting candidate process nodes whose matching scores meet the score requirements as target process nodes.
[0008] According to embodiments of this disclosure, process nodes have hierarchical information; based on the target process node, the retrieval constraints of the query request are determined, including: obtaining user profile information of the object initiating the query request, the user profile information including at least one of the hierarchical information allowed for object access, the object's semantic completion strategy, the object's knowledge data preference type, and the object's output content constraints; based on the user profile information and the target process node, the retrieval constraints of the query request are determined.
[0009] According to embodiments of this disclosure, the retrieval constraints for a query request are determined based on user profile information and target process nodes, including: filtering target process nodes using user profile information to obtain filtered target process nodes, and matching the filtered target process nodes with user profile information; generating retrieval constraints based on the filtered target process nodes and nodes in the process semantic information set that satisfy association conditions with the filtered target process nodes, and using the retrieval constraints to constrain the acquisition of knowledge data in the knowledge base that matches the filtered target process nodes or nodes that satisfy association conditions.
[0010] According to embodiments of this disclosure, process nodes also have domain information, and the process semantic information set also includes rule constraint nodes. The associations between process nodes in the process semantic information set are execution order associations, hierarchical inclusion associations, or cross-domain dependency associations. There are constraint associations between process nodes and rule constraint nodes. It also includes: taking process nodes that have execution order associations, hierarchical inclusion associations, or cross-domain dependency associations with the target process node and that meet the distance requirements between nodes as nodes that satisfy the association conditions of the target process node.
[0011] According to embodiments of this disclosure, retrieving a knowledge base using retrieval constraints and a query request to obtain retrieval results for the query request includes: adjusting the query request using retrieval constraints to obtain an adjusted query request, wherein the adjusted query request includes a knowledge retrieval scope generated based on the retrieval constraints; and retrieving knowledge data in the knowledge base that falls within the knowledge retrieval scope according to the adjusted query request's knowledge retrieval scope.
[0012] According to embodiments of this disclosure, the process semantic information set is determined in the following manner: based on the business processing processes of multiple business domains, the process nodes, business action nodes, participating role nodes, system nodes associated with the operation, and rule constraint nodes of each business processing process, as well as the associations between nodes, are obtained to generate a process semantic information set; multiple process nodes representing the same semantics are subjected to semantic normalization processing to generate a process node and semantic mapping information including the process node.
[0013] According to embodiments of this disclosure, it further includes: in response to the existence of cross-domain execution dependencies between process nodes in different business domains, establishing cross-domain dependency associations between process nodes in different business domains that have cross-domain execution dependencies in a process semantic information set.
[0014] Another aspect of this disclosure provides a knowledge query enhancement device based on process semantics, comprising: a first determining module, configured to, in response to receiving a query request, determine a target process node matching the query request in a process semantic information set, wherein the process semantic information set includes at least one process node included in at least one business domain and the association between process nodes, each process node having semantic mapping information, and the target process node being a process node whose semantic mapping information matches the query request; a second determining module, configured to determine retrieval constraints of the query request based on the target process node; and a retrieval module, configured to use the retrieval constraints and the query request to retrieve a knowledge base and obtain the retrieval result of the query request.
[0015] Another aspect of this disclosure provides an electronic device, including: a storage device for storing one or more programs; and one or more processors for running the programs to perform the following operations: in response to receiving a query request, determining a target process node matching the query request in a process semantic information set, the process semantic information set including at least one process node included in at least one business domain and the association between process nodes, each process node having semantic mapping information, and the target process node being the process node whose semantic mapping information matches the query request; determining retrieval constraints for the query request based on the target process node; and retrieving a knowledge base using the retrieval constraints and the query request to obtain retrieval results for the query request.
[0016] Another aspect of this disclosure provides a computer program product, including a computer program that, when executed by a processor, implements the above-described method.
[0017] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of this disclosure, nor is it intended to limit the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Attached Figure Description
[0018] The above and other objects, features and advantages of this disclosure will become clearer from the following description of embodiments with reference to the accompanying drawings, in which:
[0019] Figure 1 This is an exemplary system architecture diagram of an electronic device to which a process semantics-based knowledge query enhancement method and electronic device can be applied according to an embodiment of the present disclosure;
[0020] Figure 2 A flowchart illustrating a knowledge query enhancement method based on process semantics according to an embodiment of the present disclosure is shown.
[0021] Figure 3 This schematically illustrates a structural block diagram of a knowledge query enhancement apparatus based on process semantics according to an embodiment of the present disclosure;
[0022] Figure 4 A schematic block diagram of an electronic device that can be used to implement the process semantics-based knowledge query enhancement method of the present disclosure is shown. Detailed Implementation
[0023] The embodiments of this disclosure will now be described with reference to the accompanying drawings. Various details of the embodiments of this disclosure are included to aid understanding and should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of this disclosure. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.
[0024] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of this disclosure. 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.
[0025] 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.
[0026] When using expressions such as "at least one of A, B, and C," the meaning should generally be interpreted according to the understanding of someone skilled in the art. For example, "a system having at least one of A, B, and C" should include, but is not limited to, systems having A alone, having B alone, having C alone, having A and B, having A and C, having B and C, and / or having A, B, and C. Similarly, when using expressions such as "at least one of A, B, or C," the meaning should generally be interpreted according to the understanding of someone skilled in the art. For example, "a system having at least one of A, B, or C" should include, but is not limited to, systems having A alone, having B alone, having C alone, having A and B, having A and C, having B and C, and / or having A, B, and C.
[0027] It should also be noted that the directional terms mentioned in the embodiments, such as "up," "down," "front," "back," "left," and "right," are only for reference to the directions in the accompanying drawings and are not intended to limit the scope of protection of this disclosure. Throughout the drawings, the same elements are represented by the same or similar reference numerals. Conventional structures or constructions will be omitted where they may cause confusion in understanding this disclosure.
[0028] In the technical solutions disclosed herein, the collection, storage, use, processing, transmission, provision, disclosure, and application of data (including but not limited to user personal information) comply with the provisions of relevant laws and regulations, necessary confidentiality measures have been taken, and they do not violate public order and good morals.
[0029] In knowledge-based question-answering scenarios, knowledge bases are typically used to enhance the knowledge query for questions, identify relevant information within the knowledge base, and guide the question-answering model to generate the corresponding answer based on this information. Knowledge bases primarily store large amounts of documents, process specifications, case studies, and professional knowledge.
[0030] However, among the technologies that use knowledge bases to enhance knowledge retrieval, there is a problem that the knowledge in the knowledge base is difficult to be correctly understood, located, and reused in actual business scenarios.
[0031] Related technologies primarily employ tag retrieval, keyword retrieval, full-text retrieval, and vector retrieval techniques to enhance knowledge retrieval and improve knowledge availability. However, these technologies mainly focus on literal or semantic similarity in the knowledge retrieval enhancement process, failing to understand the underlying business logic, process structure, organizational division of labor, and upstream and downstream dependencies. Consequently, knowledge query enhancement solutions from these technologies often return a large amount of seemingly relevant but business-irrelevant content.
[0032] Furthermore, in the knowledge-based question-and-answer process, the content of the questions entered by users sometimes fails to accurately represent their intent. Users not only expect answers that match the question itself, but also that include content related to the relevant processes or systems. However, in the knowledge query enhancement process, the content determined by enhancement mainly consists of semantic or vector matching of the question content, failing to capture content with upstream and downstream process connections to the question. This results in the final answer lacking business logic, completeness, and operability.
[0033] Furthermore, the technologies used to enhance knowledge queries often overlook user attributes, leading to inconsistencies between the retrieved content and the user's actual attributes. This results in incomplete or misguided answers that ignore the user's focus and priorities. For instance, execution roles focus on system guidelines and operational procedures related to processes, approval roles focus on criteria and limitations, and management roles focus on higher-dimensional aspects such as the scope of impact and collaborative workflows.
[0034] Based on this, the embodiments of this disclosure improve the accuracy of the question-answering model in understanding questions by performing knowledge query retrieval on user queries or questions within a restricted information framework, wherein the restricted information framework includes business process nodes and mapping information between business process nodes, and can provide answers related to upstream and downstream business processes for questions.
[0035] Figure 1 This is a schematic diagram of an exemplary system architecture for an electronic device to which a process semantics-based knowledge query enhancement method and an electronic device can be applied, according to an embodiment of this disclosure. It should be noted that... Figure 1 The examples shown are merely examples of system architectures that can be applied to the embodiments of this disclosure, in order to help those skilled in the art understand the technical content of this disclosure, but do not mean that the embodiments of this disclosure cannot be used in other devices, systems, environments or scenarios.
[0036] like Figure 1 As shown, the system architecture 100 according to this embodiment may include terminal devices 101, 102, and 103, a network 104, and a server 105. The network 104 serves as a medium for providing a communication link between the terminal devices 101, 102, and 103 and the server 105. The network 104 may include various connection types, such as wired and / or wireless communication links, etc.
[0037] Users can use terminal devices 101, 102, and 103 to interact with server 105 via network 104 to receive or send messages, etc. Terminal devices 101, 102, and 103 can be various electronic devices with displays and web browsing capabilities, including but not limited to smartphones, tablets, laptops, and desktop computers.
[0038] Server 105 can be a server that provides various services, such as a backend management server that supports websites browsed by users using terminal devices 101, 102, and 103 (for example only). The backend management server can analyze and process data such as received user requests, and feed back the processing results (such as web pages, information, or data obtained or generated according to user requests) to the terminal devices.
[0039] It should be noted that the knowledge query enhancement method based on process semantics provided in this disclosure can generally be executed by server 105. Correspondingly, the knowledge query enhancement device based on process semantics provided in this disclosure can generally be located in server 105. The knowledge query enhancement method based on process semantics provided in this disclosure can also be executed by a server or server cluster that is different from server 105 and capable of communicating with terminal devices 101, 102, 103 and / or server 105. Correspondingly, the knowledge query enhancement device based on process semantics provided in this disclosure can also be located in a server or server cluster that is different from server 105 and capable of communicating with terminal devices 101, 102, 103 and / or server 105.
[0040] Figure 2 The flowchart illustrates a knowledge query enhancement method based on process semantics according to an embodiment of the present disclosure.
[0041] like Figure 2 As shown, the method includes operations S210~S230.
[0042] In operation S210, in response to receiving a query request, a target process node matching the query request is determined in the process semantic information set. The process semantic information set includes at least one process node included in at least one business domain and the association between process nodes. Each process node has semantic mapping information, and the target process node is the process node whose semantic mapping information matches the query request.
[0043] A query request can be a question or query information entered by the querying party. A query request can be unimodal data such as text or audio, or multimodal data including at least two of the following: text, audio, image, and video.
[0044] The process semantic information set can be information pre-extracted in a structured manner based on the process regulations and system metadata provided by the management object. The process semantic information set can include multiple process nodes, which can have relationships. These relationships can include execution order relationships and hierarchical inclusion relationships. Hierarchical inclusion relationships are used to represent the execution order of each node in the process or the subordinate relationship between nodes at different granularities. If there are relationships between process nodes at different levels, a higher-level process node can be the parent node of a lower-level node, and a lower-level node can be the child node of a higher-level node.
[0045] For example, nodes 1, 2, 3, and 4 can be generated in the process semantic information set based on information related to the production, sales, and after-sales service of product A. Node 1 includes a general introduction to product A, node 2 includes information related to the production of product A, node 3 includes information related to the sales of product A, and node 4 includes information related to the after-sales service of product A. Node 1 has a first association relationship with nodes 2, 3, and 4, indicating that nodes 2, 3, and 4 are used to provide detailed explanations of node 1. Node 2 has a second association relationship with node 3, and node 3 has a second association relationship with node 4, indicating that nodes 2, 3, and 4 have a sequential execution relationship.
[0046] Semantic mapping information represents different textual expressions of semantic matching with process nodes. For example, semantic mapping information can be a node mapping table, including process nodes and nodes in the process semantic information set that have a mapping relationship with the process nodes. Semantic mapping information can also include non-standard query content, used to map non-standard query content to process nodes. For example, semantic mapping information can be an alias mapping table, including keyword information of process nodes and alias information of keyword information.
[0047] The target process node can be the process node in the process semantic information set whose node information or semantic mapping information has the highest matching degree with the query request or whose matching degree is greater than the matching degree threshold. The node information can be the process node's metadata or tag information, such as including at least one of the following: node name, the business domain to which the node belongs, and the unique identifier of the process node.
[0048] In operation S220, based on the target process node, the retrieval constraints of the query request are determined.
[0049] Search constraints include structured instructions or parameters used to limit the scope and rules of knowledge retrieval.
[0050] Search constraints can be determined based on the attribute information of the target process node. For example, if the attribute of the target process node represents the process information of salary verification, then the corresponding search constraint can be determined as retrieving knowledge of the operation guide type in the human resources management knowledge domain.
[0051] In operation S230, the knowledge base is searched using retrieval constraints and query requests to obtain the retrieval results of the query requests.
[0052] A knowledge base can be used to enhance the knowledge retrieval of a query request, obtaining knowledge associated with the query request. This knowledge can then be used to improve the effectiveness of the question-answering model in responding to the query request.
[0053] The retrieval results of the query request are knowledge information in the knowledge base that meets the retrieval constraints and matches the query request.
[0054] In the embodiments of this disclosure, by determining the target process node matching the query request from the process semantic information set, generating retrieval constraints corresponding to the query request using the target process node, and using the retrieval constraints and the query request to perform a retrieval in the knowledge base, the retrieval scope in the knowledge base is limited to the space corresponding to the retrieval constraints. This avoids knowledge that semantically matches the keywords of the query request but is irrelevant to the business process during the retrieval process, thus improving the quality of knowledge query enhancement. Furthermore, by utilizing the process nodes in the process semantic information set and the semantic mapping information of the process nodes to determine the target process node matching the query request, the accuracy of determining the process node matching the intent expressed by the query request is improved, enabling the determination of the corresponding target process node based on the non-canonical representation in the query request.
[0055] Candidate process nodes that match the semantic features of the query request can be initially screened from the process semantic information set. Then, the candidate process nodes can be further screened in detail to obtain the target process node that matches the query request.
[0056] According to embodiments of this disclosure, determining a target process node matching a query request from a process semantic information set includes: determining the semantic element features of the query request; then, determining candidate process nodes matching the semantic element features from the process semantic information set; and then, determining the target process node based on the candidate process nodes and the query request.
[0057] Process nodes in the process semantic information set can have multi-dimensional label information, which can represent semantic feature elements of different dimensions. For example, a process node can include at least one of the following label information: domain feature, action feature, business object feature, time feature, condition feature, and user role feature.
[0058] Semantic element features represent the semantic elements included in a query request. Semantic element features can be feature information from at least one of a set of predefined dimensions. For example, semantic element features can include at least one of the following: domain features, action features, business object features, time features, condition features, and user role features.
[0059] Domain features are used to represent the business domain to which the semantic element features of a process node or query request belong. For example, they may include human resources management, finance, sales, and development domains.
[0060] Action features are used to represent business actions represented by process nodes or query requests, such as application, approval, and adjustment actions.
[0061] Business object characteristics can represent the business object corresponding to the business represented by the process node or query request, such as the target business name, target product name, or target personnel name.
[0062] Time features are used to represent the valid time of a process node or the target time information that a query request wants to retrieve.
[0063] Conditional features can include triggering conditions for process nodes or triggering conditions included in the query request. For example, a conditional feature for a process node could be that the transaction amount is greater than m, indicating that the process node will only be triggered if the transaction amount is greater than m.
[0064] User role characteristics can be used to represent the role information of the object initiating the query request. For example, if the role of the object initiating the query request is an administrator, then the semantic element characteristics used to determine the query request can include user role characteristics representing the administrator.
[0065] A process node that matches at least one of the feature information represented by the semantic element feature of the query request in the process semantic information set can be identified as a candidate process node.
[0066] By extracting semantic feature characteristics from the query request and performing a coarse screening of process nodes in the process semantic information set based on these semantic feature characteristics, candidate process nodes with semantic feature characteristics are obtained, reducing the number of process nodes to be screened in detail later, thereby reducing the computational load for determining the target process node.
[0067] The target process node can be determined from the candidate process nodes by comprehensively considering the semantic mapping hit rate, the degree of process level adaptation, and the consistency between the candidate process nodes and the current query context in the upstream and downstream structure of the process semantic skeleton.
[0068] Process nodes can have hierarchical information, which is used to represent the granularity of information within the process node and facilitates the representation of the hierarchical relationships between process nodes. This hierarchical information can be stored as tag information within the node information of the process node.
[0069] According to embodiments of this disclosure, determining a target process node based on candidate process nodes and a query request includes: obtaining the allowed access level of the object initiating the query request. Then, a matching score is determined between the query request and each candidate process node. The matching score is determined based on at least one of the following: text matching score between the query request and each candidate node; vector similarity between the query request and each candidate node; matching degree of semantic mapping information between the query request and each candidate node; fit between the allowed access level and each candidate process node; and distance between the process node corresponding to the context information of the query request and each candidate process node. Finally, candidate process nodes whose matching scores meet the score requirements are selected as the target process node.
[0070] Text matching score represents the degree of match between a query request and the text information of each candidate node. For example, the text matching score can be determined based on the keyword hit rate between the text content of the query request and the text content of each candidate node, or by using the Best Matching 25 (BM25) algorithm.
[0071] Vector similarity can be determined based on the encoded vectors of the query request and the candidate process nodes in the same vector space. For example, the cosine similarity between the vector of the query request and the vector of each candidate process node can be used as the vector similarity between the query request and each candidate node.
[0072] The matching degree between the query request and the semantic mapping information of each candidate node is used to represent the degree of hit between the query request and the semantic mapping information of the candidate process nodes.
[0073] The mapping between objects and allowed access levels can be pre-configured. In response to a received query request, the allowed access level of the object that initiated the query request can be determined based on the mapping between objects and allowed access levels.
[0074] The fit between the allowed access level and each candidate process node represents the granularity of the match between the level of the candidate process node and the allowed access level of the object that initiated the query request. The closer the level of the candidate process node is to or falls within the allowed range of the allowed access level, the higher the fit between the candidate process node and the allowed access level.
[0075] The distance between the process node corresponding to the context information of the query request and each candidate process node includes the degree of association between the target process node corresponding to the historical query request and each candidate process node.
[0076] The matching score can be obtained by weighted summing at least one of the following: the text matching score between the query request and each candidate node, the vector similarity between the query request and each candidate node, the matching degree of the semantic mapping information between the query request and each candidate node, the adaptability between the allowed access level and each candidate process node, and the distance between the process node corresponding to the context information of the query request and each candidate process node.
[0077] For example, the matching score can be determined using the following formula:
[0078] Where q represents the query request, p represents a candidate process node, C represents the context of the query request, and b1, b2, b3, b4, and b5 are preset weight values. The score is the text matching score between the query request and each candidate node. To calculate the vector similarity between the query request and each candidate node, To determine the degree of match between the query request and the semantic mapping information of each candidate node, To determine the compatibility between the allowed access level of the object initiating the query and each candidate process node, The distance between the process node corresponding to the context information C of the query request and the candidate process node p.
[0079] The score requirement can be a preset score threshold or the target number of selected target process nodes. You can select a target number of candidate process nodes with a matching score greater than the score threshold or the highest matching score as the target process nodes.
[0080] By comprehensively evaluating at least one dimension—text, semantics, aliases, hierarchy, and structure—the matching score between candidate process nodes and the query request is determined, enabling the accurate selection of the target process node with the highest matching degree from among the candidate process nodes. Furthermore, by determining the target process node based on the allowed access level of the object initiating the query request, information that does not match the object's permissions is avoided, ensuring the information security of knowledge retrieval enhancement for the query request. By introducing contextual structure information, the output weight of nodes related to the context of the query request is increased, improving the consistency between the knowledge provided for the query request and the knowledge of historical query requests.
[0081] In some embodiments, after determining at least one target process node, role constraint information can be generated based on the role profile of the object initiating the query request, and retrieval constraints can be obtained by combining the role constraint information and the target process node.
[0082] According to embodiments of this disclosure, determining retrieval constraints for a query request based on a target process node includes: obtaining user profile information of the object initiating the query request; the user profile information includes at least one of the following: hierarchical information allowed for object access, semantic completion strategy of the object, knowledge data preference type of the object, and output content constraints of the object; and determining retrieval constraints for the query request based on the user profile information and the target process node. The user profile information is preset and includes information describing the role, responsibilities, and scope of permissions of the object initiating the query request.
[0083] The allowed access level information of the object initiating the query request can be declared again in the user profile information, and the target process nodes can be filtered to avoid the target process nodes including nodes whose access level information does not match the allowed access level due to the small weight of the fit between the allowed access level and the candidate process nodes in the matching score.
[0084] Semantic completion strategies can be pre-designed for different user roles. After a query request is initiated for an object of that user role, the corresponding semantic completion strategy can be used to generate prompt words for retrieval or for question answering, or to perform semantic completion processing on the content of the query request.
[0085] The knowledge data preference type of an object represents the type of knowledge that a user role prioritizes when acquiring knowledge. For example, it can be used to implement role preference operation guidelines, approve role preference policy terms, and manage role preference process architecture and impact analysis.
[0086] The output content constraints of the object are rules used to limit the consistency verification of content generated by knowledge query enhancement, and to remove content that does not match the user role.
[0087] By combining the user profile information of the object initiating the query request with the target process node, the retrieval constraints of the query request are comprehensively determined. This enables knowledge query enhancement based on the information constraints of the object initiating the query request. In the process of knowledge query enhancement, knowledge matching their respective user profile information is provided to the objects, thereby improving the effectiveness of knowledge query enhancement.
[0088] In some embodiments, user role profiles and target process nodes can be combined to determine the retrieval constraints of a query request.
[0089] In some embodiments, user profile information can be used to filter target process nodes, and the filtered target process nodes can be used to generate retrieval constraints for query requests.
[0090] According to embodiments of this disclosure, determining retrieval constraints for a query request based on user profile information and target process nodes includes: filtering target process nodes using user profile information to obtain filtered target process nodes, and matching the filtered target process nodes with the user profile information. Based on the filtered target process nodes and nodes in the process semantic information set that satisfy association conditions with the filtered target process nodes, retrieval constraints are generated. These retrieval constraints are used to constrain the acquisition of knowledge data in the knowledge base that matches the filtered target process nodes or nodes that satisfy association conditions.
[0091] User profile information can be determined based on a unified identity system, organizational structure information, and the historical interaction behavior of the object that initiated the query request.
[0092] It can exclude process nodes in the target process that do not match the user profile information, and use the remaining target process nodes as the filtered target process nodes.
[0093] For example, the user profile information for object A includes a node that allows object A to access level 2 information, and a semantic completion strategy that adds the prompt text "1" to the search constraints. The user profile information for object B includes a node that allows object B to access level 3 information, and an output content constraint that excludes information from the management domain. If object A and object B initiate the same query request and their corresponding target process nodes are the same, then based on object A's user profile information, target process nodes with a level of no more than 2 can be selected to generate search constraints for object A. These search constraints for object A include target process nodes with a level of no more than 2 and the prompt text "1". Furthermore, based on object B's user profile information, target process nodes with a level of no more than 3 and whose domain information is not in the management domain can be determined, and search constraints for object B can be generated. By generating different search constraints for object A and object B based on their user profile information, different knowledge data can be provided to object A and object B respectively during the search phase according to these different search constraints.
[0094] By utilizing user profile information to filter target process nodes, it ensures that the hierarchy of the target process nodes used by users to generate search constraints matches the user's allowed access hierarchy, avoiding the return of knowledge that the user is not authorized to access during the search process. Furthermore, by combining the filtered target process nodes with nodes that meet the association conditions, search constraints are generated based on information within the process link in which the target process node resides, ensuring that the knowledge data returned during the search matches the constraint requirements of the business link in which the target process node is located. Moreover, by constructing user role profiles based on a unified identity system, organizational structure information, and users' historical interaction behavior, strong constraints are imposed on the semantic state of the process, causing the same query to enter different internal state spaces under different role conditions, thereby consistently outputting differentiated results that conform to role responsibilities and concerns.
[0095] In some embodiments, a node that meets the association criteria can be a process node that is associated with the filtered target process node, and can be a predecessor node, successor node, or child node of the filtered target process node. A child node is a node that is associated with the filtered target process node but has a smaller granularity than the filtered target process node.
[0096] According to embodiments of this disclosure, process nodes also possess domain information, and the process semantic information set further includes rule constraint nodes. The associations between process nodes in the process semantic information set may also include cross-domain dependency associations, and there are constraint associations between process nodes and rule constraint nodes. The knowledge query enhancement method based on process semantics further includes: identifying process nodes that have execution order associations, hierarchical inclusion associations, or cross-domain dependency associations with the target process node and that satisfy distance requirements as nodes that satisfy the association conditions of the target process node.
[0097] Rule constraint nodes are used to characterize the conditional branches or triggering rules that trigger the execution of associated process nodes.
[0098] Cross-domain dependency associations represent the execution order between process nodes with different domain information, and are used to reflect the execution process of cross-domain business.
[0099] A process node that is associated with the target process node in terms of execution order may include at least one of the preceding and following nodes of the target process node.
[0100] A process node that has a hierarchical relationship with the target process node can be a child node of the target process node.
[0101] The distance requirement between nodes can be either a graph distance threshold or a path weight threshold between the node and the target process node.
[0102] By setting rules to constrain nodes and defining association types, including execution order association, hierarchical inclusion association, and cross-domain dependency association, the system effectively selects nodes that meet the association conditions with the target process node. This maps query requests to the target process node and its upstream and downstream process dependencies, enabling the understanding of the true business intent of the query request based on business logic rather than textual surface similarity, thus avoiding semantic drift across processes and scenarios. Furthermore, by limiting the distance between associated process nodes, the system avoids the introduction of process nodes with low relevance due to unlimited expansion of associated nodes. Additionally, by setting cross-domain dependency associations, the system links process nodes from different business domains, ensuring that the process semantic information set accurately represents the execution order of processes across different domains.
[0103] According to embodiments of this disclosure, a knowledge base is retrieved using retrieval constraints and a query request to obtain the retrieval results for the query request, including:
[0104] The query request is adjusted using search constraints to obtain a revised query request, which includes the knowledge retrieval scope generated based on the search constraints. Then, knowledge data within the knowledge retrieval scope is retrieved from the knowledge base according to the revised query request.
[0105] Adjusting a query request can include injecting search constraints into the query request so that the adjusted query request includes the knowledge retrieval scope represented by the search constraints.
[0106] By transforming constraints into the knowledge retrieval scope within the query request and implementing constraint injection by adjusting the query request before retrieval execution, no adjustments to the module used for knowledge retrieval are required. Furthermore, the knowledge retrieval scope represents the constraint range of at least one dimension, including domain, node, hierarchy, and document type. This narrows the knowledge domain being retrieved from the knowledge base and ensures that the retrieved results do not include knowledge that is semantically similar to the query request but irrelevant to the corresponding business logic, thus improving retrieval accuracy.
[0107] The process semantic information set may also include business action nodes, participating role nodes, and system nodes associated with the operation. Business action nodes, participating role nodes, and system nodes associated with the operation can be associated with process nodes.
[0108] According to embodiments of this disclosure, the process semantic information set is determined in the following manner: based on the business processing processes of multiple business domains, the process nodes, business action nodes, participating role nodes, system nodes associated with the operation, and rule constraint nodes of each business processing process, as well as the associations between nodes, are obtained to generate a process semantic information set; multiple process nodes representing the same semantics are subjected to semantic normalization processing to generate a process node and semantic mapping information including the process node.
[0109] Business action nodes represent the operational behaviors of process nodes, establish relationships between process nodes and business action nodes, and indicate the action type of a process node. Participant role nodes represent the responsible roles in performing business actions, and there are associations between business action nodes and participant role nodes. System nodes associated with operations represent the system platform that executes the business action or process.
[0110] After generating nodes and their relationships based on each business process, nodes from different sources or with different representations but representing the same semantic process are merged, retaining one process node, and using the information of the removed nodes as the semantic mapping information of that process node.
[0111] By extracting process nodes, action nodes, role nodes, system nodes, and rule nodes from multiple business domains and organizing them into a graph that includes information about nodes and their inter-node relationships, a multi-dimensional and structured process semantic model is obtained. This enables the extraction and construction of a computable process semantic skeleton from multi-source business data. Furthermore, a Process Semantic Interlayer (PSI) is introduced between query understanding and knowledge retrieval, mapping natural language queries to process semantic states constrained by business process structure, user roles, and rules. This state serves as the unified control input for subsequent retrieval, sorting, and generation. In addition, semantic normalization processes unify the same business process concept from different sources and using different expressions into a single standard process node. Semantic mapping information establishes a mapping relationship between non-standard user expressions and standard nodes.
[0112] In some embodiments, the knowledge query enhancement method based on process semantics further includes:
[0113] In response to the existence of cross-domain execution dependencies between process nodes in different business domains, cross-domain dependency associations are established between process nodes in different business domains that have cross-domain execution dependencies in the process semantic information set.
[0114] Based on the business domain attributes of the process nodes, it can be determined whether the process nodes belong to different business domains.
[0115] If a process node in one business domain triggers or depends on a process node in another business domain during business processing, then it is determined that there is a cross-domain execution dependency between the two process nodes.
[0116] By establishing cross-domain dependency associations between process nodes that depend on words from different business domains, the constructed process semantic information set can represent the complete business process chain across business domains, and solves the problem of non-interoperability between business processes in different domains.
[0117] In one example, for a target enterprise, process nodes, business actions, participating roles, involved system nodes and rule constraints can be extracted from the enterprise's internal policy documents, standard operating procedures (SOPs), and system configuration metadata to form a process definition diagram (corresponding to a set of process semantic information) and an alias table (corresponding to semantic mapping information).
[0118] For cross-business domain enterprise processes, the system uses business tasks or business objects as a unified main line identifier. First, it extracts business process event sequences with clear business semantics from process instance data, business object status change records, or business audit trajectories in each business domain. Then, based on business object mapping relationships and process dependencies, it aligns and splices the event sequences of different business domains to generate a cross-business domain process definition diagram, completing the process sequence, branch structure, and weights.
[0119] In response to receiving a natural language query from a target user within the enterprise (corresponding to a query request), the system first performs a preliminary screening of the query information using at least one of the following: a domain keyword table, a lightweight classification model, and vector clustering results. This quickly determines whether the query belongs to a specific domain such as human resources, finance, sales, or technology development, or a cross-domain process involving multiple domains. Next, semantic elements such as actions, business objects, time, conditions, and roles are identified from the query information. These elements are then combined with keyword matching, alias mapping, and vector similarity to generate a set of candidate process nodes within the process semantic skeleton. Finally, within this candidate process node set, ambiguity is resolved by combining process semantic matching and upstream / downstream structural consistency information from the process semantic skeleton to determine the target process node.
[0120] After identifying the target process nodes, user role profiles built upon the enterprise's unified identity system, organizational structure information, and user historical interaction memories can be introduced to impose strong constraints on the process semantic state, allowing the same query to enter different internal state spaces under different role conditions. These strong constraints can include process hierarchy constraints, query semantic completion constraints, and evidence type constraints. Process hierarchy constraints limit the scope of process levels that can participate in subsequent processing based on the user's role responsibilities, allowing different roles to enter process semantic spaces of different granularities. Query semantic completion constraints select different semantic completion strategies based on roles to differentiate the completion of implicit behavioral goals, system entry points, or decision conditions in the original query. Evidence type constraints define the types of knowledge evidence that receive higher weight in subsequent retrieval and generation stages, ensuring that the output content is consistent with role responsibilities. For example, execution roles focus on operational steps and system guidelines, approval roles focus on policy terms and approval rules, and management roles focus on process architecture and impact analysis.
[0121] Define the types of knowledge evidence that receive higher weight in the retrieval and generation stages to ensure that the output content aligns with the user's role and responsibilities. For example, the execution role focuses on operational procedures and system guidelines, the approval role focuses on policy terms and approval rules, and the management role focuses on process architecture and impact analysis.
[0122] After generating the results, the consistency standard of user role differentiation is verified, and process nodes with low relevance to the role's focus are removed from the target process nodes, so as to realize the role profile as a key control variable affecting the system's information processing path.
[0123] After identifying the target process node, a restricted traversal is performed along the process semantic skeleton, following the relationships between preceding and subsequent processes, sub-processes, and rule constraints, starting from the target process node. This yields a set of process dependency chains related to the current query. Then, the target process node, process dependency chains, role constraints, system action scope, and rule conditions are uniformly encapsulated into a Process Semantic Context Package (PSCP). This context package serves as a strong constraint in the subsequent knowledge retrieval stage, explicitly defining the scope of process nodes, knowledge scope, and constraints that can participate in retrieval and generation.
[0124] After determining the context package, within the business domain defined by the context package, a hybrid search and recall process is performed around the target process node, dependency chain, and system scope. Candidate results are then filtered and comprehensively ranked to ensure that semantic similarity conforms to process matching, role suitability, and dependency chain consistency. During the answer generation phase, only search results defined by the context package are allowed to be used as knowledge evidence, preventing free generation that deviates from process semantics.
[0125] Figure 3A schematic block diagram of a knowledge query enhancement apparatus based on process semantics according to an embodiment of the present disclosure is shown.
[0126] like Figure 3 As shown, the knowledge query enhancement device 300 based on process semantics in this embodiment includes a first determining module 310, a second determining module 320, and a retrieval module 330.
[0127] The first determining module 310 is used to, in response to receiving a query request, determine a target process node that matches the query request in a process semantic information set. The process semantic information set includes at least one process node included in at least one business domain and the association between process nodes. Each process node has semantic mapping information, and the target process node is the process node whose semantic mapping information matches the query request.
[0128] The second determining module 320 is used to determine the retrieval constraints of the query request based on the target process node.
[0129] The retrieval module 330 is used to retrieve the knowledge base using retrieval constraints and query requests, and obtain the retrieval results of the query requests.
[0130] According to an embodiment of this disclosure, the first determining module 310 determines the target process node matching the query request in the process semantic information set in the following manner: determining the semantic element features of the query request; determining candidate process nodes matching the semantic element features in the process semantic information set; and determining the target process node based on the candidate process nodes and the query request.
[0131] According to an embodiment of this disclosure, the first determining module 310 determines the target process node based on the candidate process nodes and the query request in the following manner: obtaining the allowed access level of the object initiating the query request; determining the matching score between the query request and each candidate process node, wherein the matching score is determined based on at least one of the following: text matching score between the query request and each candidate node, vector similarity between the query request and each candidate node, matching degree of semantic mapping information between the query request and each candidate node, adaptability between the allowed access level and each candidate process node, and distance between the process node corresponding to the context information of the query request and each candidate process node; and selecting the candidate process node whose matching score meets the score requirement as the target process node.
[0132] According to embodiments of this disclosure, process nodes have hierarchical information. The second determining module 320 determines the retrieval constraints of the query request based on the target process node in the following manner: obtaining user profile information of the object initiating the query request, the user profile information including at least one of the hierarchical information allowed for object access, the object's semantic completion strategy, the object's knowledge data preference type, and the object's output content constraints; and determining the retrieval constraints of the query request based on the user profile information and the target process node.
[0133] According to embodiments of this disclosure, the second determining module 320 determines the retrieval constraints of the query request based on user profile information and target process nodes in the following manner: filtering target process nodes using user profile information to obtain filtered target process nodes, and matching the filtered target process nodes with user profile information; generating retrieval constraints based on the filtered target process nodes and nodes in the process semantic information set that satisfy association conditions with the filtered target process nodes, and using the retrieval constraints to constrain the acquisition of knowledge data in the knowledge base that matches the filtered target process nodes or nodes that satisfy association conditions.
[0134] According to embodiments of this disclosure, process nodes also have domain information, and the process semantic information set also includes rule constraint nodes. The associations between process nodes in the process semantic information set are execution order associations, hierarchical inclusion associations, or cross-domain dependency associations. There are constraint associations between process nodes and rule constraint nodes. The knowledge query enhancement device 300 based on process semantics also includes an association node determination module, which is used to: identify process nodes that have execution order associations, hierarchical inclusion associations, or cross-domain dependency associations with the target process node and that meet the distance requirements between nodes as nodes that satisfy the association conditions of the target process node.
[0135] According to embodiments of this disclosure, the retrieval module 330 retrieves the knowledge base using retrieval constraints and a query request in the following manner to obtain the retrieval results of the query request: the query request is adjusted using retrieval constraints to obtain an adjusted query request, the adjusted query request including a knowledge retrieval scope generated based on the retrieval constraints; and knowledge data within the knowledge retrieval scope in the knowledge base is retrieved according to the knowledge retrieval scope of the adjusted query request.
[0136] According to embodiments of this disclosure, the knowledge query enhancement device 300 based on process semantics further includes a process semantic information set determination module, which is used to: obtain the process nodes, business action nodes, participating role nodes, system nodes associated with the operation and rule constraint nodes, and the associations between nodes for each business processing flow based on the business processing flows of multiple business domains, and generate a process semantic information set; perform semantic normalization processing on multiple process nodes representing the same semantics to generate a process node and semantic mapping information including the process node.
[0137] According to embodiments of this disclosure, the process semantic information set determination module is further configured to: in response to the existence of cross-domain execution dependencies between process nodes in different business domains, establish cross-domain dependency associations between process nodes in different business domains that have cross-domain execution dependencies in the process semantic information set.
[0138] Figure 4 A schematic block diagram of an electronic device that can be used to implement the process semantics-based knowledge query enhancement method of the present disclosure is shown.
[0139] like Figure 4 As shown, an electronic device 400 according to an embodiment of the present disclosure includes a processor 401, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 402 or a program loaded from a storage portion 408 into a random access memory (RAM) 403. The processor 401 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 401 may also include onboard memory for caching purposes. The processor 401 may include a single processing unit or multiple processing units for performing different actions of the method flow according to an embodiment of the present disclosure.
[0140] RAM 403 stores various programs and data required for the operation of electronic device 400. Processor 401, ROM 402, and RAM 403 are interconnected via bus 404. Processor 401 performs various operations of the method flow according to embodiments of this disclosure by executing programs in ROM 402 and / or RAM 403. It should be noted that programs may also be stored in one or more memories other than ROM 402 and RAM 403. Processor 401 may also implement the methods provided in embodiments of this disclosure by executing programs stored in one or more memories.
[0141] According to embodiments of this disclosure, the electronic device 400 may further include an input / output (I / O) interface 405, which is also connected to a bus 404. The electronic device 400 may also include one or more of the following components connected to the I / O interface 405: an input section 406 including a keyboard, mouse, etc.; an output section 407 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 408 including a hard disk, etc.; and a communication section 409 including a network interface card such as a LAN card, modem, etc. The communication section 409 performs communication processing via a network such as the Internet. A drive 410 is also connected to the I / O interface 405 as needed. A removable medium 411, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on the drive 410 as needed so that computer programs read from it can be installed into the storage section 408 as needed.
[0142] This disclosure 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 computer-readable storage medium carries one or more programs that, when executed, implement the method according to the embodiments of this disclosure.
[0143] According to embodiments of this disclosure, the computer-readable storage medium may be a non-volatile computer-readable storage medium, such as including, but 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 disclosure, the computer-readable storage medium may be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. For example, according to embodiments of this disclosure, the computer-readable storage medium may include ROM 402 and / or RAM 403 and / or one or more memories other than ROM 402 and RAM 403 described above.
[0144] Embodiments of this disclosure also include a computer program product comprising a computer program containing program code for performing the methods shown in the flowchart. When the computer program product is run on a computer system, the program code is used to cause the computer system to implement the methods provided in the embodiments of this disclosure.
[0145] When the computer program is executed by the processor 401, it performs the functions defined in the system / apparatus of this disclosure embodiments. According to embodiments of this disclosure, the systems, apparatuses, modules, units, etc., described above can be implemented by computer program modules.
[0146] 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 downloaded and installed via communication section 409, and / or installed from removable medium 411. 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.
[0147] In such an embodiment, the computer program can be downloaded and installed from a network via communication section 409, and / or installed from removable medium 411. When the computer program is executed by processor 401, it performs the functions defined in the system of this disclosure embodiment. According to embodiments of this disclosure, the systems, devices, apparatuses, modules, units, etc., described above can be implemented by computer program modules.
[0148] Embodiments of this disclosure also include an electronic device, comprising: a storage device for storing one or more programs; and one or more processors for running the programs to perform the following operations: in response to receiving a query request, determining a target process node matching the query request in a process semantic information set, the process semantic information set including at least one process node included in at least one business domain and the association between process nodes, each process node having semantic mapping information, and the target process node being the process node whose semantic mapping information matches the query request; determining retrieval constraints for the query request based on the target process node; and retrieving a knowledge base using the retrieval constraints and the query request to obtain retrieval results for the query request.
[0149] It should be noted that the collection, storage, use, processing, transmission, provision, disclosure, and application of user personal information in this disclosed technical solution comply with relevant laws and regulations, necessary confidentiality measures have been taken, and it does not violate public order and good morals. In this disclosed technical solution, user authorization or consent has been obtained before acquiring or collecting user personal information.
[0150] According to embodiments of this disclosure, program code for executing the computer programs provided in embodiments of this disclosure 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 execute entirely on a user's computing device, partially on a 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).
[0151] 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 disclosure. 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.
[0152] Those skilled in the art will understand that the features described in the various embodiments and / or claims of this disclosure can be combined and / or combined in various ways, even if such combinations or combinations are not explicitly described in this disclosure. In particular, the features described in the various embodiments and / or claims of this disclosure can be combined and / or combined in various ways without departing from the spirit and teachings of this disclosure. All such combinations and / or combinations fall within the scope of this disclosure.
[0153] The embodiments of this disclosure have been described above. However, these embodiments are for illustrative purposes only and are not intended to limit the scope of this disclosure. Although various embodiments have been described above, this does not mean that the measures in the various embodiments cannot be used advantageously in combination. The scope of this disclosure is defined by the appended claims and their equivalents. Various substitutions and modifications can be made by those skilled in the art without departing from the scope of this disclosure, and all such substitutions and modifications should fall within the scope of this disclosure.
Claims
1. A knowledge query enhancement method based on process semantics, comprising: In response to receiving a query request, a target process node matching the query request is determined in a process semantic information set. The process semantic information set includes at least one process node included in at least one business domain and the association between process nodes. Each process node has semantic mapping information. The target process node is the process node whose semantic mapping information matches the query request. Based on the target process node, determine the retrieval constraints for the query request; The knowledge base is searched using the search constraints and the query request to obtain the search results.
2. The method according to claim 1, wherein determining the target process node matching the query request in the process semantic information set includes: Determine the semantic element features of the query request; In the set of process semantic information, candidate process nodes that match the semantic element features are identified; The target process node is determined based on the candidate process nodes and the query request.
3. The method according to claim 2, wherein determining the target process node based on the candidate process node and the query request comprises: Obtain the allowed access level of the object that initiated the query request; The matching score between the query request and each candidate process node is determined. The matching score is determined based on at least one of the following: the text matching score between the query request and each candidate node, the vector similarity between the query request and each candidate node, the matching degree of the semantic mapping information between the query request and each candidate node, the adaptability between the allowed access level and each candidate process node, and the distance between the process node corresponding to the context information of the query request and each candidate process node. Candidate process nodes whose matching scores meet the score requirements are selected as the target process nodes.
4. The method according to any one of claims 1 to 3, wherein the process node has hierarchical information; the step of determining the retrieval constraints of the query request based on the target process node includes: Obtain user profile information of the object that initiated the query request. The user profile information includes at least one of the following: hierarchical information that the object is allowed to access, the object's semantic completion strategy, the object's knowledge data preference type, and the object's output content constraints. Based on the user profile information and the target process node, the retrieval constraints of the query request are determined.
5. The method according to claim 4, wherein determining the retrieval constraints of the query request based on the user profile information and the target process node includes: The target process nodes are filtered using the user profile information to obtain filtered target process nodes, and the filtered target process nodes are matched with the user profile information. Based on the filtered target process nodes and the nodes in the process semantic information set that satisfy the association conditions with the filtered target process nodes, the retrieval constraints are generated. The retrieval constraints are used to constrain the acquisition of knowledge data in the knowledge base that matches the filtered target process nodes or the nodes that satisfy the association conditions.
6. The method according to claim 5, wherein the process node further has domain information, the process semantic information set further includes rule constraint nodes, the association between process nodes in the process semantic information set is an execution order association, a hierarchical inclusion association, or a cross-domain dependency association, and the process node and the rule constraint node have a constraint association; Also includes: Process nodes that have execution order association, hierarchical inclusion association, or cross-domain dependency association with the target process node and that meet the distance requirements between nodes are considered as nodes that satisfy the association conditions of the target process node.
7. The method according to claim 1, wherein the step of retrieving the knowledge base using the retrieval constraints and the query request to obtain the retrieval result of the query request includes: The query request is adjusted using the retrieval constraints to obtain an adjusted query request, which includes the knowledge retrieval scope generated based on the retrieval constraints. According to the knowledge retrieval scope of the adjusted query request, retrieve knowledge data in the knowledge base that falls within the knowledge retrieval scope.
8. The method according to claim 1, wherein the process semantic information set is determined in the following manner: Based on the business processing flow of each of the multiple business domains, obtain the process nodes, business action nodes, participating role nodes, system nodes and rule constraint nodes associated with the operation for each business processing flow, as well as the associations between the nodes, and generate the process semantic information set. Semantic normalization is performed on multiple process nodes that represent the same semantics to generate a process node and semantic mapping information including the process node.
9. The method according to claim 8, further comprising: In response to the existence of cross-domain execution dependencies between process nodes in different business domains, cross-domain dependency associations are established between process nodes in different business domains that have cross-domain execution dependencies in the process semantic information set.
10. An electronic device, comprising: Storage device for storing one or more programs; One or more processors are used to run the program to perform the following operations: In response to receiving a query request, a target process node matching the query request is determined in a process semantic information set. The process semantic information set includes at least one process node included in at least one business domain and the association between process nodes. Each process node has semantic mapping information. The target process node is the process node whose semantic mapping information matches the query request. Based on the target process node, determine the retrieval constraints for the query request; The knowledge base is searched using the search constraints and the query request to obtain the search results.
