Data processing method and device based on atlas retrieval

Abstract

The embodiment of the invention provides a data processing method and device based on atlas retrieval, and the method comprises the steps: carrying out the question type recognition of an input question in a process of carrying out the answer generation of a question inputted by an access user in a sub-service of a resource service, and obtaining a question type, the method comprises the following steps: extracting question characteristics of a question, inputting a graph retrieval model corresponding to a question type, and on the basis of a sub-graph obtained by performing entity selection and graph editing according to a knowledge graph of resource service in advance, calling a graph retrieval interface through the graph retrieval model to retrieve in the sub-graph, and an answer is generated according to a retrieval result returned by interface calling, so that the question answer is generated on the basis of map retrieval.

Description

Technical Field

[0001] This document relates to the field of data processing technologies, and in particular, to a data processing method and apparatus based on graph retrieval. Background Art

[0002] With the continuous development and popularization of the Internet, the application scope of various online services provided based on the Internet is getting wider and wider. And with the continuous increase in the number of users participating in online services, the user data generated by online services is also increasing. In this case, when the service provider of online services makes service decisions based on user data, due to the large amount of user data, it takes a lot of manpower and a long time cost to analyze user data, which brings great trouble to the service provider. Summary of the Invention

[0003] One or more embodiments of this specification provide a data processing method based on graph retrieval, including: identifying the problem type of the problem input by the accessing user in the sub-service of the resource service to obtain the problem type. Extracting the problem features of the problem and inputting them into the graph retrieval model corresponding to the problem type. Invoking the graph retrieval interface through the graph retrieval model to perform a search in the sub-graph of the sub-service, and generating an answer according to the retrieval result returned by the interface call. The sub-graph is obtained by performing entity selection and graph editing on the knowledge graph of the resource service. The knowledge graph is obtained by performing data structure conversion on the user service data of the resource service.

[0004] One or more embodiments of this specification provide a data processing apparatus based on graph retrieval, including: a problem identification module configured to identify the problem type of the problem input by the accessing user in the sub-service of the resource service to obtain the problem type. A problem input module configured to extract the problem features of the problem and input them into the graph retrieval model corresponding to the problem type. A graph retrieval module configured to invoke the graph retrieval interface through the graph retrieval model to perform a search in the sub-graph of the sub-service, and generate an answer according to the retrieval result returned by the interface call. The sub-graph is obtained by performing entity selection and graph editing on the knowledge graph of the resource service. The knowledge graph is obtained by performing data structure conversion on the user service data of the resource service.

[0005] One or more embodiments of this specification provide a data processing device based on graph retrieval, including: a processor; and a memory configured to store computer-executable instructions, which, when executed, cause the processor to: identify the type of a question entered by an accessing user in a sub-service of a resource service to obtain the question type. Extract the question features of the question and input them into a graph retrieval model corresponding to the question type. Call a graph retrieval interface through the graph retrieval model to perform a search in a sub-graph of the sub-service, and generate an answer based on the search result returned by the interface call. The sub-graph is obtained by performing entity selection and graph editing on the knowledge graph of the resource service. The knowledge graph is obtained by performing data structure conversion on the user service data of the resource service.

[0006] One or more embodiments of this specification provide a computer-readable storage medium for storing computer-executable instructions, which, when executed, implement the following process: identify the type of a question entered by an accessing user in a sub-service of a resource service to obtain the question type. Extract the question features of the question and input them into a graph retrieval model corresponding to the question type. Call a graph retrieval interface through the graph retrieval model to perform a search in a sub-graph of the sub-service, and generate an answer based on the search result returned by the interface call. The sub-graph is obtained by performing entity selection and graph editing on the knowledge graph of the resource service. The knowledge graph is obtained by performing data structure conversion on the user service data of the resource service. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] In order to more clearly illustrate the technical solutions in one or more embodiments of this specification or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments recorded in this specification. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings. Figure 1 It is a schematic diagram of an implementation environment of a data processing method based on graph retrieval provided by one or more embodiments of this specification; Figure 2 It is a processing flow chart of a data processing method based on graph retrieval provided by one or more embodiments of this specification; Figure 3 It is a processing flow chart of a data processing method based on graph retrieval applied to a resource service scenario provided by one or more embodiments of this specification; Figure 4 It is a schematic diagram of an embodiment of a data processing device based on graph retrieval provided by one or more embodiments of this specification; Figure 5 Schematic structural diagram of a data processing device based on graph retrieval provided for one or more embodiments of this specification. Specific implementation manners

[0008] In order to enable those skilled in the art to better understand the technical solutions in one or more embodiments of this specification, the following will clearly and completely describe the technical solutions in one or more embodiments of this specification with reference to the accompanying drawings in one or more embodiments of this specification. Obviously, the described embodiments are only a part of the embodiments of this specification, rather than all the embodiments. Based on one or more embodiments of this specification, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of this document.

[0009] The graph retrieval-based data processing method provided by one or more embodiments of this specification is applicable to the implementation environment of a data management engine. Refer to Figure 1 , this implementation environment at least includes: A graph retrieval model 101 and a database system 102; Among them, the database system 102 is used for data storage, such as storing the knowledge graph of the resource service and the sub-graphs of the sub-services. At the same time, the database system 102 is also used to provide a graph retrieval interface for graph retrieval to the graph retrieval model 101; The graph retrieval model 101 is used to call the graph retrieval interface provided by the database system 102 to perform retrieval in the sub-graphs of the sub-services, and generate an answer to the question according to the retrieval results returned by the interface call; specifically, the graph retrieval model 101 may specifically include an entity retrieval model 101-1 and / or a data retrieval model 101-2. Among them, the entity retrieval model 101-1 is used to perform entity and / or entity relationship retrieval in the sub-graph, and the data retrieval model 101-2 is used to perform instance data retrieval in the sub-graph; In addition, this implementation environment may further include a data conversion sub-engine 103, which is used to perform data structure conversion on the user service data of the resource service to obtain the knowledge graph of the resource service; and this implementation environment may further include a graph editing module 104. The configurator of the sub-service can perform entity selection and graph editing on the basis of the knowledge graph of the resource service through the graph editing module 104 to obtain the sub-graph of this sub-service.

[0010] In this implementation environment, based on the data structure conversion of the user service data of the resource service by the data conversion sub-engine 103 to obtain the knowledge graph of the resource service, and based on the entity selection and graph editing in the knowledge graph of the resource service by the graph editing module 104 to obtain the sub-graph of the sub-service, for the question input by the accessing user in the sub-service, the question type is obtained by identifying the question type of the input question. Further, the question features of the question are extracted and input into the graph retrieval model 101 corresponding to the question type. The graph retrieval model 101 calls the graph retrieval interface provided by the database system 102 to perform a search in the sub-graph of the sub-service, and an answer is generated according to the search result returned by the interface call. Thus, based on the graph retrieval in the sub-graph of the sub-service, the question answer of the sub-service is generated.

[0011] It should be noted that considering that the user service data involved in this specification may, to a certain extent, belong to the privacy of the applying user. Therefore, if you want to collect relevant data such as the user service data of the user, you can obtain the authorization of the user before collecting the data to make the operation of collecting data comply with relevant data management regulations. For example, data authorization can be carried out during the process of the user accessing the resource service, or data authorization can also be carried out during the process of the applying user's first access to the resource service. The specific method of data authorization can be to send a user data authorization reminder to the user. After the user confirms the reminder through an instruction, the data collection authorization can be obtained. Or, the method of data authorization can also be to obtain the data collection authorization by signing a data authorization agreement.

[0012] One or more embodiments of a data processing method based on graph retrieval provided in this specification are as follows: Refer to Figure 2 In this embodiment, the data processing method based on graph retrieval is provided, and the method specifically includes steps S202 to S206.

[0013] Step S202: Identify the question type of the question input by the accessing user in the sub-service of the resource service to obtain the question type.

[0014] The resource service described in this embodiment refers to a service related to resources. For example, a resource service provided by an application program for performing related processing such as recommendation and transaction of resource objects. The resource object can be an equity resource, such as tradable asset varieties (funds, stocks, bonds), etc., or a capital resource or a virtual resource. The virtual resource can be a virtual resource such as carbon emission reduction amount and points.

[0015] The sub-service can be any one or more of the multiple services provided by the resource service. For example, the sub-service for resource trading among the multiple services provided by the resource service. The accessing user can access the sub-service through a sub-program within the application, such as the sub-program for accessing the sub-service within the application. In addition, the sub-service can also be any one or more of the services for multiple resources provided by the resource service. For example, the sub-service for a certain rights and interests resource among the services for multiple resources provided by the resource service.

[0016] In specific implementation, during the process of the accessing user accessing the resource service, the accessing user can ask questions regarding the currently accessed sub-service. In this embodiment, responses are made to the questions of the accessing user to answer the problems encountered by the accessing user during the sub-service access process, so as to improve the access experience of the accessing user in the resource service. Specifically, the accessing user can ask questions by inputting the questions. Here, for the questions input by the accessing user in the sub-service of the resource service, the question types are identified for the questions to obtain the question types, so as to be able to start from the question types and generate corresponding answers for the questions of different question types.

[0017] During the specific execution process, during the process of identifying the question types of the input questions, the question intention can be identified for the questions to obtain the question intention, and the question types are determined according to the question intention. Among them, the question types include entity question types and / or data question types; the entity question type refers to the questions input by the accessing user that are questions corresponding to the entities and / or entity relationships in the knowledge graph. For example, the question input by the accessing user is "How can I query the purchased rights and interests resources of the user?"; the data question type refers to the questions input by the accessing user that are questions corresponding to the instance data in the knowledge graph. For example, the question input by the accessing user is "What is the specific amount of the xx rights and interests resources purchased by the user?"

[0018] Specifically, in an optional implementation manner provided by this embodiment, identifying the question types of the questions input by the accessing user in the sub-service of the resource service to obtain the question types includes: performing semantic recognition on the questions through an identification model, and performing intention recognition according to the semantic recognition result to obtain the question intention; determining the question types based on the question intention.

[0019] Among them, the recognition model can be a model for natural language semantic recognition obtained through pre-training. For example, the recognition model can be a model using a neural network architecture with a large number of parameters. In the process of determining the question type based on the question intention, the question type mapped by the question intention can be determined according to the pre-established mapping relationship between the question intention and the question type. Additionally, a type determination model obtained through pre-training can be used. By inputting the question intention into the type determination model, the question type can be output. For example, the question intention can be input into a binary classification model for classifying two question types: entity question type and data question type, and the entity question type or the data question type can be output.

[0020] Step S204: Extract the question features of the question and input them into the graph retrieval model corresponding to the question type.

[0021] As described above, the question type includes entity question type and / or data question type. To make the generation of answers to questions input by accessing users more accurate, corresponding graph retrieval models are set for different question types respectively. That is, the question type can correspond to the graph retrieval model. The graph retrieval model refers to a model used for graph retrieval in the knowledge graph and generating question answers based on the retrieval results. The graph retrieval model can be deployed in the data engine. Specifically, the graph retrieval model includes an entity retrieval model and / or a data retrieval model. Among them, the entity retrieval model is used for retrieving entities and / or entity relationships in the knowledge graph; the data retrieval model retrieves instance data in the knowledge graph.

[0022] To improve the efficiency of graph retrieval and also to enhance the service differentiation and personalization of different sub-services of the resource service, the graph retrieval model can specifically be a graph retrieval model for graph retrieval in the knowledge graph of the sub-service. The knowledge graph of the sub-service can be composed of local graphs in the knowledge graph of the resource service, or can be composed of local graphs in the knowledge graph of the resource service and graphs configured by the configurators of the sub-service. Or, it can also be composed of graphs configured by the configurators of the sub-service. In this case, the knowledge graph of the sub-service is called a sub-knowledge graph (sub-graph). Correspondingly, the graph retrieval model includes: an entity retrieval model for retrieving entities and / or entity relationships in the sub-graph of the sub-service, and / or a data retrieval model for retrieving instance data in the sub-graph of the sub-service. The entity retrieval model and / or the data retrieval model can also be deployed in the data engine.

[0023] In specific implementation, based on the above-mentioned recognition of the question type of the question input by the accessing user in the sub-service to obtain the question type, in this step, the question features of the question are extracted, and the extracted question features are input into the graph retrieval model corresponding to the question type. Thus, the graph retrieval interface is called through the graph retrieval model to perform retrieval in the sub-graph of the sub-service, and answers are generated according to the retrieval results returned by the interface call.

[0024] In this embodiment, the sub-graph of the sub-service is obtained by entity selection and graph editing in the knowledge graph of the resource service. The knowledge graph of the resource service can be obtained by performing data structure conversion on the user service data of the resource service. The following specifically describes the acquisition process of the knowledge graph of the resource service and the acquisition process of the sub-graph of the sub-service.

[0025] (1) Acquisition process of the knowledge graph of the resource service In the process of obtaining the knowledge graph of the resource service by performing data structure conversion on the user service data of the resource service, the knowledge graph can be constructed according to the data structure object obtained by parsing the data structure of the user service data. Specifically, in an optional implementation manner provided in this embodiment, the knowledge graph of the resource service is obtained in the following manner: the data structure of the user service data is parsed by a large language model to obtain a data structure object; the knowledge graph of the resource service is obtained based on the data structure object for graph construction. Among them, the large language model can adopt an open-source large language model (Large Language Model, LLM), or a large language model obtained by fine-tuning the open-source large language model, or a large language model obtained by constructing and training a large language model.

[0026] Optionally, the user service data includes: structured service data and / or unstructured service data of each sub-service of the resource service; the structured service data refers to user service data that can be represented by a specific data structure, where the structured service data may include service data tables; The unstructured data refers to user service data that cannot be represented by a data structure or has not been represented by a data structure. Specifically, the unstructured data may be user service data that may be different for each accessing user, such as text data custom-input by each accessing user, or event data related to the value change of the rights and interests resources, or industry data related to the transaction or management of the rights and interests resources.

[0027] Based on the structured service data and unstructured service data of user service data, data structure conversion is performed on the structured service data and unstructured service data to obtain the knowledge graph of resource services, that is: in the process of constructing the knowledge graph based on the structured service data and unstructured service data, by fusing the structured service data and unstructured service data, the effective connection of the structured service data and unstructured service data is realized with the help of the knowledge graph. Specifically, the structured service data and unstructured service data are realized with the help of the knowledge graph, so that the retrieval can be carried out in two dimensions of the structured service data and unstructured service data during the graph retrieval process, improving the data richness of the knowledge graph, and the retrieval results of graph retrieval based on the knowledge graph are also more accurate and comprehensive.

[0028] Specifically, in the process of performing data structure conversion on user service data, for the service data table in user service data, since the service data table often stores a large amount of data and the data structure is relatively clear, the triple elements (entities, relationships, and objects) of the service data table can be parsed, and the knowledge graph can be constructed according to the parsed triple elements.

[0029] Specifically, in an optional implementation manner provided in this embodiment, the data structure conversion includes: Perform data structure parsing on the service data table recording user service data to obtain entities, relationships, and / or objects; Construct graph nodes based on the obtained entities and / or objects, and construct edges of graph nodes based on relationships to obtain a knowledge graph.

[0030] Among them, the data structure parsing of the service data table can be realized through a large language model. By using the large language model to perform data structure parsing on the service data table to obtain entities, relationships, and / or objects, the input service data table can also be parsed through a pre-trained data structure parsing model to output entities, relationships, and / or objects.

[0031] Specifically, the data structure conversion of the user service data of the resource service can be realized through the data conversion sub-engine deployed by the data engine. Specifically, the conversion sub-engine can call a large language model or a data structure parsing model to perform data structure conversion on the user service data of the resource service, and construct a knowledge graph according to the parsing output by the model.

[0032] After the construction of the knowledge graph for the above-mentioned resource service, in order to implement graph retrieval based on the knowledge graph, specifically to retrieve instance data of entities or edges in the knowledge graph, the instance data mapped by the indication graph of the resource service can also be imported into the knowledge graph. Optionally, the instance data mapped by the knowledge graph is imported into the knowledge graph by calling a data import interface; the data import interface can be provided by a database system. And after the instance data is imported into the knowledge graph, the knowledge graph obtained after the instance data import can be stored in the database system. Subsequently, during the process of retrieving in the knowledge graph, graph retrieval can be performed in the knowledge graph by calling the graph retrieval interface provided by the database system.

[0033] (2) Process of obtaining the sub-graph of the sub-service The sub-graph of the sub-service can be obtained by entity selection and / or graph editing in the knowledge graph of the resource service. The obtained sub-graph of the sub-service can be stored in the database system. Specifically, in an optional implementation manner provided in this embodiment, entity selection and graph editing include: Query the associated entities of the entity and the entity relationships of the associated entities in the knowledge graph according to the entities submitted by the configurator of the sub-service. Based on the entity, associated entities, and entity relationships, construct a knowledge graph, and edit the obtained knowledge graph according to the editing instructions submitted by the configurator to obtain a sub-graph.

[0034] In addition, during the process of constructing the sub-graph of the sub-service based on the knowledge graph of the resource service, the sub-knowledge graph corresponding to the sub-service in the knowledge graph can also be displayed according to the access request of the configurator of the sub-service, and the displayed sub-knowledge graph can be edited according to the editing instructions submitted by the configurator to obtain a sub-graph; among them, the sub-knowledge graph corresponding to the sub-service can be obtained by screening or pruning the knowledge graph of the resource service according to the service relationship between the sub-service and the resource service, or the sub-knowledge graph of the sub-service can also be obtained by screening or pruning the knowledge graph of the resource service according to the service data relationship between the sub-service and the resource service.

[0035] In the implementation method of obtaining the sub-graph of the sub-service based on the knowledge graph of the resource service provided above, since the user service data of the resource service contains the user service data of all sub-services of the resource service, therefore, the data connection between different sub-services of the resource service is established through the construction of the knowledge graph of the resource service. By querying the associated entities of the entity and the entity relationship of the associated entity in the knowledge graph, and constructing and editing the graph based on the entity, the associated entity and the entity relationship to obtain the sub-graph, the content of the sub-graph of the sub-service is enriched. Based on this, the comprehensiveness of the graph retrieval performed in the sub-graph of the sub-service; and, by providing the method of knowledge graph editing to the configurator of the sub-service, the flexibility of the sub-graph of the sub-service is improved, and the graph retrieval performed in the sub-graph of the sub-service can be more consistent with the actual service requirements of the sub-service.

[0036] It should be noted that the above provides the process of obtaining the knowledge graph of the resource service and the process of obtaining the sub-graph of the sub-service. In addition, a method similar to the process of obtaining the knowledge graph can also be used to construct the sub-graph of the sub-service. The specific construction process is similar to the construction process of the knowledge graph of the resource service above. The difference is that the user service data of the resource service is used to construct the knowledge graph of the resource service, and the user association data of the sub-service can be used to construct the sub-graph of the sub-service, or the user service data related to the sub-service can be queried in the user service data of the resource service. The sub-graph obtained by constructing according to the queried user service data has a higher data correlation with the overall resource service and other sub-services of the resource service. For example, the following method can be used to construct the sub-graph of the sub-service: parse the data structure of the user service data of the sub-service through a large language model to obtain a data structure object, and construct a sub-graph based on the data structure object; or parse the data structure of the service data table recording the user service data of the sub-service to obtain entities, relationships and / or objects, construct graph nodes based on the obtained entities and / or objects, and construct edges of the graph nodes based on the relationships to obtain a sub-graph.

[0037] Step S206, retrieve in the sub-graph of the sub-service through the graph retrieval model by calling the graph retrieval interface, and generate an answer according to the retrieval result returned by the interface call.

[0038] As described above, the sub-graph is obtained by entity selection and graph editing in the knowledge graph of the resource service, and the knowledge graph is obtained by data structure conversion of the user service data of the resource service. On this basis, after extracting the problem features of the problem and inputting them into the graph retrieval model corresponding to the problem type, retrieve in the sub-graph of the sub-service through the graph retrieval model by calling the graph retrieval interface, and generate an answer according to the retrieval result returned by the interface call.

[0039] In this embodiment, during the retrieval process in the sub-graph of the sub-service, different graph retrieval models perform retrieval in the sub-graph of the sub-service by calling different graph retrieval interfaces. When the above problem types correspond to the graph retrieval models, the graph retrieval models can correspond to the graph retrieval interfaces. Specifically, when the problem types include entity problem types and / or data problem types, and the graph retrieval models include entity retrieval models and / or data retrieval models, the graph retrieval interfaces can include entity retrieval interfaces and / or data retrieval interfaces. Among them, the entity problem type, the entity retrieval model, and the entity retrieval interface correspond to each other, and the data problem type, the data retrieval model, and the data retrieval interface correspond to each other.

[0040] During the specific execution process, when the sub-graph of the sub-service is stored in the database system, the graph retrieval interface can be provided by the database system. Specifically, the graph retrieval model calls the corresponding graph retrieval interface to perform graph retrieval in the sub-graph of the sub-service to obtain the retrieval result. In an optional implementation manner provided by this embodiment, calling the graph retrieval interface to perform retrieval in the sub-graph of the sub-service includes: calling the graph retrieval interface in the database system that matches the problem type to perform retrieval in the sub-graph data stored in the database system to obtain the problem retrieval data; correspondingly, generating an answer according to the retrieval result returned by the interface call includes: generating an answer according to the problem retrieval data to obtain the answer to the question.

[0041] When the graph retrieval model includes an entity retrieval model and the graph retrieval interface includes an entity retrieval interface corresponding to the entity retrieval model, in an optional implementation manner provided by this embodiment, the graph retrieval model calls the graph retrieval interface to perform retrieval in the sub-graph of the sub-service, and generates an answer according to the retrieval result returned by the interface call, including: the entity retrieval model calls the entity retrieval interface of the database system to perform entity retrieval and / or entity relationship retrieval in the sub-graph to obtain the entity retrieval result; generating an answer according to the entity retrieval result to obtain the answer to the question.

[0042] When the graph retrieval model includes a data retrieval model and the data retrieval interface includes a data retrieval interface corresponding to the data retrieval model, in an optional implementation manner provided by this embodiment, the graph retrieval model calls the graph retrieval interface to perform retrieval in the sub-graph of the sub-service, and generates an answer according to the retrieval result returned by the interface call, including: the data retrieval model calls the data retrieval interface of the database system to perform instance data retrieval in the instance data of the sub-graph to obtain the data retrieval result; generating an answer according to the data retrieval result to obtain the answer to the question.

[0043] In practical applications, during the process of generating a knowledge graph and answers for the questions input by accessing users, there may also be situations where the questions input by accessing users involve entities and data in the sub-graph of the sub-service. In view of this, during the process of retrieving in the sub-graph of the sub-service through the graph retrieval model by calling the graph retrieval interface and generating answers based on the retrieval results returned by the interface call, the entity retrieval model can also be used to call the entity retrieval interface of the database system to perform entity retrieval and / or entity relationship retrieval in the sub-graph to obtain entity retrieval results. Moreover, the data retrieval model can be used to call the data retrieval interface of the database system to perform instance data retrieval in the instance data of the sub-graph to obtain data retrieval results, and generate answers based on the entity retrieval results and the data retrieval results to obtain the answers to the questions.

[0044] In summary, the data processing method based on graph retrieval provided in this embodiment obtains the knowledge graph of the resource service by performing data structure conversion on the user service data of the resource service, and obtains the sub-graph of the sub-service by performing entity selection and graph editing on the knowledge graph of the resource service. In this way, the data connection between different sub-services of the resource service is established through the construction of the knowledge graph of the resource service, and at the same time, the flexibility of the sub-graph of the sub-service is improved through the method of graph editing. On this basis, during the process of generating answers for the questions input by the accessing user in the sub-service, the question type is obtained by identifying the question type of the input question, the question features of the question are extracted and input into the graph retrieval model corresponding to the question type, so as to retrieve in the sub-graph of the sub-service through the graph retrieval model by calling the graph retrieval interface, and generate answers based on the retrieval results returned by the interface call, so as to generate question answers based on graph retrieval, thereby improving the accuracy of question answer generation through more accurate and efficient graph retrieval. Specifically, during the construction process of the knowledge graph of the resource service, the knowledge graph is obtained by performing data structure conversion on the structured service data and the unstructured service data, and the effective connection between the structured service data and the unstructured service data is realized with the help of the knowledge graph, which improves the richness of the knowledge graph, thereby contributing to improving the accuracy of graph retrieval.

[0045] The following takes the application of the data processing method based on graph retrieval provided in this embodiment in the resource service scenario as an example, combined with Figure 3 , to further illustrate the data processing method based on graph retrieval provided in this embodiment. See Figure 3 , the data processing method based on graph retrieval applied to the resource service scenario specifically includes the following steps.

[0046] Step S302, identify the question type of the question input by the accessing user in the sub-service of the resource service to obtain the question type.

[0047] Step S304: Extract the problem features of the problem and input them into the graph retrieval model corresponding to the problem type.

[0048] If the graph retrieval model corresponding to the problem type is an entity retrieval model, execute the following steps S306 to S308; If the graph retrieval model corresponding to the problem type is a data retrieval model, execute the following steps S310 to S312.

[0049] Step S306: Invoke the entity retrieval interface of the database system through the entity retrieval model to perform entity retrieval in the sub-graph of the sub-service, and obtain the entity retrieval result.

[0050] Step S308: Generate an answer based on the entity retrieval result to obtain the answer to the problem.

[0051] Step S310: Invoke the data retrieval interface of the database system through the data retrieval model to perform instance data retrieval in the instance data of the sub-graph, and obtain the data retrieval result.

[0052] Step S312: Generate an answer based on the data retrieval result to obtain the answer to the problem.

[0053] It should be noted that any one step or any combination of multiple steps in steps S302 to S312 can be combined with any one step or any combination of multiple steps in the above steps S202 to S206 to form a new implementation method according to the needs of implementation and deployment; in addition, according to the actual deployment needs, any one or any combination of technical features in steps S302 to S312 can be combined with any one or more technical features provided in the above steps S202 to S206 to form a new implementation method; or, any one or any combination of technical features in steps S302 to S312 can also be replaced by any one or more technical features provided in the above steps S202 to S206 according to the actual deployment needs to form a new implementation method, which will not be elaborated here one by one.

[0054] An embodiment of a data processing device based on graph retrieval provided in this specification is as follows: In the above embodiment, a data processing method based on graph retrieval is provided. Correspondingly, a data processing device based on graph retrieval is also provided, which will be described below with reference to the drawings.

[0055] Refer to Figure 4 , which shows a schematic diagram of an embodiment of a data processing device based on graph retrieval provided in this embodiment.

[0056] Since the device embodiments correspond to the method embodiments, they are described relatively simply. For relevant parts, please refer to the corresponding descriptions of the method embodiments provided above. The device embodiments described below are merely illustrative.

[0057] This embodiment provides a data processing device based on graph retrieval. The device includes: A problem identification module 402, configured to identify the problem type of the problem input by the accessing user in the sub-service of the resource service, and obtain the problem type; A problem input module 404, configured to extract the problem features of the problem and input them into the graph retrieval model corresponding to the problem type; A graph retrieval module 406, configured to call a graph retrieval interface through the graph retrieval model to perform a search in the sub-graph of the sub-service, and generate an answer according to the search result returned by the interface call; the sub-graph is obtained by entity selection and graph editing in the knowledge graph of the resource service, and the knowledge graph is obtained by performing data structure conversion on the user service data of the resource service.

[0058] An embodiment of a data processing device based on graph retrieval provided in this specification is as follows: Corresponding to the above-described data processing method based on graph retrieval, based on the same technical concept, one or more embodiments of this specification also provide a data processing device based on graph retrieval. This data processing device based on graph retrieval is used to execute the above-provided data processing method based on graph retrieval. Figure 5 It is a schematic structural diagram of a data processing device based on graph retrieval provided by one or more embodiments of this specification.

[0059] A data processing device based on graph retrieval provided in this embodiment includes: Such as Figure 5As shown, data processing devices based on graph retrieval can vary significantly due to differences in configuration or performance. They can include one or more processors 501 and a memory 502. The memory 502 can store one or more stored application programs or data. Among them, the memory 502 can be short-term storage or persistent storage. The application programs stored in the memory 502 can include one or more modules (not shown in the figure), and each module can include a series of computer-executable instructions in the data processing device based on graph retrieval. Further, the processor 501 can be set to communicate with the memory 502 and execute a series of computer-executable instructions in the memory 502 on the data processing device based on graph retrieval. The data processing device based on graph retrieval can also include one or more power supplies 503, one or more wired or wireless network interfaces 504, one or more input / output interfaces 505, one or more keyboards 506, etc.

[0060] In a specific embodiment, the data processing device based on graph retrieval includes a memory and one or more programs. One or more of the programs are stored in the memory, and one or more of the programs can include one or more modules. Each module can include a series of computer-executable instructions in the data processing device based on graph retrieval and is configured to be executed by one or more processors. The one or more programs include the following computer-executable instructions: Identify the question type for the question entered by the accessing user in the sub-service of the resource service to obtain the question type; Extract the question features of the question and input them into the graph retrieval model corresponding to the question type; Call the graph retrieval interface through the graph retrieval model to perform a search in the sub-graph of the sub-service, and generate an answer according to the search result returned by the interface call; the sub-graph is obtained by entity selection and graph editing in the knowledge graph of the resource service, and the knowledge graph is obtained by performing data structure conversion on the user service data of the resource service.

[0061] An embodiment of a computer-readable storage medium provided in this specification is as follows: Corresponding to the above-described data processing method based on graph retrieval, based on the same technical concept, one or more embodiments of this specification also provide a computer-readable storage medium.

[0062] The computer-readable storage medium provided in this embodiment is used to store computer-executable instructions, and the computer-executable instructions, when executed, implement the following process: Identify the question type for the question entered by the accessing user in the sub-service of the resource service to obtain the question type; Extract the problem features of the problem and input them into the graph retrieval model corresponding to the problem type; Call the graph retrieval interface through the graph retrieval model to perform a search in the sub-graph of the sub-service, and generate an answer based on the search result returned by the interface call; the sub-graph is obtained by entity selection and graph editing in the knowledge graph of the resource service, and the knowledge graph is obtained by performing data structure conversion on the user service data of the resource service.

[0063] It should be noted that the embodiments of a computer-readable storage medium in this specification and the embodiments of a data processing method based on graph retrieval in this specification are based on the same inventive concept. Therefore, the specific implementation of this embodiment can refer to the implementation of the corresponding method described above, and the repeated parts will not be elaborated.

[0064] An embodiment of a computer program product provided in this specification is as follows: Corresponding to the data processing method based on graph retrieval described above, based on the same technical concept, one or more embodiments of this specification also provide a computer program product.

[0065] A computer program product includes a computer program / instructions, and when the computer program / instructions are executed by a processor, the following steps are implemented: Identify the problem type of the problem input by the accessing user in the sub-service of the resource service to obtain the problem type; Extract the problem features of the problem and input them into the graph retrieval model corresponding to the problem type; Call the graph retrieval interface through the graph retrieval model to perform a search in the sub-graph of the sub-service, and generate an answer based on the search result returned by the interface call; the sub-graph is obtained by entity selection and graph editing in the knowledge graph of the resource service, and the knowledge graph is obtained by performing data structure conversion on the user service data of the resource service.

[0066] It should be noted that the embodiments of a computer program product in this specification and the embodiments of a data processing method based on graph retrieval in this specification are based on the same inventive concept. Therefore, the specific implementation of this embodiment can refer to the implementation of the corresponding method described above, and the repeated parts will not be elaborated.

[0067] Each embodiment in this specification is described in a progressive manner. For the same or similar parts among the embodiments, reference can be made to each other. The key point of each embodiment is to illustrate the differences from other embodiments. For example, the apparatus embodiment, the device embodiment, the computer-readable storage medium embodiment, and the computer program product embodiment are all similar to the method embodiment, so the description is relatively simple. To read the relevant content in the apparatus embodiment, the device embodiment, the computer-readable storage medium embodiment, and the computer program product embodiment, please refer to the corresponding description in the method embodiment.

[0068] The specific embodiments of this specification have been described above. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than in the embodiments and still achieve the desired result. Additionally, the processes depicted in the figures do not necessarily require the particular order or sequential order shown to achieve the desired result. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0069] In the 1930s, it was obvious to distinguish whether an improvement to a technology was a hardware improvement (e.g., improvement to circuit structures such as diodes, transistors, switches, etc.) or a software improvement (improvement to method flows). However, with the development of technology, many method flow improvements today can be regarded as direct improvements to hardware circuit structures. Almost all designers obtain the corresponding hardware circuit structure by programming the improved method flow into the hardware circuit. Therefore, it cannot be said that an improvement to a method flow cannot be implemented with a hardware entity module. For example, a Programmable Logic Device (PLD) (such as a Field Programmable Gate Array (FPGA)) is such an integrated circuit whose logic function is determined by the user's programming of the device. The designer can program by themselves to "integrate" a digital system on a single PLD, without the need to ask a chip manufacturer to design and fabricate a dedicated integrated circuit chip. Moreover, nowadays, instead of manually fabricating integrated circuit chips, this programming is mostly implemented using "logic compiler" software, which is similar to the software compiler used in program development and writing. The original code before compilation also has to be written in a specific programming language, which is called a Hardware Description Language (HDL). And there is not only one type of HDL, but many types, such as ABEL (Advanced Boolean Expression Language), AHDL (Altera Hardware Description Language), Confluence, CUPL (Cornell University Programming Language), HDCal, JHDL (Java Hardware Description Language), Lava, Lola, MyHDL, PALASM, RHDL (Ruby Hardware Description Language), etc. Currently, the most commonly used ones are VHDL (Very-High-Speed Integrated Circuit Hardware Description Language) and Verilog. Those skilled in the art should also be aware that by simply making a little logical programming of the method flow with the above-mentioned several hardware description languages and programming it into the integrated circuit, it is easy to obtain the hardware circuit that implements the logical method flow.

[0070] The controller can be implemented in any suitable manner. For example, the controller can take the form of, for example, a microprocessor or a processor and a computer-readable medium storing computer-readable program code (such as software or firmware) executable by the (micro)processor, logic gates, switches, an application specific integrated circuit (ASIC), a programmable logic controller, and an embedded microcontroller. Examples of the controller include, but are not limited to, the following microcontrollers: ARC 625D, Atmel AT91SAM, Microchip PIC18F26K20, and Silicone Labs C8051F320. The memory controller can also be implemented as part of the control logic of the memory. Those skilled in the art also know that in addition to implementing the controller in the form of pure computer-readable program code, it is entirely possible to logically program the method steps to enable the controller to be implemented in the form of logic gates, switches, application specific integrated circuits, programmable logic controllers, embedded microcontrollers, etc. to achieve the same function. Therefore, such a controller can be considered a hardware component, and the devices included therein for implementing various functions can also be regarded as the structures within the hardware component. Or even, the devices for implementing various functions can be regarded as either software modules for implementing the method or structures within the hardware component.

[0071] The systems, devices, modules, or units illustrated in the above embodiments can be specifically implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, the computer can be, for example, a personal computer, a laptop computer, a cellular phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or any combination of these devices.

[0072] For the convenience of description, when describing the above devices, they are described separately as various units according to their functions. Of course, when implementing the embodiments of this specification, the functions of each unit can be implemented in the same or multiple software and / or hardware.

[0073] Those skilled in the art should understand that one or more embodiments of this specification can be provided as a method, a system, or a computer program product. Therefore, one or more embodiments of this specification can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, this specification can take the form of a computer program product implemented on one or more computer-readable storage media (including but not limited to disk memories, CD-ROMs, optical memories, etc.) containing computer-usable program code.

[0074] This specification is described with reference to the flowcharts and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the specification. It should be understood that each flow and / or block in the flowchart and / or block diagram, and combinations of flows and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processors of general-purpose computers, special-purpose computers, embedded processors, or other programmable data processing devices based on atlas retrieval to produce a machine, such that the instructions executed by the processors of the computer or other programmable data processing devices based on atlas retrieval produce means for implementing the functions specified in the Figure 1 one or more flows and / or blocks Figure 1 means for implementing the functions specified in one or more blocks or multiple blocks.

[0075] These computer program instructions can also be stored in a computer-readable memory capable of guiding a computer or other programmable data processing device based on atlas retrieval to work in a specific manner, such that the instructions stored in the computer-readable memory produce a manufactured article including instruction means that implement the functions specified in the Figure 1 one or more flows and / or blocks Figure 1 means for implementing the functions specified in one or more blocks or multiple blocks.

[0076] These computer program instructions can also be loaded onto a computer or other programmable data processing device based on atlas retrieval, such that a series of operation steps are executed on the computer or other programmable device to produce a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in the Figure 1 one or more flows and / or blocks Figure 1 means for implementing the functions specified in one or more blocks or multiple blocks.

[0077] In a typical configuration, a computing device includes one or more processors (CPUs), an input / output interface, a network interface, and memory.

[0078] Memory may include non-permanent memory in the form of computer-readable media, random access memory (RAM), and / or non-volatile memory such as read-only memory (ROM) or flash memory (flash RAM). Memory is an example of computer-readable media.

[0079] A computer-readable medium includes permanent and non-permanent, removable and non-removable media that can implement information storage by any method or technology. The information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer-readable storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassette tapes, disk storage or other magnetic storage devices, or any other non-transitory medium that can be used to store information accessible by a computing device. As defined herein, a computer-readable medium does not include transitory computer-readable media, such as modulated data signals and carrier waves.

[0080] It should also be noted that the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, such that a process, method, commodity or device comprising a series of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, commodity or device. Without further limitation, an element defined by the statement "comprising at least one..." does not exclude the presence of additional identical elements in the process, method, commodity or device comprising the element.

[0081] One or more embodiments of this specification can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. One or more embodiments of this specification can also be practiced in a distributed computing environment, where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0082] The above are only embodiments of this document and are not used to limit this document. For those skilled in the art, this document can have various changes and modifications. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of this document should be included within the scope of the claims of this document.

Claims

1. A data processing method based on graph retrieval, comprising: Identifying the problem type of the problem input by the accessing user in the sub-service of the resource service to obtain the problem type; Extracting the problem features of the problem and inputting them into the graph retrieval model corresponding to the problem type; Invoking the graph retrieval interface through the graph retrieval model to perform a search in the sub-graph of the sub-service, and generating an answer according to the retrieval result returned by the interface call; The sub-graph is obtained by entity selection and graph editing in the knowledge graph of the resource service, and the knowledge graph is obtained by performing data structure conversion on the user service data of the resource service.

2. The data processing method based on graph retrieval according to claim 1, wherein the knowledge graph is obtained in the following manner: Performing data structure parsing on the user service data through a large language model to obtain a data structure object; Constructing a graph based on the data structure object to obtain the knowledge graph.

3. The data processing method based on graph retrieval according to claim 1, wherein the invoking the graph retrieval interface to perform a search in the sub-graph of the sub-service includes: Invoking the graph retrieval interface in the database system that matches the problem type to perform a search in the sub-graph data stored in the database system to obtain problem retrieval data; Correspondingly, the generating an answer according to the retrieval result returned by the interface call includes: generating an answer according to the problem retrieval data to obtain the answer to the problem.

4. The data processing method based on atlas retrieval according to claim 1, wherein the user service data includes: The structured service data and / or unstructured service data of each sub-service of the resource service; Wherein, the structured service data includes service data tables.

5. The data processing method based on graph retrieval according to claim 1, after the knowledge graph is constructed, the instance data mapped by the knowledge graph is imported into the knowledge graph by invoking the data import interface; The data import interface and the graph retrieval interface are provided by the database system.

6. The data processing method based on graph retrieval according to claim 1, wherein the entity selection and graph editing include: Querying the associated entities of the entity and the entity relationships of the associated entities in the knowledge graph according to the entities submitted by the configurator of the sub-service; Constructing a graph based on the entity, the associated entity and the entity relationship, and editing the constructed knowledge graph according to the editing instructions submitted by the configurator to obtain the sub-graph.

7. The data processing method based on atlas retrieval according to claim 1, wherein the atlas retrieval model includes: An entity retrieval model for performing entity and / or entity relationship retrieval in the sub-graph, and a data retrieval model for performing instance data retrieval in the sub-graph.

8. The data processing method based on graph retrieval according to claim 7, wherein the invoking the graph retrieval interface through the graph retrieval model to perform a search in the sub-graph of the sub-service and generating an answer according to the retrieval result returned by the interface call includes: Invoking the entity retrieval interface of the database system through the entity retrieval model to perform entity retrieval and / or entity relationship retrieval in the sub-graph to obtain an entity retrieval result; Generating an answer according to the entity retrieval result to obtain the answer to the problem.

9. The data processing method based on graph retrieval according to claim 7, wherein retrieving in the sub-graph of the sub-service through the graph retrieval model by invoking the graph retrieval interface, and generating an answer according to the retrieval result returned by the interface call, includes: Invoking the data retrieval interface of the database system through the data retrieval model to perform instance data retrieval in the instance data of the sub-graph, and obtaining a data retrieval result; Generating an answer according to the data retrieval result to obtain the answer to the question.

10. The data processing method based on graph retrieval according to claim 1, wherein identifying the type of the question input by the accessing user in the sub-service of the resource service, and obtaining the question type, includes: Performing semantic recognition on the question through an identification model, and performing intention recognition according to the semantic recognition result to obtain the question intention; Determining the question type based on the question intention; The question type corresponds to the graph retrieval model and the graph retrieval interface.

11. The data processing method based on graph retrieval according to claim 1, wherein the data structure conversion includes: Performing data structure parsing on the service data table recording user service data to obtain entities, relationships, and objects; Constructing graph nodes based on the obtained entities and objects, and constructing edges of the graph nodes based on the relationships, to obtain the knowledge graph.

12. The data processing method based on graph retrieval according to claim 7, wherein the entity retrieval model and the data retrieval model are deployed in a data engine; The data structure conversion of the user service data of the resource service is implemented through a data conversion sub-engine deployed by the data engine.

13. A data processing apparatus based on graph retrieval, including: A question identification module, configured to identify the type of a question input by an accessing user in the sub-service of a resource service, and obtain the question type; A question input module, configured to extract the question features of the question and input them into the graph retrieval model corresponding to the question type; A graph retrieval module, configured to retrieve in the sub-graph of the sub-service through the graph retrieval model by invoking the graph retrieval interface, and generate an answer according to the retrieval result returned by the interface call; the sub-graph is obtained by performing entity selection and graph editing on the knowledge graph of the resource service, and the knowledge graph is obtained by performing data structure conversion on the user service data of the resource service.

14. A data processing device based on graph retrieval, including: A processor; And a memory configured to store computer-executable instructions, where the computer-executable instructions, when executed, cause the processor to: Identify the type of a question input by an accessing user in the sub-service of a resource service, and obtain the question type; Extract the question features of the question and input them into the graph retrieval model corresponding to the question type; Retrieve in the sub-graph of the sub-service through the graph retrieval model by invoking the graph retrieval interface, and generate an answer according to the retrieval result returned by the interface call; The sub-graph is obtained by entity selection and graph editing in the knowledge graph of the resource service, and the knowledge graph is obtained by performing data structure conversion on the user service data of the resource service.

15. A computer-readable storage medium for storing computer-executable instructions, which implement the steps of the method according to claim 1 when executed.

Images