Data space-oriented data retrieval enhancement generation method, device and equipment

Abstract

The invention provides a data space-oriented data retrieval enhancement generation method, device and equipment, relates to the technical field of data generation, and aims to improve the accuracy of a retrieval task in a data space of a data network. The method is applied to a data space retrieval enhancement generation system of the Internet of Things, the system comprises a proxy node and a plurality of nodes, and the method comprises the following steps: receiving a query request in a natural language form through the proxy node, and determining a query intention according to the query request; the query intention is matched with the query capability of each node, a plurality of candidate nodes are determined according to the matching degree, the nodes represent data sources in a network data space, and the query capability represents query contents and support fields provided by the nodes; sending the query request to the plurality of candidate nodes for retrieval generation to obtain query results of the plurality of candidate nodes; and the proxy node generates a final query result according to the query results of the plurality of candidate nodes.

Description

Technical Field

[0001] The present application relates to the technical field of data generation, and in particular to a data retrieval enhancement generation method, device and equipment for data space. Background Art

[0002] In the data space built on the Internet of Things, data is multi-source heterogeneous, its storage is scattered and cannot be processed in a centralized manner; and there is a weak coupling relationship between data and applications, and users cannot predict the specific data content in the data space before searching. At present, the query of the data space mainly relies on keyword search, which has significant semantic limitations: first, keywords are difficult to accurately express user intentions, and the system needs to convert them into potential structured queries, but the conversion process is uncertain; second, the query results usually contain a large amount of irrelevant data, and the system needs to evaluate the results to filter out valid information. In addition, since the data space lacks a strict data model, the relationship between data needs to be established dynamically, resulting in the possibility that the query results are only suboptimal or approximate solutions.

[0003] The metadata stored in each node in the data space is publicly accessible, while entity data must be authorized to obtain and is prohibited from flowing freely across nodes. The relevant technology is to perform retrieval enhancement generation for a single node in the Internet of Things, that is, the user submits a metadata query request to a single node, and the node returns the metadata results it stores; if a global query is required, it is necessary to rely on the top-level node (the top-level node aggregates all metadata). However, the information contained in metadata is limited, and it is difficult to fully reflect the content of the data entity, especially when it involves multiple data from multiple sources, which will result in low query accuracy. Therefore, how to improve the accuracy of retrieval tasks in the Internet of Things data space is a technical problem that needs to be solved urgently. Summary of the invention

[0004] In view of the above problems, embodiments of the present application provide a data space-oriented data retrieval enhancement generation method, apparatus and device to overcome the above problems or at least partially solve the above problems.

[0005] In a first aspect of an embodiment of the present application, a data space-oriented data retrieval enhancement generation method is disclosed, which is applied to a data space retrieval enhancement generation system for a data network. The system includes a proxy node and multiple nodes. The method includes: Receiving a query request in natural language through a proxy node, and determining a query intent according to the query request; Matching the query intent with the query capability of each node, and determining multiple candidate nodes according to the matching degree, wherein the node represents the data source in the data space of the Internet of Things, and the query capability represents the query content and support field provided by the node; Sending the query request to the multiple candidate nodes for retrieval and generation to obtain query results of the multiple candidate nodes; The proxy node generates a final query result according to the query results of the multiple candidate nodes.

[0006] Optionally, the method further comprises: The node generates a capability summary of the node according to the retrieval enhancement scheme deployed by the node, wherein the capability summary describes the query content and supported fields provided by the node in a natural language form; Registering the capability summary and node basic information to the proxy node; Matching the query intent with the query capabilities of each node includes: The proxy node determines the query capability of each node according to the registered capability summary and the basic information of the node, and matches the query intent with the query capability of each of the nodes.

[0007] Optionally, the method further comprises: The node constructs a vector library according to the retrieval enhancement scheme deployed by the node, and the vector library represents the data space of the node; The query request is sent to the multiple candidate nodes for retrieval and generation, and query results of the multiple candidate nodes are obtained, including: The proxy node processes the query request into a target query request and sends it to the multiple candidate nodes, wherein the target query request includes a query description and keywords in a natural language form; Each candidate node performs retrieval generation based on the retrieval enhancement scheme deployed by itself, the target query request and the vector library to obtain query results of multiple candidate nodes.

[0008] Optionally, each candidate node performs retrieval generation based on the retrieval enhancement scheme deployed by itself, the target query request and the vector library to obtain query results of multiple candidate nodes, including: In the case where the retrieval enhancement scheme deployed by the candidate node itself has the retrieval enhancement generation capability, searching the vector library according to the query description of the target query request, and generating the query result of the candidate node according to the large model; In the case that the retrieval enhancement scheme deployed by the candidate node itself does not have the retrieval enhancement generation capability, metadata retrieval is performed in the vector library according to the keywords in the target query request to obtain the query result of the candidate node.

[0009] The query description of the target query request is searched in the vector library, and query results of candidate nodes are generated according to the large model, including: Searching the vector library according to the query description of the target query request to obtain a first target digital object, and generating a query result of a candidate node according to the first target digital object and an identifier of the first target digital object through the large model; Performing metadata retrieval in the vector library according to the keywords in the target query request to obtain query results for candidate nodes, including: Perform metadata retrieval in the vector library according to the keywords in the target query request to obtain a second target digital object, and obtain a query result of a candidate node according to the metadata of the second target digital object and the identifier of the second target digital object.

[0010] Optionally, the query intent is matched with the query capability of each node, and a plurality of candidate nodes are determined according to the matching degree, including: Matching the query intent with the query capability of each node, and sorting the nodes in descending order according to the matching degree to obtain a first sorting result; The first N nodes in the first sorting result are taken as candidate nodes, where N is a positive integer greater than 0.

[0011] Optionally, the proxy node generates a final query result according to the query results of the multiple candidate nodes, including: Matching the query result of each candidate node with the query request, and sorting the query results of the candidate nodes in descending order of matching degree to obtain a second sorting result; The query results of the first M candidate nodes in the second sorting result are used as the target query results, where M is a positive integer greater than 0; A final query result is generated according to the target query result.

[0012] Optionally, generating a final query result according to the target query result includes: A final query result is generated according to the target query result and the candidate nodes corresponding to the target query result.

[0013] In a second aspect of the embodiments of the present application, a data space-oriented data retrieval enhancement generation device is disclosed, the device comprising: An intention determination module, configured to receive a query request in natural language form and determine the query intention according to the query request; A node matching module is used to match the query intent with the query capability of each node, and determine multiple candidate nodes according to the matching degree, wherein the node represents the data source in the data space of the Internet of Things, and the query capability represents the query content and support field provided by the node; A search and generation module, used for sending the query request to the multiple candidate nodes for search and generation, and obtaining query results of the multiple candidate nodes; The query generation module is used to generate a final query result according to the query results of the multiple candidate nodes.

[0014] According to a third aspect of an embodiment of the present application, an electronic device is disclosed, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the data retrieval enhancement generation method for data space described in the first aspect of the embodiment of the present application are implemented.

[0015] According to a fourth aspect of an embodiment of the present application, a computer-readable storage medium is disclosed, on which a computer program is stored. When the computer program is executed by a processor, the steps of the data retrieval enhancement generation method for data space described in the first aspect of the embodiment of the present application are implemented.

[0016] According to a fifth aspect of an embodiment of the present application, a computer program product is disclosed, including a computer program, which, when executed by a processor, implements the steps of the data space-oriented data retrieval enhancement generation method described in the first aspect of an embodiment of the present application.

[0017] The embodiments of the present application include the following advantages: In an embodiment of the present application, a query request in natural language form is received by a proxy node and the query intent is parsed, which overcomes the semantic limitations of keyword retrieval, can more accurately capture user needs, and reduce query deviations caused by vague keyword expressions; the query intent is matched with the query capability of each node, and multiple candidate nodes are determined based on the matching degree to avoid blindly querying all nodes, reduce computing overhead, and improve retrieval efficiency; the query request is sent to multiple candidate nodes for retrieval generation, and query results of multiple candidate nodes are obtained, and the query results of multiple candidate nodes are integrated through the proxy node to generate the final query result, which can cover a wider range of data sources, reduce query omissions due to the data limitations of a single node, and improve the comprehensiveness and accuracy of the results.

[0018] In this way, by setting proxy nodes and nodes in the data space of the Internet of Things, two-level content generation is achieved. Since the final query result is generated by the proxy node based on the query results of multiple candidate nodes, the transmission of original data between nodes is avoided, the data security of each node in the data space of the Internet of Things is guaranteed, and the accuracy of the final query result is guaranteed. Therefore, a cross-node retrieval enhancement generation in the data space of the Internet of Things is realized, and the accuracy of the retrieval task in the data space of the Internet of Things is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the description of the embodiments of the present application will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0020] Figure 1 It is a schematic diagram of an existing digital network data space retrieval enhancement system; Figure 2 It is a schematic diagram of a digital network data space retrieval enhancement system provided by an embodiment of the present application; Figure 3 It is a flowchart of the steps of a data retrieval enhancement generation method for data space provided by an embodiment of the present application; Figure 4 It is a flowchart of the steps of another data space-oriented data retrieval enhancement generation method provided by an embodiment of the present application; Figure 5 It is a structural schematic diagram of a data retrieval enhancement generation device for data space provided in an embodiment of the present application; Figure 6 It is a structural schematic diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0021] In order to make the above-mentioned purposes, features and advantages of the present application more obvious and understandable, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0022] In order to better understand the technical solution of the present application, the technical concepts involved in the present application are explained.

[0023] Digital Internet: Based on software definition, various heterogeneous data platforms and systems are connected through a data-centric open software architecture and standardized interoperability protocols to form a "virtual / data" network on the "physical / machine" Internet. Under the architecture of the digital Internet, a data space of "data interconnection, on-demand scheduling, intra-domain autonomy, and inter-domain collaboration" can be formed.

[0024] Data space: Data space is a collection of data related to a subject and its relationships. The data in the data space comes from multiple different data sources and the data formats are diverse. At the same time, the relationships between data are complex, dynamic, and evolving. Data space can be constructed through data networking technology, in which data has a unified data model and supports a unified interoperability protocol.

[0025] Retrieval-Augmented Generation: Retrieval-Augmented Generation (RAG) refers to a technique that refers to private or authoritative knowledge bases other than training data to assist in the output of large models. Large Language Models (LLMs) are trained with massive amounts of data and use billions of parameters to generate raw output for tasks such as answering questions, translating languages, and completing sentences. Building on LLM's already powerful general task execution capabilities, RAG extends it to access internal knowledge bases in specific fields or organizations, all without retraining the model. Therefore, retrieval-augmented generation is a cost-effective way to improve LLM output, keeping it relevant, accurate, and useful in a variety of scenarios.

[0026] Digital object architecture: Digital object architecture is a typical technical route for the Internet of Things. Digital object architecture standardizes Internet data resources through digital objects, uses digital object interface protocol and digital object identification resolution protocol to standardize data interaction behavior, and forms an open software architecture based on the three core systems of digital objects to achieve the interconnection and interoperability of heterogeneous, remote and heterogeneous data, forming a data space with the characteristics of "data interconnection, on-demand scheduling, intra-domain autonomy and inter-domain collaboration". The three elements of digital objects respectively undertake some tasks: the identifier is the only network-wide unique identifier of the digital object, which plays a positioning role; the data entity plays the role of digitizing and persisting the original information of the data resource, and is the resource content actually referred to by the digital object; metadata is a collection of descriptive information about data items. Metadata is used in the actual retrieval process to retrieve digital objects under specific descriptive conditions.

[0027] Digital object architecture is a new data-centric software architecture system. For access to data objects, the most applicable solution currently is to use metadata as a search domain set, and perform some natural language matching with metadata names, descriptions, etc. through the user's actual query words and phrases. This matching method uses the word segmentation of the input and the metadata part field control model, and uses the search middleware to perform text-level matching to match predefined metadata content such as industries and businesses.

[0028] like Figure 1 As shown, Figure 1This is a schematic diagram of an existing data space retrieval enhancement system for the Internet of Things. The metadata stored in each node in the data space is publicly accessible, but entity data can only be accessed after authorization, and entity data cannot flow freely between nodes. The existing retrieval process: the user queries a node in the Internet of Things for metadata, and the node returns the query results based on the metadata stored in it; if global data needs to be queried, it is necessary to query through the top-level node (the top-level node aggregates all metadata). However, due to the limited information contained in the metadata, it is difficult to fully reflect the content of the data entity, especially when it involves multiple data from multiple sources, which will result in low accuracy of the query results.

[0029] The related technology is to generate retrieval enhancement for a single node in the digital network, without considering the data query scenario in the data space formed by multiple nodes. Therefore, although this single-node retrieval enhancement generation scheme based on text metadata matching can complete the task of using metadata for digital object retrieval, there are still some problems that can be improved: (1) The metadata information is limited, which limits the retrieval effect: Currently, users can add industry metadata, business metadata and other content to assist actual users in retrieval. This method has disadvantages: a. There may be cognitive differences between users who define metadata and users who use data retrieval functions, and users cannot use predefined metadata for matching retrieval well; b. Other methods for generating metadata can be supplemented. This method can only be manually input and lacks a metadata generation method that can be efficiently retrieved; c. Combined with the current development trend of natural language processing, the existing methods have the optimization point of introducing large models to improve overall performance. (2) Unable to handle cross-node retrieval enhancement problems: The current digital object retrieval enhancement generation method is only carried out for data in a single node. When it is necessary to use data from multiple nodes, the existing digital object retrieval enhancement generation method cannot work because data entities usually cannot be aggregated to a single node for processing.

[0030] Since metadata carries limited information and has strong domain characteristics, in order to further improve the accuracy of queries, it is necessary to consider using entity data. With the emergence of big model technology, RAG technology has been proposed to improve the presentation of retrieval. However, there are certain problems in directly using RAG technology in Internet of Things scenarios. When multiple data sources are involved, this method requires aggregating multiple matching data into a big model, and then the big model generates content based on the matching data. Since entity data cannot usually be directly aggregated, this brings difficulties to the application of RAG technology.

[0031] In order to overcome the limitations of related technologies, an embodiment of the present application provides a data retrieval enhancement generation method for data space, which can solve the following problems: (1) Solving the problem of cross-node retrieval enhancement generation in data space. Since data is scattered among different subjects, each data subject has its own control requirements for the use of data. When performing retrieval enhancement generation, it is not possible to directly aggregate the relevant original data retrieved from the data controlled by each subject to a certain subject and then perform enhancement generation. The embodiment of the present application performs retrieval enhancement generation on the data in the networked data space without aggregating the original data, which can achieve efficient retrieval of data in the data space while protecting the user's effective control over their own data.

[0032] (2) Solve the problem of multi-level data tracing in the retrieval process. Since the data in the data space comes from multiple data sources and has multiple different types, the correlation between the data is weak, which makes it difficult to use the data. After the introduction of the retrieval enhancement generation function, due to the illusion problem of the large model, the returned results may contain content that is fabricated out of thin air by the large model. Therefore, in order to improve the credibility of the generated results, it is necessary to trace the source of the returned results. Since there are multiple levels of data generation processes in the generation process, the embodiment of the present application implements a multi-level data tracing method to ensure that the generated content has a basis.

[0033] The following describes the data space-oriented data retrieval enhancement generation method provided in the embodiment of the present application in conjunction with the accompanying drawings.

[0034] The data space-oriented data retrieval enhancement generation method provided in the embodiment of the present application is applied to the data space retrieval enhancement generation system of the Internet of Things, such as Figure 2 As shown, Figure 2 This is a schematic diagram of a digital network data space retrieval enhancement system provided by an embodiment of the present application. The system includes a proxy node and multiple nodes, wherein a large model is deployed in the proxy node and each node, the node represents the data source in the digital network data space, each node can generate a query result corresponding to the query request based on its own data, and the proxy node can generate a final query result based on the query result from the node. This method does not require the aggregation of entity data, and performs two-level content generation in the data space based on the proxy node and the node.

[0035] Reference Figure 3 As shown, Figure 3 is a flowchart of a data space-oriented data retrieval enhancement generation method provided by an embodiment of the present application, such as Figure 3 As shown, the data space-oriented data retrieval enhancement generation method may include steps S310 to S340: Step S310: receiving a query request in natural language form through a proxy node, and determining the query intent according to the query request.

[0036] In an embodiment of the present application, a user sends a query request in natural language to a proxy node, the proxy node receives the query request in natural language, and parses the query request through intent recognition to determine the query intent corresponding to the query request. In some embodiments, the proxy node can parse the query request using an intent recognition model, for example, the proxy node can parse the query request based on a BERT text classifier, where BERT refers to Bidirectional Encoder Representations from Transformers, a pre-trained natural language processing model based on the Transformer architecture.

[0037] Since the query request is provided in the form of natural language, it overcomes the semantic limitations of traditional keyword retrieval, can capture user needs more accurately, and reduce query deviations caused by ambiguous keyword expressions.

[0038] Step S320: Match the query intent with the query capability of each node, and determine multiple candidate nodes according to the matching degree, wherein the node represents the data source in the Internet data space, and the query capability represents the query content and support field provided by the node.

[0039] In the embodiment of the present application, considering that different data sources have different requirements for data preparation (segmentation), embedding models, and reordering algorithms, different nodes use their own independent data segmentation, embedding models, and reordering algorithms, that is, each node deploys its own independent retrieval enhancement generation solution according to its own needs, so the query capabilities of each node may be different. Among them, the node support field can be understood as the field or industry served by the node.

[0040] The proxy node matches the query intent with the query capabilities (query content, supported fields) of each node, which can be achieved by calculating the semantic similarity between the query intent and the query capabilities of the node (for example, using cosine similarity to match keyword vectors). The higher the matching degree, the more the query capability of the node matches the query intent. According to the matching degree between the query intent and the query capabilities of each node, multiple candidate nodes are determined, wherein the candidate nodes represent nodes whose query capabilities match the query intent.

[0041] In some embodiments, the query intent is matched with the query capability of each node, and multiple candidate nodes are determined based on the matching degree, specifically including: matching the query intent with the query capability of each node, and sorting the nodes in descending order according to the matching degree to obtain a first sorting result; the first N nodes in the first sorting result are used as candidate nodes, where N is a positive integer greater than 0.

[0042] Among them, the first N nodes in the first sorting result can be understood as the N nodes whose matching degree between the query intent and the query capability of the node is greater than or equal to the target matching degree threshold. For example, if there are 6 nodes, the matching degrees between the query intent and the 6 nodes (node ​​1, node 2, node 3, node 4, node 5, node 6) are calculated as: 0.8, 0.3, 0.5, 0.9, 0.95, 0.7, then the nodes are sorted in descending order according to the matching degree, that is, 0.95 (node ​​5), 0.9 (node ​​4), 0.8 (node ​​1), 0.7 (node ​​6), 0.5 (node ​​3), 0.3 (node ​​2). If the target matching degree threshold is 0.8, then the three nodes of node 1, node 4, and node 5 can be used as candidate nodes.

[0043] In this way, by matching the query intent with the query capabilities of each node and determining multiple candidate nodes based on the matching degree, we can avoid blindly querying all nodes, reduce computing overhead, and improve retrieval efficiency.

[0044] Step S330: sending the query request to the multiple candidate nodes for retrieval and generation, and obtaining query results of the multiple candidate nodes.

[0045] In the embodiment of the present application, after the proxy node determines multiple candidate nodes, it sends a query request to the multiple candidate nodes. After receiving the query request, each candidate node performs a search and generation based on the search enhancement scheme deployed by itself to obtain the query result of the candidate node. Specifically, the candidate node performs a local data search according to the query request and generates a query result according to the search result.

[0046] Step S340: the proxy node generates a final query result according to the query results of the multiple candidate nodes.

[0047] In the embodiment of the present application, after generating the query result, each candidate node sends the query result to the proxy node. The proxy node generates the final query result based on the query results of multiple candidate nodes based on the retrieval enhancement scheme deployed by itself.

[0048] Specifically, the proxy node generates the final query result based on the query results of multiple candidate nodes. The final query result can be generated by combining indicators such as data relevance and authority, effectively filtering out noise data, providing high-quality answers that better meet user needs, and avoiding the problem of traditional methods returning suboptimal or approximate solutions. In this way, by integrating the query results of multiple candidate nodes through the proxy node to generate the final query result, a wider range of data sources can be covered, reducing query omissions caused by the limitations of a single node data, and improving the comprehensiveness and accuracy of the results.

[0049] By adopting the technical solution of the embodiment of the present application, two-level content generation is achieved by setting proxy nodes and nodes in the data space of the Internet of Things. Since the final query result is generated by the proxy node based on the query results of multiple candidate nodes, the transmission of original data between nodes is avoided, the data security of each node in the data space of the Internet of Things is guaranteed, and the accuracy of the final query result is guaranteed. Therefore, a cross-node retrieval enhancement generation in the data space of the Internet of Things is achieved, and the accuracy of the retrieval task in the data space of the Internet of Things is improved.

[0050] In combination with the above embodiments, in one implementation, the present application embodiment further provides a data space-oriented data retrieval enhancement generation method, in which, in addition to the above steps, the following steps are also included: Step A1: The node generates a capability summary of the node according to the retrieval enhancement scheme deployed by itself, wherein the capability summary describes the query content and supported fields provided by the node in natural language form; and registers the capability summary and basic node information to the proxy node.

[0051] Step A2: The node constructs a vector library according to the retrieval enhancement solution deployed by itself, and the vector library represents the data space of the node.

[0052] In the embodiment of the present application, the proxy node and each node are initialized before the search, and data preparation and related registration work are completed. Each node generates a node capability summary according to the search enhancement scheme deployed by itself, and registers the capability summary and node basic information to the proxy node, so that the proxy node can determine the query capability of each node and the basic information of each node according to the registered information, where the node basic information can be node address information, etc. At the same time, each node can also complete data accuracy and vector library construction according to the search enhancement scheme deployed by itself. When the subsequent node performs search generation, it can perform vector search based on the constructed vector library and generate query results according to the large model.

[0053] Furthermore, the "matching of the query intent with the query capability of each node" in the above step S320 specifically includes: the proxy node determines the query capability of each node based on the registered capability summary and node basic information, and matches the query intent with the query capability of each node.

[0054] In an embodiment of the present application, when performing a search, after receiving a query request and determining the query intent, the proxy node can determine the query capability (query content, supported fields) of each node based on the registered capability summary and the basic information of the node, thereby determining multiple candidate nodes based on the matching degree between the query intent and the query capability of each node, avoiding blindly querying all nodes, reducing computing overhead, and improving retrieval efficiency.

[0055] Furthermore, the above step S330 of “sending the query request to the multiple candidate nodes for retrieval and generation to obtain query results of the multiple candidate nodes” may specifically include sub-steps S330-1 to S330-2: Step S330 - 1 : the proxy node processes the query request into a target query request and sends it to the multiple candidate nodes, where the target query request includes a query description and keywords in a natural language form.

[0056] Step S330 - 2 : Each candidate node performs retrieval generation based on the retrieval enhancement solution deployed by itself, the target query request and the vector library to obtain query results of multiple candidate nodes.

[0057] In an embodiment of the present application, the proxy node uses keyword technology to extract keywords from the query request, obtains keywords, and sends the keywords and the corresponding query description as a target query request to the candidate node. After receiving the target query request, the candidate node performs retrieval generation according to the retrieval enhancement scheme and vector library deployed by itself, and obtains the corresponding query result. Among them, the retrieval enhancement scheme deployed by the node itself may have retrieval enhancement generation capability, or may not have retrieval enhancement generation capability, and the query result generation methods of those with retrieval enhancement generation capability and those without retrieval enhancement generation capability are different.

[0058] Specifically, in the above step S330-2, "each candidate node performs retrieval generation based on the retrieval enhancement scheme deployed by itself, the target query request and the vector library to obtain query results for multiple candidate nodes" may specifically include: Step B1: When the retrieval enhancement scheme deployed by the candidate node itself has the retrieval enhancement generation capability, the vector library is searched according to the query description of the target query request, and the query results for the candidate nodes are generated according to the large language model; Step B2: When the retrieval enhancement scheme deployed by the candidate node itself does not have the retrieval enhancement generation capability, metadata retrieval is performed in the vector library according to the keywords in the target query request to obtain the query results for the candidate nodes.

[0059] That is to say, if the candidate node has the ability to generate enhanced retrieval, the large model can be used to generate query results that match the target query request based on the retrieved data (digital objects); if the candidate node does not have the ability to generate enhanced retrieval, the metadata of the digital object retrieved based on the keywords will be used as the query result of the candidate node.

[0060] By adopting the technical solution of the embodiment of the present application, each candidate node can perform retrieval generation according to the retrieval enhancement solution deployed by itself to obtain the corresponding query results, so that the proxy node can generate the final query results according to the query results of each candidate node, so that the final query results can cover a wider range of data sources, reduce query omissions caused by the data limitations of a single node, and improve the comprehensiveness and accuracy of the results.

[0061] In combination with the above embodiments, in one implementation, the embodiment of the present application further provides a data space-oriented data retrieval enhancement generation method. In the method, the "the proxy node generates a final query result according to the query results of the multiple candidate nodes" in the above step S340 may specifically include sub-steps S340-1 to S340-3: Step S340 - 1 : Match the query result of each candidate node with the query request, and sort the query results of the candidate nodes in descending order of matching degree to obtain a second sorting result.

[0062] Step S340-2: Use the query results of the first M candidate nodes in the second sorting result as the target query result, where M is a positive integer greater than 0.

[0063] Step S340-3: Generate a final query result based on the target query result.

[0064] In an embodiment of the present application, considering that the query results of the candidate nodes may not meet the query intent of the query request, in order to obtain more accurate query results, the query results of each candidate node are matched with the query request (for example, the semantic similarity of the query intent and the query capability of the node is calculated) to select the query results that meet the query intent to generate the final query result.

[0065] Among them, the query results of the first M candidate nodes in the second sorting result can be understood as the query results of the M candidate nodes whose matching degree between the query results of the candidate nodes and the query request is greater than or equal to the target matching degree threshold. For example, from the query results of 4 candidate nodes, the matching degrees between the query results of the 4 candidate nodes (candidate node 1, candidate node 2, candidate node 3) and the query request are calculated as: 0.9, 0.4, 0.8, then the query results of the candidate nodes are sorted in descending order according to the matching degree, that is, 0.9, 0.8, 0.4. If the target matching degree threshold is 0.8, the query results of candidate node 1 and the query results of candidate node 3 can be used as the target query results.

[0066] By adopting the technical solution of the embodiment of the present application, the query results of multiple candidate nodes that match the query request are integrated through the proxy node, so that the generated final query result is more in line with the query intent. At the same time, the final query result can cover a wider range of data sources, reduce query omissions caused by the data limitations of a single node, and improve the comprehensiveness and accuracy of the results.

[0067] In combination with the above embodiments, in one implementation, the embodiment of the present application also provides a data retrieval enhancement generation method for data space. In this method, in order to improve the credibility of the generated results, the source of the query results is traced, so the identifier of the associated digital object is included in the query results, and the associated nodes are included in the final query results.

[0068] Specifically, in the above step B1, "searching in the vector library according to the query description of the target query request, and generating query results for candidate nodes according to the big model" may specifically include: searching in the vector library according to the query description of the target query request to obtain a first target digital object, and generating query results for candidate nodes according to the first target digital object and the identifier of the first target digital object through the big model.

[0069] In the above step B1, "performing a metadata search in the vector library according to the keywords in the target query request to obtain a query result of a candidate node" may specifically include: performing a metadata search in the vector library according to the keywords in the target query request to obtain a second target digital object, and obtaining a query result of a candidate node according to the metadata of the second target digital object and the identifier of the second target digital object.

[0070] In the above step S340-3, "generating a final query result according to the target query result" may specifically include: generating a final query result according to the target query result and the candidate nodes corresponding to the target query result.

[0071] In the embodiment of the present application, when the candidate node is performing retrieval generation, for the digital object retrieved for generating the query result, the identifier of the digital object is also carried as content in the query result of the candidate node, so that the data can be traced back based on the identifier of the digital object in the query result; and only the identifier of the digital object is added to the query result, that is, the content can be traced and the control of the data content can be achieved. Similarly, when the proxy node generates the final query result, the node corresponding to the screened target query result is also carried in the final query result, so that the data can be traced back based on the node information in the final query result.

[0072] For example, Table 1 is a schematic result of the traceability relationship of the generated content. For example, it can be determined that the final query result is generated based on the query results fed back by nodes 1, 3, and 4, and the query result 1 of node 1 is generated based on digital objects 1 and 2, the query result 2 of node 3 is generated based on digital objects 5 and 7, and the query result 3 of node 4 is generated based on digital objects 9 and 11.

[0073] Table 1 shows the traceability relationship of the query results

[0074] By adopting the technical solution implemented in this application, the digital object identification and node information associated with the query results are added to the query results, and the enhanced generation process of retrieval based on digital objects and the enhanced generation process based on node retrieval content are traced, thereby realizing data tracing of multi-level data generation in the entire retrieval process.

[0075] The following is a description of the data space-oriented data retrieval enhancement generation method in the embodiment of the present application in conjunction with a specific embodiment. Figure 4 As shown, Figure 44 is a flowchart of another data space-oriented data retrieval enhancement generation method provided by an embodiment of the present application, the method comprising the following steps S410 to S470: Step S410: The node generates a capability summary of the node and constructs a vector library according to the retrieval enhancement solution deployed by itself. The capability summary describes the query content and support fields provided by the node in natural language, and the vector library represents the data space of the node.

[0076] Step S420: receiving a query request in natural language form through a proxy node, and determining the query intent according to the query request.

[0077] Step S430: The proxy node determines the query capability of each node based on the registered capability summary and node basic information, matches the query intent with the query capability of each node, and sorts the nodes in descending order according to the matching degree to obtain a first sorting result; the first N nodes in the first sorting result are used as candidate nodes, where N is a positive integer greater than 0.

[0078] Step S440: the proxy node processes the query request into a target query request and sends it to the multiple candidate nodes, where the target query request includes a query description and keywords in a natural language form.

[0079] Step S450: In the case where the retrieval enhancement scheme deployed by the candidate node itself has the retrieval enhancement generation capability, a search is performed in the vector library according to the query description of the target query request to obtain a first target digital object, and a query result for the candidate node is generated through the large model according to the first target digital object and the identifier of the first target digital object.

[0080] Step S460: When the retrieval enhancement scheme deployed by the candidate node itself does not have the retrieval enhancement generation capability, metadata retrieval is performed in the vector library according to the keywords in the target query request to obtain the second target digital object, and the query result of the candidate node is obtained according to the metadata of the second target digital object and the identifier of the second target digital object.

[0081] Step S470: Match the query result of each candidate node with the query request, and sort the query results of the candidate nodes in descending order according to the matching degree to obtain a second sorting result; use the query results of the first M candidate nodes in the second sorting result as the target query result, where M is a positive integer greater than 0; generate a final query result based on the target query result and the candidate nodes corresponding to the target query result.

[0082] In the embodiment of the present application, by setting proxy nodes and nodes in the digital network data space, two-level content generation is achieved. Since the final query result is generated by the proxy node based on the query results of multiple candidate nodes, the transmission of original data between nodes is avoided, the data security of each node in the digital network data space is guaranteed, and the accuracy of the final query result is guaranteed; and the enhanced generation process of retrieval based on digital objects and the enhanced generation process based on node retrieval content is traced, so as to realize data tracing of multi-level data generation in the entire retrieval process. Therefore, a cross-node enhanced generation of retrieval in the digital network data space is realized, and the accuracy of retrieval tasks in the digital network data space is improved.

[0083] The present application also provides a data space-oriented data retrieval enhancement generation device, referring to Figure 5 As shown, Figure 5 : is a structural schematic diagram of a data space-oriented data retrieval enhancement generation device provided in an embodiment of the present application, the device comprising: The intention determination module 510 is used to receive a query request in natural language form and determine the query intention according to the query request; A node matching module 520 is used to match the query intent with the query capability of each node, and determine multiple candidate nodes according to the matching degree, wherein the node represents the data source in the data space of the Internet of Things, and the query capability represents the query content and support field provided by the node; A search and generation module 530 is used to send the query request to the multiple candidate nodes for search and generation, and obtain query results of the multiple candidate nodes; The query generation module 540 is used to generate a final query result according to the query results of the multiple candidate nodes.

[0084] In an optional embodiment, the device further includes: A summary generation module, used for the node to generate a capability summary of the node according to the retrieval enhancement scheme deployed by the node, wherein the capability summary describes the query content and supported fields provided by the node in a natural language form; A registration module, used to register the capability summary and node basic information to the proxy node; The node matching module is specifically used for: the proxy node determines the query capability of each node according to the registered capability summary and node basic information, and matches the query intention with the query capability of each node.

[0085] In an optional embodiment, the device further includes: A construction module, used for the node to construct a vector library according to the retrieval enhancement scheme deployed by the node itself, wherein the vector library represents the data space of the node; The retrieval generation module is specifically used for: the proxy node processes the query request into a target query request and sends it to the multiple candidate nodes, the target query request includes a query description and keywords in natural language form; each candidate node performs retrieval generation based on its own deployed retrieval enhancement scheme, the target query request and the vector library to obtain query results for multiple candidate nodes.

[0086] In an optional embodiment, the retrieval generation module includes: A first generation submodule is used to search the vector library according to the query description of the target query request and generate a query result of the candidate node according to the large model when the retrieval enhancement scheme deployed by the candidate node itself has the retrieval enhancement generation capability; The second generation submodule is used to perform metadata search in the vector library according to the keywords in the target query request to obtain the query result of the candidate node when the retrieval enhancement solution deployed by the candidate node itself does not have the retrieval enhancement generation capability.

[0087] In an optional embodiment, the first generating submodule is specifically used to: search the vector library according to the query description of the target query request to obtain a first target digital object, and generate a query result of a candidate node according to the first target digital object and the identifier of the first target digital object through the large model; The second generation submodule is specifically used to perform metadata retrieval in the vector library according to the keywords in the target query request to obtain the second target digital object, and obtain the query result of the candidate node according to the metadata of the second target digital object and the identifier of the second target digital object.

[0088] In an optional embodiment, the node matching module includes: A first matching module is used to match the query intent with the query capability of each node, and sort the nodes in descending order according to the matching degree to obtain a first sorting result; The first sorting module is used to use the first N nodes in the first sorting result as candidate nodes, where N is a positive integer greater than 0.

[0089] In an optional embodiment, the query generation module includes: A second matching module is used to match the query result of each candidate node with the query request, and sort the query results of the candidate nodes in descending order of matching degree to obtain a second sorting result; A second sorting module, used to use the query results of the first M candidate nodes in the second sorting result as the target query result, where M is a positive integer greater than 0; The third generation submodule is used to generate a final query result according to the target query result.

[0090] In an optional embodiment, the third generating submodule is specifically configured to generate a final query result according to the target query result and the candidate nodes corresponding to the target query result.

[0091] The present application also provides an electronic device, referring to Figure 6 , Figure 6 Schematic diagram of the structure of an electronic device provided in an embodiment of the present application. Figure 6 As shown, the electronic device 600 includes: a memory 610 and a processor 620. The memory 610 and the processor 620 are connected via a bus communication. A computer program is stored in the memory 610. The computer program can be run on the processor 620 to implement the steps of the data retrieval enhancement generation method for data space described in the embodiment of the present application.

[0092] The embodiment of the present application also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of the data retrieval enhancement generation method for data space described in the embodiment of the present application are implemented.

[0093] The embodiment of the present application also provides a computer program product, including a computer program, which, when executed by a processor, implements the steps of the data space-oriented data retrieval enhancement generation method described in the embodiment of the present application.

[0094] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0095] The embodiments of the present application are described with reference to the flowcharts and / or block diagrams of the methods and devices according to the embodiments of the present application. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of the processes 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 a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing terminal device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing terminal device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0096] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing terminal device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce a manufactured product including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0097] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal device so that a series of operating steps are executed on the computer or other programmable terminal device to produce a computer-implemented process, thereby providing instructions for executing on the computer or other programmable terminal device to implement the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0098] Although the preferred embodiments of the present application have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the embodiments of the present application.

[0099] Finally, it should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or terminal device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or terminal device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or terminal device including the elements.

[0100] The above is a detailed introduction to the data retrieval enhancement generation method, device and equipment for data space provided by the present application. Specific examples are used in this article to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method of the present application and its core idea; at the same time, for general technical personnel in this field, according to the idea of ​​the present application, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present application.

Claims

1. A data retrieval enhancement generation method for data space, characterized in that: Applied to a spatial retrieval enhancement generation system for Internet data, the system includes an agent node and multiple nodes, and the method includes: Receiving a query request in natural language through a proxy node, and determining a query intent according to the query request; Matching the query intent with the query capability of each node, and determining multiple candidate nodes according to the matching degree, wherein the node represents the data source in the data space of the Internet of Things, and the query capability represents the query content and support field provided by the node; Sending the query request to the multiple candidate nodes for retrieval and generation to obtain query results of the multiple candidate nodes; The proxy node generates a final query result according to the query results of the multiple candidate nodes.

2. The data space-oriented data retrieval enhancement generation method according to claim 1, characterized in that: The method further comprises: The node generates a capability summary of the node according to the retrieval enhancement scheme deployed by the node, wherein the capability summary describes the query content and supported fields provided by the node in a natural language form; Registering the capability summary and node basic information to the proxy node; Matching the query intent with the query capabilities of each node includes: The proxy node determines the query capability of each node according to the registered capability summary and the basic information of the node, and matches the query intent with the query capability of each node.

3. The data space-oriented data retrieval enhancement generation method according to claim 1 or 2, characterized in that: The method further comprises: The node constructs a vector library according to the retrieval enhancement scheme deployed by the node, and the vector library represents the data space of the node; The query request is sent to the multiple candidate nodes for retrieval and generation, and query results of the multiple candidate nodes are obtained, including: The proxy node processes the query request into a target query request and sends it to the multiple candidate nodes, wherein the target query request includes a query description and keywords in a natural language form; Each candidate node performs retrieval generation based on the retrieval enhancement scheme deployed by itself, the target query request and the vector library to obtain query results of multiple candidate nodes.

4. The data space-oriented data retrieval enhancement generation method according to claim 3, characterized in that: Each candidate node performs retrieval generation based on the retrieval enhancement scheme deployed by itself, the target query request and the vector library to obtain query results of multiple candidate nodes, including: In the case where the retrieval enhancement scheme deployed by the candidate node itself has the retrieval enhancement generation capability, searching the vector library according to the query description of the target query request, and generating the query result of the candidate node according to the large model; In the case that the retrieval enhancement scheme deployed by the candidate node itself does not have the retrieval enhancement generation capability, metadata retrieval is performed in the vector library according to the keywords in the target query request to obtain the query result of the candidate node.

5. The data space-oriented data retrieval enhancement generation method according to claim 4, characterized in that: The query description of the target query request is searched in the vector library, and query results of candidate nodes are generated according to the large model, including: Searching the vector library according to the query description of the target query request to obtain a first target digital object, and generating a query result of a candidate node according to the first target digital object and an identifier of the first target digital object through the large model; Performing metadata retrieval in the vector library according to the keywords in the target query request to obtain query results for candidate nodes, including: Perform metadata retrieval in the vector library according to the keywords in the target query request to obtain a second target digital object, and obtain a query result of a candidate node according to the metadata of the second target digital object and the identifier of the second target digital object.

6. The data space-oriented data retrieval enhancement generation method according to claim 1, characterized in that: The query intent is matched with the query capability of each node, and multiple candidate nodes are determined according to the matching degree, including: Matching the query intent with the query capability of each node, and sorting the nodes in descending order according to the matching degree to obtain a first sorting result; The first N nodes in the first sorting result are taken as candidate nodes, where N is a positive integer greater than 0.

7. The data space-oriented data retrieval enhancement generation method according to claim 1, characterized in that: The proxy node generates a final query result according to the query results of the multiple candidate nodes, including: Matching the query result of each candidate node with the query request, and sorting the query results of the candidate nodes in descending order of matching degree to obtain a second sorting result; The query results of the first M candidate nodes in the second sorting result are used as the target query results, where M is a positive integer greater than 0; A final query result is generated according to the target query result.

8. The data space-oriented data retrieval enhancement generation method according to claim 7, characterized in that: According to the target query result, a final query result is generated, including: A final query result is generated according to the target query result and the candidate nodes corresponding to the target query result.

9. A data retrieval enhancement generation device for data space, characterized in that: The device comprises: An intention determination module, configured to receive a query request in natural language form and determine the query intention according to the query request; A node matching module is used to match the query intent with the query capability of each node, and determine multiple candidate nodes according to the matching degree, wherein the node represents the data source in the data space of the Internet of Things, and the query capability represents the query content and support field provided by the node; A search and generation module, used for sending the query request to the multiple candidate nodes for search and generation, and obtaining query results of the multiple candidate nodes; The query generation module is used to generate a final query result according to the query results of the multiple candidate nodes.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the data space-oriented data retrieval enhancement generation method described in any one of claims 1 to 8 are implemented.

Images