Stamping process knowledge retrieval method and device, and electronic equipment

Abstract

The application provides a stamping process knowledge retrieval method and device and electronic equipment. A stamping process knowledge retrieval request is received, and a query text carried by the stamping process knowledge retrieval request is determined. At least one expansion query set is generated according to a core term of the query text. Weight information matched with a target retrieval scene corresponding to the stamping process knowledge retrieval request is determined, and the weight information represents the importance of different expansion query sets in the at least one expansion query set. Based on the weight information and the at least one expansion query set, knowledge documents in a stamping process knowledge base are retrieved, and a retrieval result of the stamping process knowledge retrieval request is generated. Through expansion of the core term of the query text and determination of the weight information matched with the target retrieval scene, the generated retrieval result is more comprehensive and closer to the retrieval requirement, thereby effectively ensuring the accuracy of the stamping process knowledge retrieval.

Description

Technical Field

[0001] This application relates to the field of automotive manufacturing technology, specifically to a method, apparatus, and electronic device for retrieving knowledge of stamping processes. Background Technology

[0002] In the application of stamping process knowledge in automotive manufacturing, stamping process knowledge retrieval is a core bridge connecting the "accumulated knowledge" in the stamping process knowledge base with the "practical problems" encountered by engineers in their work. With the expansion of the stamping process knowledge base, how to provide a method for retrieving stamping process knowledge to achieve accurate retrieval is a pressing technical problem that needs to be solved in this field. Summary of the Invention

[0003] Based on the defects and shortcomings of the existing technology, this application proposes a stamping process knowledge retrieval method, device and electronic equipment to achieve accurate retrieval of stamping process knowledge in the stamping process knowledge base.

[0004] The first aspect of this application provides a method for retrieving knowledge of stamping processes, including: Receive a stamping process knowledge retrieval request and determine the query text carried by the stamping process knowledge retrieval request; Generate at least one extended query set based on the core terms of the query text; Determine the weight information that matches the target retrieval scenario corresponding to the stamping process knowledge retrieval request, wherein the weight information characterizes the importance of different extended query sets in the at least one extended query set; Based on the weight information and the at least one extended query set, the knowledge documents in the stamping process knowledge base are retrieved to generate the retrieval results of the stamping process knowledge retrieval request.

[0005] In one possible implementation, generating at least one extended query set based on the core terms of the query text includes: The query text is optimized to generate a target query text, which consists of process terms of at least one process category; At least one extended query set is generated based on the core terms of the target query text, wherein the core terms include process terms of the at least one process category.

[0006] In one possible implementation, generating at least one extended query set based on the core terms of the target query text includes: Determine at least one semantic relationship of the core terms of the target query text, wherein the at least one semantic relationship includes at least one of term variant relationship, concept hierarchy relationship, and question parameter causal relationship; Generate an extended query set corresponding to each of the at least one semantic relations.

[0007] In one possible implementation, determining the weight information matching the target retrieval scenario corresponding to the stamping process knowledge retrieval request includes: Determine the target retrieval scenario corresponding to the stamping process knowledge retrieval request, whereby the target retrieval scenario represents the urgency of the retrieval. Query the pre-set scene correction information configuration table to determine the scene correction information corresponding to the target retrieval scene; Based on the pre-set basic weight information and the scene correction information, weight information matching the target retrieval scene is generated, and the weight information includes the weight of each of the at least one extended query set.

[0008] In one possible implementation, the step of retrieving knowledge documents in the stamping process knowledge base based on the weight information and the at least one extended query set, and generating retrieval results for the stamping process knowledge retrieval request, includes: Based on the weight information and the at least one extended query set, the matching degree of knowledge documents in the stamping process knowledge base is calculated; the matching degree of knowledge documents represents the degree of matching between the knowledge documents and the query text. Identify the first knowledge document whose matching degree is greater than the preset target matching degree; Obtain the second knowledge document associated with the first knowledge document from the stamping process knowledge base; The search results for the stamping process knowledge retrieval request are generated based on the first knowledge document and the second knowledge document.

[0009] In one possible implementation, calculating the matching degree of knowledge documents in the stamping process knowledge base based on the weight information and the at least one extended query set includes: Based on the weight information and the at least one extended query set, calculate the first matching degree and the second matching degree of the knowledge documents in the stamping process knowledge base; the first matching degree represents the character matching degree, and the second matching degree represents the semantic similarity. The matching degree of the knowledge document is calculated based on the first matching degree and the second matching degree of the knowledge document.

[0010] In one possible implementation, calculating the first and second matching degrees of knowledge documents in the stamping process knowledge base based on the weight information and the at least one extended query set includes: Calculate the character matching degree between the knowledge documents in the stamping process knowledge base and each of the at least one extended query set; The first matching degree of the knowledge document is calculated based on the weight information and the character matching degree between the knowledge document and each of the extended query sets. Calculate the semantic similarity between the knowledge documents in the stamping process knowledge base and each of the at least one extended query set; The second matching degree of the knowledge document is calculated based on the weight information and the semantic similarity between the knowledge document and each of the extended query sets.

[0011] One possible implementation also includes: The first knowledge documents are sorted according to a pre-set sorting rule to generate a sequence of first knowledge documents; the sorting rule is related to the matching degree of the first knowledge documents, or the sorting rule is related to the decision logic information category sequence. Sort the second knowledge documents associated with the first knowledge document to generate a second knowledge document sequence for each first knowledge document; The retrieval results are displayed based on the first knowledge document sequence and the second knowledge document sequence for each of the first knowledge documents in the first knowledge document sequence.

[0012] A second aspect of this application provides a stamping process knowledge retrieval device, comprising: The query text determination unit is used to receive a stamping process knowledge retrieval request and determine the query text carried by the stamping process knowledge retrieval request. A query expansion unit is used to generate at least one expanded query set based on the core terms of the query text. The weight information determination unit is used to determine the weight information that matches the target retrieval scenario corresponding to the stamping process knowledge retrieval request, wherein the weight information characterizes the importance of different extended query sets in the at least one extended query set; The retrieval result generation unit is used to retrieve knowledge documents in the stamping process knowledge base based on the weight information and the at least one extended query set, and generate retrieval results for the stamping process knowledge retrieval request.

[0013] A third aspect of this application provides an electronic device, including a memory and a processor; The memory is connected to the processor and is used to store programs; The processor is used to implement the stamping process knowledge retrieval method as described in the first aspect of this application or any possible implementation thereof by running a program in the memory.

[0014] The fourth aspect of this application provides a chip including a processor and a data interface, wherein the processor reads and runs a program stored in a memory through the data interface to execute the stamping process knowledge retrieval method as described in the first aspect of this application or any possible implementation thereof.

[0015] The fifth aspect of this application provides a computer program product, including a computer program that, when executed by a processor, implements the stamping process knowledge retrieval method as described in the first aspect of this application or any possible implementation thereof.

[0016] The sixth aspect of this application provides a storage medium storing a computer program, which, when executed by a processor, implements the stamping process knowledge retrieval method as described in the first aspect of this application or any possible implementation thereof.

[0017] This application provides a method, apparatus, and electronic device for retrieving stamping process knowledge. The method includes receiving a stamping process knowledge retrieval request, determining the query text carried in the request, generating at least one extended query set based on the core terms of the query text, determining weight information matching the target retrieval scenario corresponding to the stamping process knowledge retrieval request (the weight information characterizes the importance of different extended query sets within the at least one extended query set), and retrieving knowledge documents from a stamping process knowledge base based on the weight information and the at least one extended query set to generate retrieval results for the stamping process knowledge retrieval request. By expanding the core terms of the query text and determining the weight information matching the target retrieval scenario, this application makes the generated retrieval results more comprehensive and closer to the retrieval needs, thereby effectively ensuring the accuracy of stamping process knowledge retrieval. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of this application. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0019] Figure 1 This is a schematic diagram of the structure of a stamping process knowledge retrieval system provided in an embodiment of this application.

[0020] Figure 2This application provides a flowchart of a stamping process knowledge retrieval method.

[0021] Figure 3 A flowchart illustrating a method for determining weight information that matches a target retrieval scenario corresponding to a stamping process knowledge retrieval request, provided in an embodiment of this application.

[0022] Figure 4 The present application provides a flowchart of a method for retrieving knowledge documents in a stamping process knowledge base based on weight information and at least one extended query set, and generating retrieval results for a stamping process knowledge retrieval request.

[0023] Figure 5 This is a schematic diagram of the structure of a stamping process knowledge retrieval device provided in an embodiment of this application.

[0024] Figure 6 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0025] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0026] Exemplary methods In the field of stamping process knowledge retrieval technology, in order to achieve accurate and efficient querying of documents in the stamping process knowledge base, the relevant technologies usually adopt the following approach: receiving natural language queries input by users and using inverted index technology to calculate the character matching degree between query keywords and documents.

[0027] However, when this solution is applied to the retrieval of stamping process knowledge, which is highly specialized, involves diverse terminology, and has varied application scenarios, its performance is less than ideal. A fundamental contradiction lies in the fact that the relevant technologies rely on "exact keyword matching" (e.g., inputting "side panel OP10 sliding line" will yield results, while inputting "right side panel drawing process line defect" will result in a failed search), and cannot understand the relationship between synonyms (e.g., "wrinkling" and "pleats") and hyponyms (e.g., "side panel A pillar" and "side panel") in the field of stamping processes, resulting in "incomplete and inaccurate searches" and low search precision.

[0028] Through in-depth analysis, the inventors discovered that semantic expansion of keywords can be performed before searching based on query keywords, allowing for information retrieval based on the expanded keywords. However, this semantic expansion method inevitably compromises the relevance and accuracy of search results across different business scenarios. For example, in a troubleshooting scenario where a "side panel OP10 sliding line" defect occurs on a production line, requiring an emergency shutdown, using a fixed semantic expansion weight might cause the search system to return a large number of theoretical documents or historical cases that are not highly relevant to the current urgent production problem. This would overwhelm documents that are most valuable for fault diagnosis and process parameter adjustment, resulting in a mismatch between the search result ranking and the urgency of the scenario, and a significant decrease in accuracy.

[0029] Further analysis by the inventors revealed that the root cause of the aforementioned contradiction lies in the fact that the semantic expansion mechanism is static and general, failing to incorporate the specific business scenario of the search request (such as urgency or production activity type) as a dynamic variable into the core logic of the search decision. This makes it unable to adapt to the differentiated needs for knowledge focus in different scenarios such as "emergency fault handling", "normal process debugging" and "new process solution research".

[0030] To overcome the aforementioned contradictions, this invention proposes a different technical approach. Its core concept lies in quantifying abstract retrieval scenarios into differentiated weight coefficients for specific semantic expansion dimensions (such as term variations, concept levels, and causal relationships of question parameters), and using these coefficients to dynamically adjust the matching degree calculation during the retrieval process. This allows the final retrieval result ranking to significantly align with the core knowledge needs of the current retrieval scenario while maintaining multi-dimensional semantic expansion to improve recall.

[0031] In other words, the embodiments of this application provide a stamping process knowledge retrieval method, apparatus and electronic device. By expanding the core terms of the query text and determining the weight information of the target retrieval scenario, the generated retrieval results are more comprehensive and closer to the retrieval needs, thereby effectively ensuring the accuracy of stamping process knowledge retrieval and achieving the technical effect of balancing high recall rate and high scenario-based accuracy in complex stamping process knowledge retrieval environments.

[0032] Figure 1 This is a schematic diagram of the structure of a stamping process knowledge retrieval system provided in an embodiment of this application. Figure 1 As shown, the system includes a user terminal, a stamping process knowledge retrieval server, and a stamping process knowledge base.

[0033] The user terminal is used to receive stamping process knowledge retrieval requests input by the user, which may be submitted through a web page, application programming interface, or dedicated client.

[0034] The stamping process knowledge base stores structured or unstructured knowledge documents in the field of stamping processes, such as defect solution work instructions, process standard parameter tables, rectification case reports, and graphic image data. The knowledge documents in the knowledge base are pre-indexed, and relationships between them can be established; for example, one work instruction may be associated with multiple process standard parameter tables.

[0035] The stamping process knowledge retrieval server is the core device for executing the stamping process knowledge retrieval method provided by this invention. After receiving a stamping process knowledge retrieval request from a user terminal, it performs a knowledge document retrieval operation in the stamping process knowledge base according to the stamping process knowledge retrieval request to generate the retrieval results of the stamping process knowledge retrieval request, and returns the retrieval results to the user terminal.

[0036] Figure 2 This application provides a flowchart of a stamping process knowledge retrieval method according to an embodiment. For example... Figure 2 The stamping process knowledge retrieval method shown can be executed by the aforementioned stamping process knowledge retrieval server, and includes the following steps: S210. Receive a stamping process knowledge retrieval request and determine the query text carried in the stamping process knowledge retrieval request. As a specific implementation, a stamping process knowledge retrieval request sent by a user is received. This stamping process knowledge retrieval request is a structured request object, which can contain multiple fields, such as the stamping process knowledge retrieval request ID (request_id) and the query text (raw_query).

[0037] In this application embodiment, "query text" refers to natural language text entered by the user to describe their knowledge needs. For example, it may include, but is not limited to: How should the OP10 side panel slip line problem be handled? For example, after a user enters the query text on the user terminal interface via voice, text, or other means, they can click the search button on the user terminal interface to send a stamping process knowledge retrieval request carrying the query text to the user terminal interface.

[0038] After receiving a stamping process knowledge retrieval request, the stamping process knowledge retrieval server parses the request to obtain the query text carried by the stamping process knowledge retrieval request.

[0039] Furthermore, the stamping process knowledge retrieval request also carries a unique identifier representing the request, namely, the stamping process knowledge retrieval request ID. Correspondingly, the stamping process knowledge retrieval server, by parsing the request, can also obtain the stamping process knowledge retrieval request ID carried within it.

[0040] S220. Generate at least one extended query set based on the core terms of the query text; After obtaining the query text in step S210, it is necessary to perform semantic expansion based on its core terms to generate at least one expanded query set.

[0041] In this application, "expanded query set" refers to any subset of queries used for retrieval generated by expanding the core terms in the query text with at least one semantic relation. For example, it may include, but is not limited to: query subsets generated based on term variant relations (such as synonyms), query subsets generated based on concept hierarchy relations (such as hyponyms and hypernyms), query subsets generated based on causal relations of question parameters, or any combination thereof. The semantic relation expansion is performed based on a pre-built domain semantic graph, and each query subset constitutes an expanded query set.

[0042] The specific means to achieve the above-mentioned "generating extended query sets" function can include: First, parsing the query text to identify the core terms. Then, invoking a pre-built domain-specific semantic graph. This semantic graph stores various semantic relationships between process terms in the stamping process domain, such as synonyms (term variant relationships), hyponyms (conceptual hierarchy relationships), and the association between problems and process parameters (problem-parameter causal relationships). Finally, for the identified core terms in the query text, traversing the domain-specific semantic graph, collecting related terms under at least two different semantic relationship dimensions, thereby forming multiple extended query sets, each corresponding to a semantic relationship.

[0043] In this embodiment of the application, generating at least one extended query set based on the core terms of the query text includes: optimizing the query text to generate a target query text, the target query text consisting of process terms of at least one process category; and generating at least one extended query set based on the core terms of the target query text, the core terms including process terms of at least one process category.

[0044] In this application, "core terms" refers to standardized terms extracted from the query text that represent process entities or key concepts. More specifically, in the preferred embodiment, core terms specifically refer to process terms obtained after the query text has been standardized.

[0045] In this embodiment, the process terminology dictionary stores process terms for various process categories in the stamping process field. The process of normalizing the query text can be considered as optimizing the query text, converting words representing process entities and key concepts in the query text into process terms corresponding to at least one process category in the process terminology dictionary, thereby generating the optimized target query text. Correspondingly, the core terms of the query text are composed of the various process terms in the target query text.

[0046] In this embodiment, the target query text consists of process terms from at least one process category, which includes parts, processes, core issues, and search requirements. The words representing process entities and key concepts in the query text are standardized to obtain process terms for each process category within the at least one process category. These process terms are then sorted according to a pre-set process category sequence to generate the target query text. The process category sequence indicates the order in which the process terms for each process category are ordered, and the process category sequence consists of parts, processes, core issues, and search requirements in that order.

[0047] It is understandable that the core issue can be understood as a process category that characterizes process problems (such as slip lines, wrinkles, etc.); the retrieval requirement can be understood as a process category that characterizes the retrieval requirement (such as defect solutions, standard parameter tables, etc.). For example, the query text is "How to handle the OP10 side panel slip line problem"; the optimized target query text obtained by optimizing the query text is "Solution to the OP10 side panel slip line defect". The target query text consists of: part (side panel) - process (OP10) - core problem (slip line) - retrieval requirement (defect solution).

[0048] Taking the target query text "Side panel OP10 slip line defect solution" as an example, the process terms in the target query text include side panel, OP10, slip line, and defect solution. Correspondingly, the core terms in the original query text include side panel, OP10, slip line, and defect solution.

[0049] The goal of query text optimization is to improve the standardization of queries from the source, avoiding colloquial or vague expressions that could interfere with semantic expansion. Its advantage lies in generating clearly structured and precisely terminological target query text by standardizing the received query text, thus ensuring the accuracy and standardization of the core terms upon which subsequent semantic expansion operations rely.

[0050] This further helps to solve the problem of inaccurate semantic expansion due to non-standard query text expression, thereby synergistically strengthening the overall technical effect of the present invention in improving retrieval accuracy. In addition, the structure of the target query text is not limited to "part-process-core issue-retrieval requirements". Those skilled in the art can adjust it to other structures according to the domain characteristics of the stamping process knowledge base, such as "equipment-fault phenomenon-maintenance action", etc., which are not limited in the embodiments of this application.

[0051] In this embodiment of the application, the query text is optimized to generate the target query text, and each process term included in the target query text is determined as the core term of the target query text. The core term of the target query text can also be considered as the core term of the query text.

[0052] Generate at least one extended query set based on the core terms of the target query text, including: determining at least one semantic relationship of the core terms of the target query text, wherein the at least one semantic relationship includes at least one of term variant relationship, concept hierarchy relationship, and question parameter causal relationship; and generating an extended query set corresponding to each of the at least one semantic relationship.

[0053] In this embodiment of the application, at least one semantic relationship is pre-set, which includes at least one of term variant relationship, concept hierarchy relationship, and problem parameter causal relationship.

[0054] In this application, "term variant relationship" refers broadly to the equivalent, synonymous, or near-synonymous relationship between different terms used for the same concept or entity under different expressions, contexts, or conventions. For example, it may include, but is not limited to: the relationship between professional terms and common names (such as "slip line" and "sliding line"), the relationship between full names and abbreviations, naming differences under different regions or standards, or terms that are generally recognized as interchangeable within the stamping process field. Specifically, in the stamping process field, identifying term variant relationships helps overcome search omissions caused by inconsistent terminology.

[0055] In this application, "conceptual hierarchy" broadly refers to the inclusion, subordination, or hierarchical relationship between terms in terms of conceptual scope. For example, it may include, but is not limited to, the relationship between a specific part and an assembly component, the relationship between a specific defect type and a major defect category, or the relationship between a specific process and a major technological category. Specifically, utilizing conceptual hierarchy for semantic expansion enables the query intent to be generalized to higher-level concepts, thereby retrieving more comprehensive or relevant knowledge documents.

[0056] In this application, "problem parameter causal relationship" broadly refers to the relationship between a quality problem (or defect) in the stamping process and the process parameters that cause or affect that problem. For example, this may include, but is not limited to, the correspondence between a "slip line" defect in a stamped part and the "drawing pressure" parameter. Specifically, identifying problem parameter causal relationships implicitly links the problem description in a query to relevant process parameter knowledge, achieving cross-semantic knowledge retrieval.

[0057] The above are merely preferred contents of at least one semantic relationship provided in the embodiments of this application. Those skilled in the art can set the specific contents of at least one semantic relationship according to their own needs, and there are no limitations here.

[0058] By expanding the core terms of the target query text from three semantic relation dimensions—term variant relation (addressing expression diversity), concept hierarchy relation (expanding the scope of abstraction), and problem parameter causal relation (mining deep process associations)—we can obtain an expanded query set corresponding to each semantic relation. The expanded query sets corresponding to each semantic relation in at least one semantic relation constitute at least one expanded query set.

[0059] Taking the query text "How to handle the OP10 side slip line problem" and the optimized target query text "Solution to the OP10 side slip line defect" as an example, the core terms in the target query text include side, OP10, slip line, and defect solution. If at least one semantic relationship includes term variation relationship, concept hierarchy relationship, and problem parameter causal relationship, the process of generating at least one extended query set is as follows: (1) Terminology variant relationship: By querying the "synonyms" or "process equivalent terms" relationship in the domain semantic graph, the equivalent terms of the same process term under different expressions are identified, the variant "sliding line" and "drawing sliding line" are found for "sliding line", and the extended query set A1 = {side wall, OP10, sliding line, sliding line, drawing sliding line, defect solution} is generated.

[0060] (2) Conceptual hierarchy: By querying the conceptual hierarchy in the domain semantic graph, the hierarchical relationship between terms is identified. The superordinate concept "side sliding line" is found for "side sliding line OP10". Furthermore, the superordinate concept "stamping surface defect" is found for "side sliding line". An extended query set A2 = {side sliding line OP10, side sliding line, stamping surface defect} is generated.

[0061] (3) Problem parameter causal relationship: By querying the "defect-parameter" mapping table in the domain semantic graph, the association between process terms and related process parameters is identified, and it is determined that "slip line" is usually related to "drawing pressure", generating an extended query set A3 = {slip line, drawing pressure}.

[0062] The three semantic relationships mentioned above can be used simultaneously, or two can be selected according to the actual scenario. By adopting the specific limiting features of generating extended query sets based on multiple preset semantic relationships, this preferred solution can enrich the query intent from different dimensions, thereby capturing relevant knowledge documents in the stamping process knowledge base more precisely and from multiple angles, and improving the recall and precision of the retrieval.

[0063] It should be noted that the setting of "at least one semantic relationship" in step S220 provides an operational object for dynamically allocating weights according to the target retrieval scenario in the subsequent step S230. The synergy between the two enables the flexible adjustment of the importance of different semantic expansion directions according to the target retrieval scenario, thereby making the ranking of retrieval results more in line with the needs of the target retrieval scenario and jointly solving the problem of poor scenario adaptability of retrieval results.

[0064] S230. Determine the weight information that matches the target retrieval scenario corresponding to the stamping process knowledge retrieval request. The weight information represents the importance of different extended query sets in at least one extended query set. Figure 3 This application provides a flowchart of a method for determining weight information that matches a target retrieval scenario corresponding to a stamping process knowledge retrieval request. (See flowchart for example.) Figure 3 As shown, the method includes: S310. Determine the target retrieval scenario corresponding to the stamping process knowledge retrieval request. The target retrieval scenario represents the urgency of the retrieval. In this embodiment of the application, at least one process scenario label is set in advance. The at least one process scenario label includes: production line shutdown, process debugging during normal production of the production line, small-scale defect rectification, batch defect output, new process scheme design, new mold development, quality spot check rectification, and similar technology reference.

[0065] The above are merely preferred contents of at least one process scenario label provided in the embodiments of this application. Those skilled in the art can set the specific contents of at least one process scenario label according to their own needs, and there are no limitations here.

[0066] In this embodiment, the retrieval scenario represents the urgency of the retrieval. The target retrieval scenario corresponding to the stamping process knowledge retrieval request is determined based on the keywords representing the urgency of the retrieval in the query text carried by the stamping process knowledge retrieval request and the process scenario tags carried by the stamping process knowledge retrieval request.

[0067] In this embodiment of the application, at least one search scenario is pre-set. The at least one search scenario includes a first search scenario, a second search scenario, and a third search scenario. The first search scenario represents an urgent search, the second search scenario represents a regular search, and the third search scenario represents a research search.

[0068] The above are merely preferred contents of at least one search scenario provided in the embodiments of this application. The specific contents of at least one search scenario can be set by those skilled in the art according to their own needs, and are not limited here.

[0069] In this embodiment of the application, when a user sends a stamping process knowledge retrieval request through a user terminal, the user can be prompted to enter the target process scenario tag corresponding to the current stamping process knowledge retrieval request, and thus the received stamping process knowledge retrieval request sent by the user carries the target process scenario tag.

[0070] Specifically, when a user sends a stamping process knowledge retrieval request, at least one process scenario tag is displayed to the user. The system responds to the user's input of query text and selection of the target process scenario tag from at least one process scenario tag, generating a stamping process knowledge retrieval request. The stamping process knowledge retrieval request carries the query text and the target process scenario tag.

[0071] It is understandable that the query text may also include keywords that indicate the urgency of the search, such as "urgent" or "very urgent".

[0072] For example, the first search scenario is: the stamping process knowledge search request carries process scenario tags that represent urgent production activities such as production line shutdown and batch defect production, or the query text of the stamping process knowledge search request carries keywords that represent the urgency of the search. In this case, the search results should be returned first. The target search scenario corresponding to the stamping process knowledge search request is the first search scenario, and the first search scenario is identified as P1. Second search scenario: The stamping process knowledge search request carries process scenario tags that represent routine production activities such as process debugging, quality spot check rectification, and small-scale defect rectification during normal production. In this case, the search priority is routine, and the target search scenario corresponding to the stamping process knowledge search request is the second search scenario, which is identified as P2. Third search scenario: Stamping process knowledge search requests carry non-production-related process scenario tags such as new process scheme design, new mold product development, and similar technology references. In this case, there is no time limit for the search, and the target search scenario corresponding to the stamping process knowledge search request is the third search scenario, which is identified as P3.

[0073] It should be noted that, generally, when the tags carried in a stamping process knowledge retrieval request are process scenario tags representing routine production activities or non-production-related process scenario tags, the query text of the stamping process knowledge retrieval request will not contain keywords indicating the urgency of the search. However, if, in this case, the query text of the stamping process knowledge retrieval request contains keywords indicating the urgency of the search, then the target retrieval scenario corresponding to the stamping process knowledge retrieval request is determined to be the first retrieval scenario.

[0074] S320. Query the pre-set scene correction information configuration table to determine the scene correction information corresponding to the target retrieval scene; In this embodiment, a scene correction information configuration table is pre-set, which includes the correspondence between search scenes and scene correction information. Based on the target search scene, the scene correction information corresponding to the target search scene is obtained by querying the scene correction information configuration table.

[0075] In this embodiment of the application, basic weight information is preset. The basic weight information includes the basic weight of the extended query set corresponding to each semantic relation in at least one semantic relation. The scene correction information corresponding to the retrieval scene includes the correction coefficient of each semantic relation in at least one semantic relation. The correction coefficient of the semantic relation is used to correct the basic weight of the extended query set corresponding to the semantic relation to obtain the weight of the extended query set corresponding to the semantic relation.

[0076] S330. Based on the pre-set basic weight information and scenario correction information, generate weight information that matches the target retrieval scenario. The weight information includes the weight of each extended query set in at least one extended query set.

[0077] In this embodiment, the basic weight information includes the basic weight of the extended query set corresponding to each semantic relation in at least one semantic relation. Taking at least one semantic relation including term variant relation, concept hierarchy relation, and problem parameter causal relation as an example, the basic weight information includes the basic weight of the extended query set corresponding to the term variant relation, the basic weight of the extended query set corresponding to the concept hierarchy relation, and the basic weight of the extended query set corresponding to the problem parameter causal relation.

[0078] For example, in the basic weight information, the basic weight of the extended query set corresponding to the term variant relationship is 3, the basic weight of the extended query set corresponding to the concept hierarchy relationship is 2, and the basic weight of the extended query set corresponding to the problem parameter causal relationship is 5.

[0079] The above is merely a preferred embodiment of basic weight information provided in this application. The specific values ​​of the basic weights of the extended query sets corresponding to each semantic relationship in the basic weight information can be set by those skilled in the art according to their own needs, and are not limited here.

[0080] The table below is an example diagram of a scene correction information configuration provided in an embodiment of this application.

[0081] The above is merely an example of a scene correction information configuration table provided in this application embodiment. The specific values ​​of the correction coefficients in the scene correction information configuration table can be set by those skilled in the art according to their own needs, and are not limited here.

[0082] Taking at least one semantic relationship, including term variant relationship, concept hierarchy relationship, and problem parameter causal relationship, and with pre-set basic weight information including a basic weight of 3 for the extended query set corresponding to the term variant relationship, a basic weight of 2 for the extended query set corresponding to the concept hierarchy relationship, and a basic weight of 5 for the extended query set corresponding to the problem parameter causal relationship, and with scenario correction information corresponding to the target retrieval scenario (P1) of the stamping process knowledge retrieval request including a correction coefficient of 0.8 for the term variant relationship, a correction coefficient of 0.8 for the concept hierarchy relationship, and a correction coefficient of 1.2 for the problem parameter causal relationship, the weight of the extended query set corresponding to the term variant relationship is 2.4 (3). 0.8 = 2.4), the weight of the extended query set corresponding to the concept hierarchy relationship is 1.6 (2 0.8 = 1.6), the weight of the extended query set corresponding to the causal relationship of the problem parameter is 6 (5 1.2 = 6).

[0083] In this embodiment, based on the target retrieval scenario, a pre-set scenario correction information configuration table is queried to obtain the correction coefficients for each semantic relationship. For example, for scenario "P1", the scenario correction information configuration table specifies that the correction coefficient for term variant relationships is 0.8, the correction coefficient for concept hierarchy relationships is 0.8, and the correction coefficient for problem parameter causal relationships is 1.2. This set of coefficients represents the scenario correction information matching the current target retrieval scenario P1. It clearly indicates that in the current P1 scenario, the most important factors should be the knowledge of process parameters related to the problem, followed by synonyms related to defect descriptions and higher-level defect category knowledge. Using this method, the requirements of the process scenario can be directly and interpretably transformed into the control parameters of the retrieval algorithm.

[0084] S240. Based on weight information and at least one extended query set, retrieve knowledge documents in the stamping process knowledge base and generate retrieval results for the stamping process knowledge retrieval request.

[0085] As an optional implementation of this application embodiment, the knowledge documents in the stamping process knowledge base are stored according to a classification system. For example, the knowledge documents in the stamping process knowledge base are divided into four categories for storage: "process standard knowledge, defect solution knowledge, case knowledge, and graphic knowledge".

[0086] When the target query text is "Solution to the P10 slip line defect in the side panel", the "defect solution" is associated with the "defect solution knowledge" knowledge category through a pre-defined rule mapping table. Then, combined with the problem term "slip line" in the target query text, the pre-established "problem-supporting knowledge type" association rule is invoked to identify that this type of defect problem usually requires supporting process parameters and similar rectification records in actual handling. Thus, the process parameters are further associated with the "process standard knowledge" knowledge category, and the similar rectification records are associated with the "case knowledge" knowledge category. Finally, knowledge documents are retrieved from the three knowledge categories of "process standard knowledge", "defect solution knowledge" and "case knowledge" stored in the stamping process knowledge base.

[0087] Figure 4 This application provides a flowchart of a method for retrieving knowledge documents from a stamping process knowledge base based on weight information and at least one extended query set, and generating retrieval results for a stamping process knowledge retrieval request. (See flowchart for example.) Figure 4 As shown, the method includes: S410. Based on weight information and at least one extended query set, calculate the matching degree of knowledge documents in the stamping process knowledge base; the matching degree of knowledge documents represents the degree of matching between knowledge documents and query text. In this embodiment of the application, the matching degree of knowledge documents in the stamping process knowledge base is calculated based on weight information and at least one extended query set, including: calculating a first matching degree and a second matching degree of knowledge documents in the stamping process knowledge base according to weight information and at least one extended query set; the first matching degree represents the character matching degree, and the second matching degree represents the semantic similarity; and calculating the matching degree of knowledge documents based on the first matching degree and the second matching degree of knowledge documents.

[0088] In this embodiment of the application, calculating the first matching degree and the second matching degree of knowledge documents in the stamping process knowledge base based on weight information and at least one extended query set includes: calculating the character matching degree between the knowledge documents in the stamping process knowledge base and each of the at least one extended query set; calculating the first matching degree of the knowledge documents based on the weight information and the character matching degree between the knowledge documents and each extended query set; calculating the semantic similarity between the knowledge documents in the stamping process knowledge base and each of the at least one extended query set; and calculating the second matching degree of the knowledge documents based on the weight information and the semantic similarity between the knowledge documents and each extended query set.

[0089] Calculate the matching degree of each knowledge document in the stamping process knowledge base. Taking a knowledge document (temporarily referred to as knowledge document D) as an example, to calculate the matching degree of knowledge document D, we need to first calculate the first matching degree and the second matching degree of knowledge document D. If at least one semantic relationship includes term variant relationship, concept hierarchy relationship, and problem parameter causal relationship, and the extended query set generated corresponding to the term variant relationship is extended query set A1, the extended query set generated corresponding to the concept hierarchy relationship is extended query set A2, and the extended query set generated corresponding to the problem parameter causal relationship is extended query set A3, then the calculation method of the first matching degree Score_char_D of knowledge document D can be as follows: Score_char_D = (W1 B1+W2 B2+W3 B3) / (W1+W2+W3) Where W1 is the weight of extended query set A1, W2 is the weight of extended query set A2, W3 is the weight of extended query set A3, B1 is the character matching degree between extended query set A1 and knowledge document D, B2 is the character matching degree between extended query set A2 and knowledge document D, and B3 is the character matching degree between extended query set A3 and knowledge document D.

[0090] For example, the character matching degree can be calculated using algorithms such as TF-IDF and BM25 based on inverted indexes, and this application embodiment does not limit it.

[0091] The second matching degree, Score_semantic_D, of knowledge document D can be calculated as follows: Score_semantic_D = (W1) C1+W2 C2+W3 C3) / (W1+W2+W3) Where W1 is the weight of extended query set A1, W2 is the weight of extended query set A2, W3 is the weight of extended query set A3, C1 is the semantic similarity between extended query set A1 and knowledge document D, C2 is the semantic similarity between extended query set A2 and knowledge document D, and C3 is the semantic similarity between extended query set A3 and knowledge document D.

[0092] For example, semantic similarity can be calculated using cosine similarity, but this application does not limit the specific implementation.

[0093] Specifically, in some embodiments, the matching degree Score_final_D of knowledge document D can be calculated as: Score_final_D = α Score_char_D+(1-α) Score_semantic_D, where Score_char_D is the first matching degree of knowledge document D, Score_semantic_D is the second matching degree of knowledge document D, and α is an adjustable parameter, for example, set to 0.5.

[0094] Understandably, the matching degree of a knowledge document represents the degree of matching between the knowledge document and the query text; the first matching degree of a knowledge document represents the character matching degree between the knowledge document and the query text; and the second matching degree of a knowledge document represents the semantic similarity between the knowledge document and the query text. By adopting the specific limiting features of the above-mentioned dual-dimensional matching calculation that integrates characters and semantics, this preferred scheme makes the matching degree evaluation mechanism more robust and improves the accuracy of retrieval.

[0095] S420. Identify the first knowledge document whose matching degree is greater than the preset target matching degree; For each knowledge document in the stamping process knowledge base, calculate its matching degree, and determine each knowledge document in the stamping process knowledge base whose matching degree is greater than the preset target matching degree as a first knowledge document.

[0096] For example, the target matching degree is 80%. The above is merely a preferred content of the target matching degree provided in the embodiments of this application. The specific value of the target matching degree can be set by those skilled in the art according to their own needs, and is not limited here.

[0097] S430: Obtain the second knowledge document associated with the first knowledge document from the stamping process knowledge base; Based on the preliminary retrieval of relevant documents (first knowledge documents), the predefined document relationships in the stamping process knowledge base are further explored, and other knowledge documents (second knowledge documents) associated with the first knowledge documents in the stamping process knowledge base are also included in the retrieval results.

[0098] This design aims to overcome the limitations of single documents. The principle is that knowledge is often interconnected in a network; the effective understanding and application of a knowledge document (such as a defect solution) often requires reference to the process standards it cites, related case studies, or parameter tables. By proactively aggregating this related knowledge, more comprehensive support can be provided to users. For example, the first knowledge document retrieved might be a "Side Panel OP10 Slip Line Rectification Operation Instruction Manual." Based on the predefined association data table in the knowledge base, a second knowledge document associated with this first knowledge document can be found, such as a "Drawing Pressure Process Standard Parameter Table" and a "Slip Line Historical Rectification Case Report."

[0099] As one implementation of this application, in the process of storing knowledge documents into the stamping process knowledge base, the association relationship between the knowledge document to be stored in the stamping process knowledge base (hereinafter referred to as the knowledge document to be stored) and each stored knowledge document in the stamping process knowledge base (hereinafter referred to as the knowledge document already in the database) is first determined. The database already in the stamping process knowledge base that has an association relationship with the knowledge document to be stored is determined, and an association relationship information is generated. The association relationship information indicates the knowledge document to be stored and the database already in the database that has an association relationship with the knowledge document to be stored. Then, the knowledge document to be stored is stored in the stamping process knowledge base, and the generated association relationship information is stored in the association relationship data table in the stamping process knowledge base.

[0100] The above is merely a preferred method for obtaining associated documents provided in the embodiments of this application. Those skilled in the art can set the specific method for obtaining associated documents according to their own needs, and it is not limited here.

[0101] It should be noted that the matching degree of the second knowledge document may be lower than that of the first knowledge document, but it is considered to have supplementary value due to its association with the first knowledge document. By adopting the specific limiting features of the above-mentioned related knowledge document expansion, this preferred solution can proactively aggregate scattered but related knowledge documents for users, providing more comprehensive and systematic knowledge support, and effectively improving the completeness and usability of search results.

[0102] S440. Generate the search results for the stamping process knowledge retrieval request based on the first knowledge document and the second knowledge document.

[0103] In this embodiment of the application, the number of first knowledge documents may be one or more, and the second knowledge documents associated with each first knowledge document are determined. Taking a first knowledge document as an example, the knowledge documents stored in the stamping process knowledge base that have a relationship with it are determined as the second knowledge documents associated with the first knowledge document.

[0104] It should be noted that if the second knowledge document is duplicated with the first knowledge document, the duplicate second knowledge document will be deleted to obtain the target second knowledge document. The first knowledge document and the target second knowledge document will then constitute the search results for the stamping process knowledge retrieval request.

[0105] For example, upon receiving a stamping process knowledge retrieval request, the first knowledge documents identified from the stamping process knowledge base include knowledge document 1, knowledge document 2, and knowledge document 3. A query of the association data table determines that the second knowledge documents associated with knowledge document 1 include knowledge document 2, knowledge document 4, and knowledge document 5; the second knowledge documents associated with knowledge document 2 include knowledge document 1, knowledge document 6, and knowledge document 7; and the second knowledge document associated with knowledge document 3 is knowledge document 8. At this point, it is determined that knowledge document 2 in the second knowledge documents associated with knowledge document 1 is a duplicate of the first knowledge document, and knowledge document 2 is deleted from the second knowledge documents associated with knowledge document 1 to obtain the target second knowledge documents associated with knowledge document 1 (knowledge document 4 and knowledge document 5). It is also determined that knowledge document 1 in the second knowledge documents associated with knowledge document 2 is a duplicate of the first knowledge document, and knowledge document 1 is deleted from the second knowledge documents associated with knowledge document 2 to obtain the target second knowledge documents associated with knowledge document 2 (knowledge document 6 and knowledge document 7). Finally, it is determined that the knowledge documents in the second knowledge documents associated with knowledge document 3 do not duplicate the first knowledge document, so no deduplication operation is performed, and the second knowledge document associated with knowledge document 3 is determined as the target second knowledge document associated with knowledge document 3 (knowledge document 8). The search results are generated based on the first knowledge document and the target second knowledge document associated with the first knowledge document. The search results include: knowledge document 1, knowledge document 2, knowledge document 3, the target second knowledge document associated with knowledge document 1 (knowledge document 4 and knowledge document 5), the target second knowledge document associated with knowledge document 2 (knowledge document 6 and knowledge document 7), and the target second knowledge document associated with knowledge document 3 (knowledge document 8).

[0106] To optimize the presentation of search results and improve users' knowledge acquisition efficiency and experience, this embodiment of the application further includes, after generating search results: sorting each first knowledge document according to a pre-set sorting rule to generate a first knowledge document sequence; the sorting rule is related to the matching degree of the first knowledge document, or the sorting rule is related to the decision logic information category sequence; sorting each second knowledge document associated with the first knowledge document to generate a second knowledge document sequence for each first knowledge document; and displaying the search results based on the first knowledge document sequence and the second knowledge document sequence of each first knowledge document in the first knowledge document sequence.

[0107] In this preferred solution, the search results are organized and sorted hierarchically. This constraint aims to clearly reveal the relationship structure between knowledge points and optimize information navigation. The principle is that users typically focus first on the most relevant core answers (first knowledge document), and then explore its extended knowledge as needed (second knowledge document). Hierarchical display aligns with this cognitive habit. The display can employ tree lists, collapsible panels, or master-slave views, among other methods.

[0108] As an optional implementation of this application, the sorting rule can be to sort the first knowledge documents in descending order of matching degree, and sort each first knowledge document in descending order of matching degree to obtain a sequence of first knowledge documents.

[0109] As another optional implementation of this application embodiment, the sorting rule can be to sort according to the order of the decision logic information category sequence, and sort each first knowledge document according to the order of the decision logic information category sequence to obtain the first knowledge document sequence.

[0110] For example, the decision logic information category sequence consists of problem location category, cause analysis category, solution category, parameter basis category, and effect verification category in sequence.

[0111] The above is merely a preferred content of the decision logic information category sequence provided in the embodiments of this application. The specific content of the decision logic information category sequence can be set by those skilled in the art according to their own needs, and is not limited here.

[0112] In this embodiment of the application, for each first knowledge document, the associated second knowledge documents are sorted to generate a sequence of second knowledge documents for the first knowledge document.

[0113] In a preferred implementation of this application, the second knowledge documents associated with the first knowledge document are sorted in descending order of their matching degree to generate a sequence of second knowledge documents for the first knowledge document. It is understood that the matching degree of the second knowledge document is the matching degree between the second knowledge document and the query text.

[0114] In another preferred implementation of this application, the second knowledge documents associated with the first knowledge document are sorted in descending order of their relevance to the first knowledge document, thereby generating a sequence of second knowledge documents for the first knowledge document. The calculation method for the relevance between the second knowledge document and the first knowledge document may be related to character matching degree and semantic similarity, and is not limited here.

[0115] The following example illustrates how search results are displayed, using the second knowledge document sequence sorted in descending order of their matching degree: The search results are presented as a hierarchical sequence. The outermost layer is the first knowledge document sequence, for example, knowledge document A (match degree 0.92) and knowledge document B (match degree 0.87) sorted sequentially. When a user expands knowledge document A, they can see its associated second knowledge document sequence, such as knowledge document A1 (match degree 0.78) and knowledge document A2 (match degree 0.65). Knowledge document B is associated with knowledge document B1, and so on. This display method allows users to quickly locate the core answer and conveniently explore extended knowledge.

[0116] Furthermore, in this embodiment of the application, when displaying a knowledge document, a summary information of the knowledge document can be displayed. The summary information is used to display the detailed content of the knowledge document in response to a trigger operation.

[0117] By adopting the aforementioned hierarchical sorting and specific limitations in result display, this preferred solution enhances the intuitiveness of result navigation and the systematic nature of knowledge acquisition.

[0118] This application provides a method for retrieving knowledge about stamping processes. The method receives a request for knowledge retrieval of stamping processes and determines the query text carried in the request. It generates at least one extended query set based on the core terms of the query text. It determines weight information matching the target retrieval scenario corresponding to the request, where the weight information characterizes the importance of different extended query sets within the at least one extended query set. Based on the weight information and at least one extended query set, it retrieves knowledge documents from a stamping process knowledge base to generate retrieval results for the request. This application, by expanding the core terms of the query text and determining the weight information matching the target retrieval scenario, makes the generated retrieval results more comprehensive and closer to the retrieval requirements, thus effectively ensuring the accuracy of stamping process knowledge retrieval. Specifically, this application dynamically maps abstract retrieval scenarios to differentiated levels of importance for different semantic expansion directions, thereby adjusting the contribution of each expansion path to the matching degree during retrieval. This ensures that the retrieval results closely match the core knowledge needs of the current scenario, effectively improving the accuracy and scenario adaptability of the retrieval.

[0119] Exemplary device Figure 5 This is a schematic diagram of the structure of a stamping process knowledge retrieval device provided in an embodiment of this application.

[0120] like Figure 5 As shown, the stamping process knowledge retrieval device 500 includes: The query text determination unit 510 is used to receive a stamping process knowledge retrieval request and determine the query text carried in the stamping process knowledge retrieval request. The query extension unit 520 is used to generate at least one extended query set based on the core terms of the query text. The weight information determination unit 530 is used to determine the weight information that matches the target retrieval scenario corresponding to the stamping process knowledge retrieval request. The weight information represents the importance of different extended query sets in at least one extended query set. The retrieval result generation unit 540 is used to retrieve knowledge documents in the stamping process knowledge base based on weight information and at least one extended query set, and generate retrieval results for the stamping process knowledge retrieval request.

[0121] In one possible implementation, the query extension unit includes: The target query text generation unit is used to optimize the query text to generate target query text, which consists of process terms of at least one process category. The query extension subunit is used to generate at least one extended query set based on the core terms of the target query text, wherein the core terms include process terms of at least one process category.

[0122] In one possible implementation, querying the extended subunit includes: The semantic relation determination unit is used to determine at least one semantic relation of the core terms of the target query text. The at least one semantic relation includes at least one of the following: term variant relation, concept hierarchy relation, and question parameter causal relation. An extended query set generation unit is used to generate extended query sets corresponding to each semantic relation in at least one semantic relation.

[0123] In one possible implementation, the weight information determination unit includes: The target retrieval scenario determination unit is used to determine the target retrieval scenario corresponding to the stamping process knowledge retrieval request. The target retrieval scenario represents the urgency of the retrieval. The scene correction information determination unit is used to query a pre-set scene correction information configuration table to determine the scene correction information corresponding to the target retrieval scene. The weight information determination subunit is used to generate weight information that matches the target retrieval scenario based on pre-set basic weight information and scenario correction information. The weight information includes the weight of each extended query set in at least one extended query set.

[0124] In one possible implementation, the retrieval result generation unit includes: The matching degree calculation unit is used to calculate the matching degree of knowledge documents in the stamping process knowledge base based on weight information and at least one extended query set; the matching degree of knowledge documents represents the degree of matching between knowledge documents and query text. The first knowledge document determination unit is used to determine first knowledge documents with a matching degree greater than a pre-set target matching degree. The second knowledge document determination unit is used to obtain the second knowledge document associated with the first knowledge document in the stamping process knowledge base; The retrieval results generation sub-unit is used to generate retrieval results for the stamping process knowledge retrieval request based on the first knowledge document and the second knowledge document.

[0125] In one possible implementation, the matching degree calculation unit includes: The first calculation unit is used to calculate the first matching degree and the second matching degree of knowledge documents in the stamping process knowledge base based on weight information and at least one extended query set; the first matching degree represents the character matching degree, and the second matching degree represents the semantic similarity. The second calculation unit is used to calculate the matching degree of the knowledge document based on the first matching degree and the second matching degree of the knowledge document.

[0126] In one possible implementation, the first computing unit includes: The character matching degree calculation unit is used to calculate the character matching degree between the knowledge documents in the stamping process knowledge base and each of the at least one extended query set; The first matching degree calculation unit is used to calculate the first matching degree of the knowledge document based on the weight information and the character matching degree between the knowledge document and each extended query set. A semantic similarity calculation unit is used to calculate the semantic similarity between knowledge documents in the stamping process knowledge base and each of the at least one extended query set; The second matching degree calculation unit is used to calculate the second matching degree of the knowledge document based on the weight information and the semantic similarity between the knowledge document and each extended query set.

[0127] In one possible implementation, the stamping process knowledge retrieval device further includes a retrieval result display unit, which includes: The first sequence generation unit is used to sort each first knowledge document according to a pre-set sorting rule to generate a first knowledge document sequence; the sorting rule is related to the matching degree of the first knowledge documents, or the sorting rule is related to the decision logic information category sequence; The second sequence generation unit is used to sort the various second knowledge documents associated with the first knowledge document and generate a sequence of second knowledge documents for each first knowledge document. The search results display sub-unit is used to display search results based on the first knowledge document sequence and the second knowledge document sequence for each first knowledge document in the first knowledge document sequence.

[0128] The stamping process knowledge retrieval device provided in this embodiment belongs to the same application concept as the stamping process knowledge retrieval method provided in the above embodiments of this application. It can execute the stamping process knowledge retrieval method provided in any of the above embodiments of this application and has the corresponding functional modules and beneficial effects of executing the stamping process knowledge retrieval method. Technical details not described in detail in this embodiment can be found in the corresponding method embodiments of this application, and will not be repeated here.

[0129] The functions implemented by each unit in the above device can be implemented by the same or different processors, and this application embodiment does not limit this.

[0130] It should be understood that the units in the above device can be implemented by a processor calling software. For example, the device includes a processor connected to a memory containing instructions. The processor calls the instructions stored in the memory to implement any of the above methods or to implement the functions of each unit in the device. The processor can be a general-purpose processor, such as a CPU or microprocessor, and the memory can be internal or external to the device. Alternatively, the units in the device can be implemented as hardware circuits. By designing the hardware circuits, some or all of the unit functions can be implemented. The hardware circuits can be understood as one or more processors. For example, in one implementation, the hardware circuit is an ASIC, and the functions of some or all of the above units are implemented by designing the logical relationships between the components within the circuit. In another implementation, the hardware circuit can be implemented using a PLD, such as an FPGA, which can include a large number of logic gates. The connection relationships between the logic gates are configured through configuration files to implement the functions of some or all of the above units. All units in the above device can be implemented entirely by a processor calling software, entirely by hardware circuits, or partially by a processor calling software with the remaining parts implemented by hardware circuits.

[0131] In this application embodiment, a processor is a circuit with signal processing capabilities. In one implementation, the processor can be a circuit with instruction reading and execution capabilities, such as a CPU, microprocessor, GPU, or DSP. In another implementation, the processor can implement certain functions through the logical relationships of hardware circuits. These logical relationships are fixed or reconfigurable. For example, the processor may be a hardware circuit implemented as an ASIC or PLD, such as an FPGA. In a reconfigurable hardware circuit, the process of the processor loading a configuration document and configuring the hardware circuit can be understood as the processor loading instructions to implement the functions of some or all of the above units. Furthermore, it can also be a hardware circuit designed for artificial intelligence, which can be understood as an ASIC, such as an NPU, TPU, or DPU.

[0132] As can be seen, each unit in the above device can be one or more processors (or processing circuits) configured to implement the above methods, such as: CPU, GPU, NPU, TPU, DPU, microprocessor, DSP, ASIC, FPGA, or a combination of at least two of these processor forms.

[0133] Furthermore, the units in the above devices can be integrated in whole or in part, or they can be implemented independently. In one implementation, these units are integrated together and implemented in the form of a System-on-Chip (SoC). The SoC may include at least one processor for implementing any of the above methods or implementing the functions of the units in the device. The at least one processor may be of different types, such as CPU and FPGA, CPU and artificial intelligence processor, CPU and GPU, etc.

[0134] Exemplary electronic devices Another embodiment of this application also proposes an electronic device. See [link to relevant documentation]. Figure 6 As shown, the electronic device may include: a memory 600 and a processor 610; wherein, the memory 600 is connected to the processor 610 and is used to store programs; the processor 610 is used to implement the stamping process knowledge retrieval method disclosed in any of the above embodiments by running the programs stored in the memory 600.

[0135] Specifically, the aforementioned electronic device may also include: a bus, a communication interface 620, an input device 630, and an output device 640.

[0136] The processor 610, memory 600, communication interface 620, input device 630, and output device 640 are interconnected via a bus. Among them: A bus can include a pathway for transmitting information between various components of a computer system.

[0137] The processor 610 can be a general-purpose processor, such as a general-purpose central processing unit (CPU), a microprocessor, etc., or an application-specific integrated circuit (ASIC), or one or more integrated circuits used to control the execution of the program of the present application. It can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), an off-the-shelf programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0138] The processor 610 may include a main processor, as well as a baseband chip, modem, etc.

[0139] The memory 600 stores a program for executing the technical solution of this application, and may also store an operating system and other key business functions. Specifically, the program may include program code, which includes computer operation instructions. More specifically, the memory 600 may include read-only memory (ROM), other types of static storage devices capable of storing static information and instructions, random access memory (RAM), other types of dynamic storage devices capable of storing information and instructions, disk storage, flash memory, etc.

[0140] Input device 630 may include a device for receiving user input data and information, such as a keyboard, mouse, camera, scanner, light pen, voice input device, touch screen, pedometer, or gravity sensor.

[0141] Output device 640 may include devices that allow information to be output to a user, such as a display screen, printer, speaker, etc.

[0142] The communication interface 620 may include a device that uses any transceiver to communicate with other devices or communication networks, such as Ethernet, Radio Access Network (RAN), Wireless Local Area Network (WLAN), etc.

[0143] The processor 610 executes the program stored in the memory 600 and calls other devices, which can be used to implement each step of any of the stamping process knowledge retrieval methods provided in the above embodiments of this application.

[0144] This application also proposes a chip, which includes a processor and a data interface. The processor reads and runs a program stored in the memory through the data interface to execute any of the stamping process knowledge retrieval methods provided in the above embodiments. For the specific processing procedure and its beneficial effects, please refer to the above-described embodiments of the stamping process knowledge retrieval method.

[0145] Exemplary computer program products and storage media In addition to the methods and devices described above, embodiments of this application may also be computer program products, which include computer program instructions that, when executed by a processor, cause the processor to perform the steps in the stamping process knowledge retrieval method according to various embodiments of this application as described in any of the above embodiments of this specification.

[0146] The computer program product can be written in any combination of one or more programming languages ​​to perform the operations of the embodiments of this application. The programming languages ​​include object-oriented programming languages ​​such as Java and C++, as well as conventional procedural programming languages ​​such as C or similar languages. The program code can be executed entirely on the user's computing device, partially on the user's computing device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0147] Furthermore, embodiments of this application may also be storage media storing a computer program, which is executed by a processor through steps in the stamping process knowledge retrieval method according to various embodiments of this application as described in any of the foregoing embodiments of this specification.

[0148] For the foregoing method embodiments, in order to simplify the description, they are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, because according to this application, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0149] It should be noted that the various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For apparatus embodiments, since they are basically similar to method embodiments, the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments.

[0150] The steps in the methods of the various embodiments of this application can be adjusted, merged, or deleted in order according to actual needs, and the technical features described in each embodiment can be replaced or combined.

[0151] The modules and sub-modules in the apparatus and terminal in the various embodiments of this application can be merged, divided, and deleted according to actual needs.

[0152] It should be understood that the disclosed terminals, devices, and methods can be implemented in other ways, given the several embodiments provided in this application. For example, the terminal embodiments described above are merely illustrative. For instance, the division of modules or sub-modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple sub-modules or modules may be combined or integrated into another module, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices, or modules, and may be electrical, mechanical, or other forms.

[0153] The modules or submodules described as separate components may or may not be physically separate. The components that constitute a module or submodule may or may not be physical modules or submodules; that is, they may be located in one place or distributed across multiple network modules or submodules. Some or all of the modules or submodules can be selected to achieve the purpose of this embodiment's solution, depending on actual needs.

[0154] Furthermore, the functional modules or sub-modules in the various embodiments of this application can be integrated into one processing module, or each module or sub-module can exist physically separately, or two or more modules or sub-modules can be integrated into one module. The integrated modules or sub-modules described above can be implemented in hardware or in the form of software functional modules or sub-modules.

[0155] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0156] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly by hardware, a software unit executed by a processor, or a combination of both. The software unit can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

[0157] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0158] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for retrieving knowledge about stamping processes, characterized in that, include: Receive a stamping process knowledge retrieval request and determine the query text carried by the stamping process knowledge retrieval request; Generate at least one extended query set based on the core terms of the query text; Determine the weight information that matches the target retrieval scenario corresponding to the stamping process knowledge retrieval request, wherein the weight information characterizes the importance of different extended query sets in the at least one extended query set; Based on the weight information and the at least one extended query set, the knowledge documents in the stamping process knowledge base are retrieved to generate the retrieval results of the stamping process knowledge retrieval request.

2. The method according to claim 1, characterized in that, The step of generating at least one extended query set based on the core terms of the query text includes: The query text is optimized to generate a target query text, which consists of process terms of at least one process category; At least one extended query set is generated based on the core terms of the target query text, wherein the core terms include process terms of the at least one process category.

3. The method according to claim 2, characterized in that, The step of generating at least one expanded query set based on the core terms of the target query text includes: Determine at least one semantic relationship of the core terms of the target query text, wherein the at least one semantic relationship includes at least one of term variant relationship, concept hierarchy relationship, and question parameter causal relationship; Generate an extended query set corresponding to each of the at least one semantic relations.

4. The method according to claim 1, characterized in that, The determination of the weight information matching the target retrieval scenario corresponding to the stamping process knowledge retrieval request includes: Determine the target retrieval scenario corresponding to the stamping process knowledge retrieval request, whereby the target retrieval scenario represents the urgency of the retrieval. Query the pre-set scene correction information configuration table to determine the scene correction information corresponding to the target retrieval scene; Based on the pre-set basic weight information and the scene correction information, weight information matching the target retrieval scene is generated, and the weight information includes the weight of each of the at least one extended query set.

5. The method according to claim 1, characterized in that, The step of retrieving knowledge documents in the stamping process knowledge base based on the weight information and the at least one extended query set, and generating retrieval results for the stamping process knowledge retrieval request, includes: Based on the weight information and the at least one extended query set, the matching degree of knowledge documents in the stamping process knowledge base is calculated; the matching degree of knowledge documents represents the degree of matching between the knowledge documents and the query text. Identify the first knowledge document whose matching degree is greater than the preset target matching degree; Obtain the second knowledge document associated with the first knowledge document from the stamping process knowledge base; The search results for the stamping process knowledge retrieval request are generated based on the first knowledge document and the second knowledge document.

6. The method according to claim 5, characterized in that, The step of calculating the matching degree of knowledge documents in the stamping process knowledge base based on the weight information and the at least one extended query set includes: Based on the weight information and the at least one extended query set, calculate the first matching degree and the second matching degree of the knowledge documents in the stamping process knowledge base; the first matching degree represents the character matching degree, and the second matching degree represents the semantic similarity. The matching degree of the knowledge document is calculated based on the first matching degree and the second matching degree of the knowledge document.

7. The method according to claim 6, characterized in that, The step of calculating the first matching degree and the second matching degree of knowledge documents in the stamping process knowledge base based on the weight information and the at least one extended query set includes: Calculate the character matching degree between the knowledge documents in the stamping process knowledge base and each of the at least one extended query set; The first matching degree of the knowledge document is calculated based on the weight information and the character matching degree between the knowledge document and each of the extended query sets. Calculate the semantic similarity between the knowledge documents in the stamping process knowledge base and each of the at least one extended query set; The second matching degree of the knowledge document is calculated based on the weight information and the semantic similarity between the knowledge document and each of the extended query sets.

8. The method according to claim 5, characterized in that, Also includes: The first knowledge documents are sorted according to a pre-set sorting rule to generate a sequence of first knowledge documents; The sorting rule is related to the matching degree of the first knowledge document, or the sorting rule is related to the decision logic information category sequence; Sort the second knowledge documents associated with the first knowledge document to generate a second knowledge document sequence for each first knowledge document; The retrieval results are displayed based on the first knowledge document sequence and the second knowledge document sequence for each of the first knowledge documents in the first knowledge document sequence.

9. A stamping process knowledge retrieval device, characterized in that, include: The query text determination unit is used to receive a stamping process knowledge retrieval request and determine the query text carried by the stamping process knowledge retrieval request. A query expansion unit is used to generate at least one expanded query set based on the core terms of the query text. The weight information determination unit is used to determine the weight information that matches the target retrieval scenario corresponding to the stamping process knowledge retrieval request, wherein the weight information characterizes the importance of different extended query sets in the at least one extended query set; The retrieval result generation unit is used to retrieve knowledge documents in the stamping process knowledge base based on the weight information and the at least one extended query set, and generate retrieval results for the stamping process knowledge retrieval request.

10. An electronic device, characterized in that, Including memory and processor; The memory is connected to the processor and is used to store programs; The processor is used to implement the stamping process knowledge retrieval method as described in any one of claims 1 to 8 by running the program in the memory.

Images