Business processing method and device, equipment and storage medium

Abstract

The embodiment of the invention discloses a business processing method and device, equipment and a storage medium. The method provided by the embodiment of the invention comprises the following steps: performing feature extraction on received service operation information to obtain a feature vector of the service operation information; performing information expansion on the service operation information according to the feature vector of the service operation information, and generating supplementary information of the service operation information; and based on the supplementary information and the business operation information, generating and executing a target business operation instruction, and returning and displaying an execution result of the target business operation instruction. According to the embodiment of the invention, feature extraction is carried out on the received service operation information, the operation intention of the user can be more intelligently understood, the supplementary information is generated after information expansion, the content input by the user can be effectively reduced, and the complexity of user operation is reduced. And on the basis of the supplementary information and the service operation information, generating the target service operation instruction, so that the accuracy and intelligent operation efficiency of service operation can be improved, and the overall efficiency of a service system is improved.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of artificial intelligence, and in particular to business processing methods, devices, equipment and storage media. Background Art

[0002] In the existing business system, users mainly operate in the traditional click interaction mode, or perform simple voice operations through the voice recognition wake-up function. However, these interaction modes have certain limitations in terms of convenience and intelligence, which are specifically reflected in the following aspects:

[0003] Users' business operations are highly dependent on their familiarity with the business system. Especially in complex business scenarios, some operation processes may be cumbersome and users need to remember specific operation paths, which increases the difficulty of use and learning cost for users. In addition, some business systems require users to open and log in to the system before they can operate. This fixed process design lacks flexibility and adaptability, and it is difficult to meet users' needs for efficient operations.

[0004] As for the voice assistants in business systems, the current voice assistants have relatively simple functions and can only execute preset fixed instructions, making it difficult to support complex business scenarios; they lack the ability to handle multiple rounds of conversations and cannot dynamically generate accurate responses or instructions based on the context, thus failing to meet users' needs for intelligent and efficient interaction; in short, the deficiencies in flexibility, intelligence, and efficiency of existing interaction models limit the user experience and overall efficiency of business systems. Summary of the invention

[0005] Based on the above problems, the embodiments of the present application provide business processing methods, devices, equipment and storage media, the purpose of which is to improve the flexibility, intelligent interaction level and user experience of the business system.

[0006] In a first aspect, an embodiment of the present application provides a service processing method, including:

[0007] Extracting features from the received business operation information to obtain a feature vector of the business operation information;

[0008] Expanding the business operation information according to the feature vector of the business operation information to generate supplementary information of the business operation information;

[0009] generating a target business operation instruction based on the supplementary information and the business operation information;

[0010] Execute the target business operation instruction, and return and display the execution result of the target business operation instruction.

[0011] In one embodiment, the step of expanding the business operation information according to the feature vector of the business operation information to generate supplementary information of the business operation information includes:

[0012] Determining the similarity between each business knowledge text in a preset knowledge base and the business operation information according to the feature vector of the business operation information;

[0013] Determine the business knowledge text with the highest similarity to the business operation information;

[0014] Based on the business knowledge text with the highest similarity, the supplementary information of the business operation information is generated in a predetermined format.

[0015] In one embodiment, if the similarity between each of the business knowledge texts in the preset knowledge base and the business operation information is less than a preset threshold, generating a target business operation instruction based on the supplementary information and the business operation information includes:

[0016] Generate supplementary questions based on the supplementary information and feed back to the user;

[0017] Determining the user intention of the business operation information based on the user's feedback result to the supplementary question;

[0018] A target business operation instruction is generated based on the user intention of the business operation information, the business operation information and the supplementary information.

[0019] In one embodiment, the generating a target business operation instruction based on the user intention of the business operation information, the business operation information and the supplementary information includes:

[0020] Based on the user intention of the business operation information, the business operation information and the supplementary information, generating a guide sentence using a large language model guide template;

[0021] The guiding sentence is input into a large language model to obtain a target business operation instruction corresponding to the guiding sentence.

[0022] In one embodiment, the large language model guidance template includes received business operation information, user authority information, supplementary information, corrected business operation information and format requirements.

[0023] In one embodiment, before generating a supplementary question based on the supplementary information and feeding it back to the user, the method further includes:

[0024] Generate a business operation information parsing result using a preset business operation information parsing template according to the similarity between each of the business knowledge texts in the preset knowledge base and the business operation information and the supplementary information;

[0025] According to the analysis result of the business operation information, it is determined whether it is necessary to generate supplementary questions according to the supplementary information and feed back to the user.

[0026] In one embodiment, the generating a target business operation instruction based on the supplementary information and the business operation information includes:

[0027] The static text features in the supplementary information and the business operation information are converted into corresponding target business operation instructions; the static text features include one of a business operation noun description, a business data noun description, and a time description related to the content of the business operation information.

[0028] In one embodiment, the method further comprises:

[0029] Establishing an association relationship among the business operation information, the supplementary information, the target business operation instruction, and the execution result;

[0030] Based on the association relationship, knowledge extraction is performed on the business operation information, the supplementary information, the target business operation instruction and the execution result, and a business knowledge text is generated and stored in a preset knowledge base.

[0031] In a second aspect, an embodiment of the present application further provides a service processing device, including:

[0032] The first module is used to extract features from the received business operation information to obtain a feature vector of the business operation information;

[0033] The second module is used to expand the business operation information according to the feature vector of the business operation information to generate supplementary information of the business operation information;

[0034] A third module is used to generate a target business operation instruction based on the supplementary information and the business operation information;

[0035] The fourth module is used to execute the target business operation instruction, and return and display the execution result of the target business operation instruction.

[0036] In a third aspect, an embodiment of the present application further provides a computer device, including:

[0037] CPU, memory, input and output interfaces;

[0038] The memory is a short-term storage memory or a persistent storage memory;

[0039] The central processing unit is configured to communicate with the memory and execute instruction operations in the memory to perform any one of the above-mentioned business processing methods.

[0040] In a fourth aspect, an embodiment of the present application further provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, any one of the above-mentioned business processing methods is executed.

[0041] It can be seen from the above technical solutions that the embodiments of the present application have the following advantages:

[0042] The embodiment of the present application extracts features from the received business operation information, so that the system can more intelligently understand the user's operation intention. The process of generating supplementary information after information expansion can effectively reduce the content input by the user and reduce the complexity of the user's operation. Based on the supplementary information and the business operation information, the target business operation instructions are generated, which can improve the accuracy and intelligent operation efficiency of the business operation and enhance the overall efficiency of the business system. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0044] Figure 1 A flowchart of a business processing method provided in an embodiment of the present application;

[0045] Figure 2 A flowchart of another business processing method provided in an embodiment of the present application;

[0046] Figure 3 A schematic diagram of the structure of a service processing device provided in an embodiment of the present application;

[0047] Figure 4 A schematic diagram of the structure of a computer device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0048] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0049] The various embodiments of the present application are described in further detail below in conjunction with the accompanying drawings.

[0050] The present application embodiment provides a service processing method, such as Figure 1 As shown, the method includes steps S101-S104.

[0051] S101: extracting features from received business operation information to obtain a feature vector of the business operation information;

[0052] Specifically, business operation information sent in text form can be received, and the form of the text includes but is not limited to voice, text and other forms. For example, a user can send business operation information in text form through a mobile application or a web page, or the user can also send business operation information in voice form. After the user sends the business operation information, the server can parse and obtain the business operation information in text form such as voice and text. When the text form of the acquired business operation information is voice, the server can also convert the business operation information in voice form into business operation information in text form. Here, the business operation information includes but is not limited to: "How do I print my account statement", "What is the profit of the business department", etc.

[0053] It can be understood that the user can input data in at least one type of form such as text, image, voice or video, and the system recognizes the directly input data to obtain natural language content. The client can receive the user input data and send it to the computer device of the embodiment of the present disclosure, and the computer device obtains the natural language content according to the input data.

[0054] If the request initiated by the user contains text (such as a query statement or action description), text feature extraction can be performed, such as decomposing the text into words or phrases, and then using TF-IDF, Word2Vec, BERT and other methods to vectorize the text. Furthermore, features can be extracted based on the context of the business operation information, such as the user's historical behavior pattern and the environment information that initiated the business operation information (such as device type, browser information, operating system, etc.).

[0055] S102: Expanding the business operation information according to the feature vector of the business operation information to generate supplementary information of the business operation information;

[0056] In some feasible embodiments, high-dimensional data (such as text, pictures, and audio) can be mapped to a low-dimensional space through an embedding processing technique to obtain an embedding vector. An embedding vector is a vector composed of real numbers, which is used to represent the input high-dimensional data as a point in a continuous numerical space.

[0057] After obtaining the feature vector of the business operation information, the retrieval augmented generation (RAG) technology can be used to perform semantic matching based on the feature vector of the business operation information to find the most relevant business knowledge text. At the same time, the system will reconstruct and optimize the semantic description of the business operation information. Specifically, after initially retrieving the relevant business knowledge text, the system can further supplement the business knowledge text of the business operation information by querying the background of the business operation information or expanding the semantic scope, and generate a more comprehensive description of the business operation information to provide support for subsequent processing. In this way, even if the user's question is not completely accurate, the system can still identify and find the most relevant knowledge points, thereby ensuring that the user's true business operation intention can be accurately identified based on the received business operation information.

[0058] S103: Generate a target business operation instruction based on the supplementary information and the business operation information;

[0059] Based on the business operation information and supplementary information, the system needs to generate target business operation instructions according to predefined rules, model reasoning or intelligent algorithms. The target business operation instruction can be a system operation instruction for the system to perform a specific computing task, data modification or scheduling task, a business process instruction for automatically processing a request based on business rules, and a task allocation instruction for distributing operation requests to appropriate services or personnel in the scenario of task allocation.

[0060] For example, suppose there is an intelligent customer service system, and a user enters a query into the system: "When will my order arrive?" This question is a type of business operation information, which expresses the user's query demand: to know the arrival time of the order.

[0061] In order to generate an accurate response, the system needs additional information to supplement and support this query:

[0062] For example, the system first needs to know the user's order number, then further needs to obtain the user's account information, and then retrieve the order data related to the user (such as the most recent completed order). Furthermore, the system also needs to query the logistics status of the order, such as the current delivery progress of the order. The above additional information can be collectively referred to as supplementary information. By obtaining supplementary information and combining it with business operation information, the system can better understand the user's business operation needs: query the estimated arrival time of the most recent order.

[0063] In other feasible embodiments, common business operation information may be associated with its supplementary information and generated instructions and stored to form a knowledge base; when the system encounters similar business operation information again, the corresponding instructions may be directly retrieved from the knowledge base and quickly executed.

[0064] After obtaining sufficient supplementary information, the system needs to generate actual target business operation instructions to perform subsequent tasks. For example, the generated instructions can be expressed in words as: "retrieve the order records in the user account and obtain the logistics information of order number 1X3Y", "call the logistics API to query the delivery progress of order number 1X3Y", "calculate and return the estimated arrival time of order number 1X3Y".

[0065] S104: Execute the target business operation instruction, and return and display the execution result of the target business operation instruction.

[0066] In the stage of executing the target business operation instruction, the reasoning ability of the Large Language Model (LLM) can be used to combine the received business operation request and / or supplementary information, and the target business operation instruction can be executed by the LLM executor. Here, the target business operation instruction includes SQL data acquisition instruction, document action execution (calling out / operating existing documents), event triggering (calling external tools or services), etc.

[0067] The final execution result includes the command execution result (pop-up chart, etc.) and the session return result (text / voice interactive answer). For example, in actual applications, the execution result can be a pop-up window showing the sales summary details of a certain time period, copying a certain document, or a responsive text or voice broadcast (such as asking someone about their sales this month).

[0068] The embodiment of the present application extracts features from the received business operation information, so that the system can more intelligently understand the user's operation intentions. Generating supplementary information after information expansion can effectively reduce the content of user input and reduce the complexity of user operations. Based on the supplementary information and the business operation information, the target business operation instructions are generated, which can improve the accuracy and intelligent operation efficiency of business operations and enhance the overall performance of the business system.

[0069] In order to improve the accuracy and contextual relevance of the execution results output by the large model, in one embodiment, step S102 specifically includes: determining the similarity between each business knowledge text in the preset knowledge base and the business operation information based on the feature vector of the business operation information; determining the business knowledge text with the highest similarity to the business operation information; and generating supplementary information of the business operation information in a predetermined format based on the business knowledge text with the highest similarity.

[0070] Specifically, the preset knowledge base includes the following four types of business knowledge texts:

[0071] (1) Frequently asked questions and answers: This type of knowledge text includes business operation information and its execution results that users may commonly encounter (in the scenario where users interact with LLM, business operation information and its execution results can also be called questions and their answers). For example, a user may ask: "Please obtain the sales summary of Business Department 1 in the past week." For such questions, the corresponding answer not only includes the final formatted output result of "Sales summary of Business Department 1 in the past week", but also includes references to the relevant business, technical and system metadata involved in the question. For example, the named entities in the question (such as names, places, organizations) should be linked to known business, technical and system metadata entries in the knowledge base to explain the basis of the output results. In addition, after the feedback of the business question is given, the system corresponding to the business question can provide related questions to further guide the subsequent user's business operations and generate expected output.

[0072] (2) Business / technical metadata knowledge:

[0073] Business metadata: refers to the business knowledge text that describes the data and its structure, content, context, and usage. Business metadata can help better utilize data for business analysis and decision-making. In the embodiments of the present application, business metadata mainly refers to the actual meaning of business fields, field usage restrictions (such as field enumeration range), and other content. For example, the parent-child relationship between the sales order header and the sales order details and their field associations, and the references to basic information (such as customer, department, material, and other fields) all fall into the category of business metadata.

[0074] Technical metadata: refers to the technical details that describe data storage, processing, and transmission. It is mainly aimed at IT professionals and technical developers to support system development, integration, management, and optimization, and provide the information required for system development, integration, management, and optimization. For example, technical metadata includes details such as the field name, field type, field length, maximum and minimum values ​​of the business system table, which can help generate technical content such as SQL queries and operation instructions. Technical metadata can be expressed in a structured manner such as JSON, which facilitates rapid matching and automatic acquisition of relevant system knowledge, ensuring that operations can be performed accurately and quickly when generating SQL or calling interfaces.

[0075] (3) System metadata knowledge:

[0076] For business system interface: System metadata related to the business system interface includes the name and function of the business system interface, calling parameters, parameter type, parameter value range, parameter data format, etc.

[0077] For business system actions: Business system action formatting defines the business actions supported by the business system (such as opening a document / list / report, closing a document / list / report, reviewing / reversing a review, etc.). These business system actions are usually exposed through interfaces, or external call interfaces are provided through SDKs.

[0078] Business system events: refers to the event mechanism supported by the business system (such as field changes, approval completion and other system events). The system metadata for business system events includes event name, event effect, event parameters and other data. These system metadata can help LLM automatically perform corresponding operations or responses based on system events.

[0079] (4) Conversation history integration: Record the current conversation history and use previous conversation information when implementing the current conversation history to improve the quality of the current conversation and the continuity of the service. By reviewing the historical conversation content, the system can build a more natural and smooth multi-round conversation experience. The automatically integrated knowledge points can be seamlessly combined with the current conversation, ensuring that the system accurately understands the user's needs and adjusts the response strategy in a timely manner.

[0080] In practical applications, a vector index can be set for the business knowledge text in the preset knowledge base, so that the business knowledge text with the highest similarity to the business operation information can be found directly based on the feature vector of the business operation information and the vector index of each business knowledge text in the preset knowledge base, and then the supplementary information of the business operation information can be generated according to the predetermined format.

[0081] In one embodiment, if the similarity between each business knowledge text and the business operation information in a preset knowledge base is less than a preset threshold, a target business operation instruction is generated based on the supplementary information and the business operation information, including: generating a supplementary question based on the supplementary information and feeding it back to the user; determining the user intention of the business operation information based on the user's feedback results on the supplementary question; and generating the target business operation instruction based on the user intention of the business operation information, the business operation information and the supplementary information.

[0082] Assume that when the system matches business operation information with business knowledge text in the knowledge base, if the similarity is lower than a preset threshold, it means that the business operation information may lack some key information, or there is not enough relevant business operation knowledge in the preset knowledge base. To solve this problem, the system needs to generate a supplementary question based on the most relevant supplementary information currently matched and feedback it to the user.

[0083] Specifically, suppose the business operation information entered by the user is "this week's sales summary", but the system does not find directly matching business knowledge in the knowledge base. At this time, the system analyzes the most relevant supplementary information (for example: find the data table "sales order" related to the business operation information, and the table contains fields such as quantity, sales amount, profit, etc.), and then can generate a supplementary question: "Sales order contains fields such as quantity, sales amount, profit, etc. Do you need to count all these fields?" This question is then fed back to the user to further determine the user's specific needs. In this way, the system can not only further clarify the user's intention through interactive feedback when the knowledge base is insufficient, but also ensure that the final generated business operation instructions can accurately meet the user's needs.

[0084] Furthermore, in one embodiment, based on the user intent, business operation information and supplementary information of the business operation information, a target business operation instruction is generated, including: based on the user intent, business operation information and supplementary information of the business operation information, a guide sentence is generated using a large language model guide template; the guide sentence is input into the large language model to obtain a target business operation instruction corresponding to the guide sentence.

[0085] In an embodiment of the present application, the main function of the large language model is to semantically understand the input natural language content and generate corresponding replies or instruction content. The large language model can adopt the ChatGPT series model, or it can be replaced by other models with semantic understanding and generation capabilities. The guide sentence (Prompt) is a prompt sentence input into the large language model, which includes business operation information, context information and keywords entered by the user, and is used to guide the model to understand the user's intentions and generate expected answers. Prompt can not only prompt the large language model how to process the received business operation information, but also specify the format and framework of the generated content.

[0086] Specifically, the system generates a guide statement based on business operation information and supplementary information. For example, when the user enters "view this week's sales summary", the prompt can be generated as: "The user needs to query this week's sales summary data. Sales orders contain fields such as quantity, sales amount, and profit. Please generate an SQL query statement to count the quantity, sales amount, and profit of sales orders this week." Enter the above prompt into LLM, and LLM will generate operation instructions that meet user needs based on the content of the prompt. For example, the generated instruction can be an SQL query statement:

[0087]

[0088] Alternatively, the target business operation instruction may be an operation of calling an interface, for example, calling the interface API / getSalesSummary, and passing in parameters: start_date = '2025-01-01', end_date = '2025-01-07'.

[0089] It can be seen that through the guidance of supplementary questions and prompts, LLM can automatically generate complex operation instructions without the need for users to master all business data to perform business operations. The embodiment of the present application combines user intentions, business operation information and supplementary information to generate guide statements and use LLM to generate target business operation instructions, which can not only accurately understand and meet user needs, but also automatically process complex business scenarios and generate accurate SQL queries, interface calls or report generation instructions, thereby improving the interaction efficiency and accuracy of the business system.

[0090] Further, in a feasible embodiment, the large language model guidance template includes received business operation information, user authority information, supplementary information, corrected business operation information and format requirements.

[0091] The "received business operation information" can be set in the large language model guidance template to ensure that LLM always generates response content around the user's original needs and avoids deviation from the topic during multiple rounds of supplementary interactions. Then, the "LLM question", that is, the "corrected business operation information", can be set to further improve the business operation information initiated by the user, so that LLM can more accurately identify user needs and output accurate response content.

[0092] Exemplarily, the large language model guidance template can be set as follows:

[0093]

[0094]

[0095] Among them, relevant business knowledge is used for LLM to understand the specific data relationship related to business operation information, that is, the "supplementary information" mentioned in the aforementioned embodiment; relevant business table and field metadata are used to specifically describe the table structure, field restrictions and other metadata in the system; variable conversion rules are used to define how the time period or variables involved in the user's natural language are mapped to system-recognizable values; goals are used to clarify the specific goals of LLM-generated content to prevent the model from deviating from the main task; LLM question supplement guidance is used to guide the model on how to further supplement the content when the information provided is insufficient; LLM answer format requirements are used to specify the format of generated content (such as structured data, specific forms of graphic reports, etc.).

[0096] The embodiment of the present application uses a large language model to guide the template to supplement context information, variable rules and other content, so that the input and output of the model become more standardized, further helping the LLM to understand and handle problems more accurately, and ensuring that the model's answer content complies with the user's authority scope and business specifications.

[0097] In order to optimize the interactive experience and avoid generating erroneous instructions due to misunderstanding or incomplete information, in some possible embodiments, before generating supplementary questions based on the supplementary information and feeding back to the user, the method also includes: generating business operation information parsing results using a preset business operation information parsing template based on the similarity and supplementary information between each business knowledge text and the business operation information in the preset knowledge base; and determining whether it is necessary to generate supplementary questions based on the supplementary information and feed back to the user based on the business operation information parsing results.

[0098] It can be understood that after receiving the input business operation information, the processor of the large language model will retrieve the relevant business knowledge text in the preset knowledge base through the business operation information to generate supplementary information. Subsequently, the system combines the business operation information with the supplementary information according to the preset large language model guidance template to generate a guidance sentence. The generated guidance sentence is passed to the executor of the large language model, and the executor generates the target business operation instruction according to the instruction information in the guidance sentence. The generated target instruction will be further processed, and its execution result will be formatted as a business operation information parsing result according to the business operation information parsing template (also called "answer template").

[0099] Exemplarily, the business operation information parsing template (answer template) may be configured as follows:

[0100]

[0101] Here, questions are used to retain the original business operation information entered by the user, ensuring that the entire processing flow is centered on user needs. "Whether metadata knowledge needs to be supplemented" is used to identify whether the current analysis needs to be supplemented with additional information, such as field definitions and time ranges, to avoid generating erroneous instructions due to insufficient information, and to optimize the interaction process. "Is the goal complete" is used to determine whether the user's intention is clear, such as detecting whether fields and conditions are missing, to ensure that all key information is clear before generating instructions. "Related knowledge" is used to record the relevant knowledge sources used in the analysis process, provide a basis for generating results, and enhance transparency. "Knowledge extraction for this question" is used to format and summarize the key information extracted in the current session, so as to facilitate the subsequent use of this knowledge to build multiple rounds of dialogue or further optimize the interaction.

[0102] In the embodiment of the present application, the system records the completeness of the user input through the business operation information parsing template, and generates supplementary questions when necessary to avoid erroneous instructions due to insufficient information; if the parsing result shows that the information input by the user is sufficiently complete (for example, the business operation information can find relevant knowledge points in the knowledge base with a similarity exceeding a preset threshold), or the user feedback confirms that the supplementary information is correct, the system can skip the supplementary question link and directly generate business instructions, thereby improving interaction efficiency and reducing user interaction costs; the business operation information parsing template records each link of the problem, parsing process and generated results, enhances the traceability of the execution results, and facilitates subsequent optimization and error diagnosis.

[0103] In one embodiment, target business operation instructions are generated based on supplementary information and business operation information, including: converting static text features in the supplementary information and business operation information into corresponding target business operation instructions; the static text features include one of a business operation noun description, a business data noun description, and a time description related to the content of the business operation information.

[0104] Here, static text features refer to the content in the business operation information that does not involve dynamic changes, and are key elements that directly describe specific operations, data or time. Among them, the business operation noun description can be a description of the type of operation that the user needs to perform, such as selection, statistics, comparison, summary, screening, query, calculation, etc. The business data noun description refers to the data entity or field involved in the description of the operation, such as "sales amount", "order quantity", "profit", etc. The time description is a description of the time range or time point of the operation, such as "this week", "January 1, 2025 to January 7, 2025", etc. After these static text features are extracted, they can be converted into specific operation instructions, such as SQL query statements, API call commands, etc., combined with system setting rules or templates, to ensure that the generated operation instructions are consistent with the user's input requirements and are executable.

[0105] In order to transform the experience accumulated during the interaction and execution process into knowledge, thereby providing support for subsequent tasks and gradually optimizing the intelligence level of the system. In one embodiment, the method further includes: constructing an association relationship between business operation information, supplementary information, target business operation instructions, and execution results; based on the association relationship, performing knowledge extraction on the business operation information, supplementary information, target business operation instructions, and execution results, generating a business knowledge text and storing it in a preset knowledge base.

[0106] The system associates the above business operation information, supplementary information, target business operation instructions and execution results by mapping, matching or serializing, extracts and summarizes the key information in the entire interaction and execution process, and generates structured business knowledge text. The results of knowledge extraction are formatted and stored according to the knowledge structure extraction template. For example, the knowledge structure extraction template is as follows:

[0107]

[0108]

[0109] Here, "final instructions / answers" are used to describe the target business operation instructions or final answers generated by the system to meet user needs, providing examples of instruction generation for subsequent similar questions.

[0110] The embodiment of the present application extracts key information from the interaction and stores it in a structured form as knowledge text, which can be directly reused in subsequent problem processing, helping the system to continuously optimize its ability to generate answers and instructions and improve user efficiency.

[0111] In order to implement the service processing method of the embodiment of the present application, the embodiment of the present application also provides a service processing device, such as Figure 3 As shown, the device comprises:

[0112] The first module 301 is used to extract features from received business operation information to obtain a feature vector of the business operation information;

[0113] The second module 302 is used to expand the business operation information according to the feature vector of the business operation information to generate supplementary information of the business operation information;

[0114] The third module 303 is used to generate a target business operation instruction based on the supplementary information and the business operation information;

[0115] The fourth module 304 is used to execute the target business operation instruction, and return and display the execution result of the target business operation instruction.

[0116] In one embodiment, the second module 302 is specifically used for:

[0117] Determining the similarity between each business knowledge text in a preset knowledge base and the business operation information according to the feature vector of the business operation information;

[0118] Determine the business knowledge text with the highest similarity to the business operation information;

[0119] Based on the business knowledge text with the highest similarity, the supplementary information of the business operation information is generated in a predetermined format.

[0120] In one embodiment, if the similarity between each of the business knowledge texts in the preset knowledge base and the business operation information is less than a preset threshold, the third module 303 is specifically configured to:

[0121] Generate supplementary questions based on the supplementary information and feed back to the user;

[0122] Determining the user intention of the business operation information based on the user's feedback result to the supplementary question;

[0123] A target business operation instruction is generated based on the user intention of the business operation information, the business operation information and the supplementary information.

[0124] In one embodiment, the third module 303 is specifically used for:

[0125] Based on the user intention of the business operation information, the business operation information and the supplementary information, generating a guide sentence using a large language model guide template;

[0126] The guiding sentence is input into a large language model to obtain a target business operation instruction corresponding to the guiding sentence.

[0127] In one embodiment, the large language model guidance template includes received business operation information, user authority information, supplementary information, corrected business operation information and format requirements.

[0128] In one embodiment, the device further includes: a fifth module; the fifth module is specifically configured to:

[0129] Generate a business operation information parsing result using a preset business operation information parsing template according to the similarity between each of the business knowledge texts in the preset knowledge base and the business operation information and the supplementary information;

[0130] According to the analysis result of the business operation information, it is determined whether it is necessary to generate supplementary questions according to the supplementary information and feed back to the user.

[0131] In one embodiment, the third module 303 is specifically used for:

[0132] The static text features in the supplementary information and the business operation information are converted into corresponding target business operation instructions; the static text features include one of a business operation noun description, a business data noun description, and a time description related to the content of the business operation information.

[0133] In one embodiment, the fifth module is specifically used for:

[0134] Establishing an association relationship among the business operation information, the supplementary information, the target business operation instruction, and the execution result;

[0135] Based on the association relationship, knowledge extraction is performed on the business operation information, the supplementary information, the target business operation instruction and the execution result, and a business knowledge text is generated and stored in a preset knowledge base.

[0136] It should be noted that: the above embodiment provides that the business processing device performs business processing, and only uses the division of the above program modules as an example. In actual applications, the above processing can be assigned to different program modules as needed, that is, the internal structure of the device is divided into different program modules to complete all or part of the above-described processing. In addition, the business processing device and the business processing method embodiment provided in the above embodiment belong to the same concept, and the specific implementation process is detailed in the method embodiment, which will not be repeated here.

[0137] Based on the hardware implementation of the above program modules, and in order to implement a business processing method provided in an embodiment of the present application, an embodiment of the present application further provides a computer device, such as Figure 4 As shown, the computer device 400 includes:

[0138] CPU 401, memory 402 and input / output interface 403;

[0139] The memory 402 is a temporary storage memory or a permanent storage memory;

[0140] The central processor 401 is configured to communicate with the memory 402 and execute instructions in the memory 402 to perform any one of the above-mentioned business processing methods.

[0141] Of course, in actual application, the various components in the computer device 400 are coupled together through the bus system 404. It can be understood that the bus system 404 is used to realize the connection and communication between these components. In addition to the data bus, the bus system 404 also includes a power bus, a control bus and a status signal bus. However, for the sake of clarity, Figure 4 Various buses are labeled as bus system 404 .

[0142] The memory 402 in the embodiment of the present application is used to store various types of data to support the operation of the computer device 400. Examples of such data include: any computer program used to operate on the computer device 400.

[0143] It can be understood that when the processor in the computer device described above executes the computer program, it can also implement the functions of each unit in the corresponding device embodiments described above, which will not be elaborated here. Exemplarily, the computer program can be divided into one or more modules / units, and one or more modules / units are stored in the memory and executed by the processor to complete various embodiments of the present application. One or more modules / units can be a series of computer program instruction segments capable of performing specific functions, and these instruction segments are used to describe the execution process of the computer program in the computer device. For example, the computer program can be divided into the units in the above computer device, and each unit can implement the specific functions as described in the corresponding computer device above.

[0144] The computer device can be a computing device such as a desktop computer, a notebook, a palm computer, and a cloud server. The computer device may include but is not limited to a processor and a memory. Those skilled in the art can understand that the processor and the memory are only examples of the computer device, and do not constitute a limitation on the computer device. It may include more or fewer components, or combine some components, or different components. For example, the computer device may also include input / output devices, network access devices, a bus, etc.

[0145] The processor can be a central processing unit (CPU, Central Processing Unit), or other general-purpose processors, digital signal processors (DSP, Digital Signal Processor), application specific integrated circuits (ASIC, Application Specific Integrated Circuit), field-programmable gate arrays (FPGA, Field-Programmable Gate Array), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc. The processor is the control center of the computer device, and connects various parts of the entire computer device through various interfaces and lines.

[0146] The memory can be used to store computer programs and / or modules. The processor implements various functions of the computer device by running or executing the computer programs and / or modules stored in the memory, and calling the data stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, at least one application required for a function, etc.; the data storage area can store data created according to the use of the terminal, etc. In addition, the memory can include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (SMC, Smart Media Card), a secure digital (SD, Secure Digital) card, a flash card (Flash Card), at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0147] An embodiment of the present application further provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, any one of the above-mentioned business processing methods is executed.

[0148] An embodiment of the present application also provides a computer program product on which a computer program / instruction is stored. When the computer program / instruction is executed by a processor, it is used to implement the business processing method described in the first aspect of the embodiment of the present application or any specific implementation method of the first aspect.

[0149] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0150] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0151] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0152] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0153] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program code.

Claims

1. A business processing method, characterized in that: include: Extracting features from the received business operation information to obtain a feature vector of the business operation information; Expanding the business operation information according to the feature vector of the business operation information to generate supplementary information of the business operation information; generating a target business operation instruction based on the supplementary information and the business operation information; Execute the target business operation instruction, and return and display the execution result of the target business operation instruction.

2. The service processing method according to claim 1, characterized in that: The step of expanding the business operation information according to the feature vector of the business operation information to generate supplementary information of the business operation information includes: Determining the similarity between each business knowledge text in a preset knowledge base and the business operation information according to the feature vector of the business operation information; Determine the business knowledge text with the highest similarity to the business operation information; Based on the business knowledge text with the highest similarity, the supplementary information of the business operation information is generated in a predetermined format.

3. The service processing method according to claim 2, characterized in that: If the similarity between each of the business knowledge texts in the preset knowledge base and the business operation information is less than a preset threshold, generating a target business operation instruction based on the supplementary information and the business operation information includes: Generate supplementary questions based on the supplementary information and feed back to the user; Determining the user intention of the business operation information based on the user's feedback result to the supplementary question; A target business operation instruction is generated based on the user intention of the business operation information, the business operation information and the supplementary information.

4. The service processing method according to claim 3, characterized in that: The generating a target business operation instruction based on the user intention of the business operation information, the business operation information and the supplementary information includes: Based on the user intention of the business operation information, the business operation information and the supplementary information, generating a guide sentence using a large language model guide template; The guiding sentence is input into a large language model to obtain a target business operation instruction corresponding to the guiding sentence.

5. The service processing method according to claim 4, characterized in that: The large language model guidance template includes received business operation information, user authority information, supplementary information, corrected business operation information and format requirements.

6. The service processing method according to claim 3, characterized in that: Before generating a supplementary question according to the supplementary information and feeding it back to the user, the method further includes: Generate a business operation information parsing result using a preset business operation information parsing template according to the similarity between each of the business knowledge texts in the preset knowledge base and the business operation information and the supplementary information; According to the analysis result of the business operation information, it is determined whether it is necessary to generate supplementary questions according to the supplementary information and feed back to the user.

7. The service processing method according to claim 1, characterized in that: The generating a target business operation instruction based on the supplementary information and the business operation information includes: The static text features in the supplementary information and the business operation information are converted into corresponding target business operation instructions; the static text features include one of a business operation noun description, a business data noun description, and a time description related to the content of the business operation information.

8. The service processing method according to claim 1, characterized in that: The method further comprises: Establishing an association relationship among the business operation information, the supplementary information, the target business operation instruction, and the execution result; Based on the association relationship, knowledge extraction is performed on the business operation information, the supplementary information, the target business operation instruction and the execution result, and a business knowledge text is generated and stored in a preset knowledge base.

9. A service processing device, characterized in that: include: The first module is used to extract features from the received business operation information to obtain a feature vector of the business operation information; The second module is used to expand the business operation information according to the feature vector of the business operation information to generate supplementary information of the business operation information; A third module is used to generate a target business operation instruction based on the supplementary information and the business operation information; The fourth module is used to execute the target business operation instruction, and return and display the execution result of the target business operation instruction.

10. A computer device, characterized in that: include: CPU, memory and input / output interface; The memory is a short-term storage memory or a persistent storage memory; The central processing unit is configured to communicate with the memory and execute instruction operations in the memory to perform the business processing method according to any one of claims 1 to 7.

Images