Intelligent question and answer data processing method and device, electronic equipment and storage medium

Abstract

The invention discloses an intelligent question and answer data processing method and device, electronic equipment and a storage medium. Comprising the steps of obtaining a to-be-processed text, and determining an initial regulation inquiry text corresponding to the to-be-processed text; performing business scene consistency verification on the initial regulation inquiry text to obtain a target regulation inquiry text meeting a business scene consistency requirement; and determining reference business rule data corresponding to the to-be-processed text and reference law and regulation data corresponding to the target law and regulation inquiry text, and determining a target processing result corresponding to the to-be-processed text through a pre-constructed first processing model based on the to-be-processed text, the target law and regulation inquiry text, the reference business rule data and the reference law and regulation data. According to the scheme, the processing result corresponding to the to-be-processed text is determined by determining the target regulation inquiry text and the reference data which correspond to the to-be-processed text and meet the business scene consistency verification, so that the processing result of the to-be-processed text meets the business scene consistency, and the accuracy and the compliance of the processing result are also improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular, to an intelligent question and answer data processing method, apparatus, electronic device, and storage medium. Background Art

[0002] With the continuous development of intelligent question and answer technology, the intelligent question and answer technology is applied in more and more business fields to process a large number of operation and maintenance consultation and question and answer processing requirements. Specifically, the processing result corresponding to the operation and maintenance consultation and question and answer processing requirements is determined through question information, domain knowledge, and historical experience. However, in the process of determining the processing result of a question, a processing result inconsistent with the business scenario corresponding to the question and a processing result not in line with the corresponding regulations are often obtained, resulting in inaccurate processing results corresponding to user questions. Summary of the Invention

[0003] The present invention provides an intelligent question and answer data processing method, apparatus, electronic device, and storage medium to solve the problems of inaccurate and non-compliant intelligent question and answer data processing results.

[0004] According to one aspect of the present invention, there is provided an intelligent question and answer data processing method, including:

[0005] Obtain a text to be processed, and determine an initial regulation query text corresponding to the text to be processed;

[0006] Perform a business scenario consistency check on the initial regulation query text to obtain a target regulation query text that meets the business scenario consistency requirements;

[0007] Determine reference business rule data corresponding to the text to be processed and reference regulation data corresponding to the target regulation query text, and determine a target processing result corresponding to the text to be processed through a pre-constructed first processing model based on the text to be processed, the target regulation query text, the reference business rule data, and the reference regulation data.

[0008] Optionally, determining the initial regulation query text corresponding to the text to be processed includes: performing text reconstruction on the text to be processed based on a pre-constructed text reconstruction model to obtain the initial regulation query text corresponding to the text to be processed, where the initial regulation query text is a text that meets regulation query terms.

[0009] Optionally, perform a business scenario consistency check on the initial regulatory inquiry text to obtain a target regulatory inquiry text that meets the business scenario consistency requirements, including: determining the business scenario similarity determination result between the text to be processed and the initial regulatory inquiry text, and determining the business scenario consistency check result of the initial regulatory inquiry text based on the business scenario similarity determination result and a preset similarity threshold; if the business scenario consistency check result of the initial regulatory inquiry text meets the business scenario consistency requirements, then determine the initial regulatory inquiry text as the target regulatory inquiry text; if the business scenario consistency check result of the initial regulatory inquiry text does not meet the business scenario consistency requirements, then perform a correction process on the initial regulatory inquiry text to obtain a target regulatory inquiry text that meets the business scenario consistency requirements.

[0010] Optionally, perform a correction process on the initial regulatory inquiry text to obtain a target regulatory inquiry text that meets the business scenario consistency requirements, including: constructing a text correction prompt model corresponding to the initial regulatory inquiry text, and determining the corrected text corresponding to the initial regulatory inquiry text based on the text correction prompt model and a pre-constructed second processing model; re-perform a business scenario consistency check on the corrected text corresponding to the initial regulatory inquiry text, and when the corrected text does not meet the business scenario consistency requirements, re-correct the corrected text until a corrected text that meets the business scenario consistency requirements is obtained.

[0011] Optionally, the method further includes: during the correction process of the initial regulatory inquiry text, determining the cumulative number of corrections corresponding to the corrected text; in the case where the cumulative number of corrections corresponding to the corrected text exceeds a preset correction number threshold and the corrected text does not meet the business scenario consistency requirements, determining the corrected text with the highest similarity score corresponding to the business scenario similarity determination result among the generated corrected texts as the target regulatory inquiry text.

[0012] Optionally, determining the reference business rule data corresponding to the text to be processed and the reference regulatory data corresponding to the target regulatory inquiry text, including: obtaining a business rule knowledge base and a regulatory knowledge base; retrieving in the business rule knowledge base based on the text to be processed to determine the reference business rule data corresponding to the text to be processed; and retrieving in the regulatory knowledge base based on the target regulatory inquiry text to obtain the reference regulatory data corresponding to the target regulatory inquiry text.

[0013] Optionally, based on the text to be processed, the target regulation query text, the reference business rule data, and the reference regulation data, determine the target processing result corresponding to the text to be processed through a pre-constructed first processing model, including: obtaining a pre-constructed answer generation prompt template, and constructing an answer generation prompt model based on the text to be processed, the target regulation query text, the reference business rule data, the reference regulation data, and the pre-constructed answer generation prompt template; inputting the answer generation prompt model into the pre-constructed first processing model, and outputting the target processing result of the text to be processed by the pre-constructed first processing model.

[0014] According to another aspect of the present invention, there is provided an intelligent question-answering data processing device, including:

[0015] An initial regulation query text determination module, configured to obtain the text to be processed and determine the initial regulation query text corresponding to the text to be processed;

[0016] A target regulation query text determination module, configured to perform a business scenario consistency check on the initial regulation query text to obtain a target regulation query text that meets the business scenario consistency requirements;

[0017] A reference data determination module, configured to determine the reference business rule data corresponding to the text to be processed and the reference regulation data corresponding to the target regulation query text;

[0018] A target processing result determination module, configured to determine the target processing result corresponding to the text to be processed through a pre-constructed first processing model based on the text to be processed, the target regulation query text, the reference business rule data, and the reference regulation data.

[0019] According to another aspect of the present invention, there is provided an electronic device, the electronic device includes:

[0020] At least one processor; and

[0021] A memory communicatively connected to the at least one processor; wherein,

[0022] The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the intelligent question-answering data processing method of any embodiment of the present invention.

[0023] According to another aspect of the present invention, there is provided a computer-readable storage medium, and the computer-readable storage medium stores computer instructions for causing a processor to implement the intelligent question-answering data processing method of any embodiment of the present invention when executed.

[0024] In the technical solution of the embodiment of the present invention, by obtaining the text to be processed, the initial regulatory inquiry text corresponding to the text to be processed is determined; the initial regulatory inquiry text is subjected to business scenario consistency verification to obtain a target regulatory inquiry text that meets the requirements of business scenario consistency; the reference business rule data corresponding to the text to be processed and the reference regulatory data corresponding to the target regulatory inquiry text are determined, and based on the text to be processed, the target regulatory inquiry text, the reference business rule data, and the reference regulatory data, the target processing result corresponding to the text to be processed is determined through a pre-constructed first processing model. In this solution, by determining the initial regulatory inquiry text corresponding to the text to be processed and performing business scenario consistency verification on the initial regulatory inquiry text to obtain a target regulatory inquiry text that meets the requirements of business scenario consistency verification. Further, the reference business rule data corresponding to the text to be processed and the reference regulatory data corresponding to the target regulatory inquiry text are determined, and then the target processing result corresponding to the text to be processed is determined according to the text to be processed, the target regulatory inquiry text, the reference business rule data, and the reference regulatory data, which solves the problems of inconsistent business scenarios and non-compliance with regulations existing in the process of intelligent question-answering data processing, and not only makes the processing result of the text to be processed meet the business scenario consistency, but also improves the accuracy and compliance of the processing result.

[0025] It should be understood that the content described in this part is not intended to identify the key or important features of the embodiments of the present invention, nor is it used to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. Brief Description of the Drawings

[0026] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0027] Figure 1 It is a flowchart of an intelligent question-answering data processing method provided in Embodiment 1 of the present invention;

[0028] Figure 2 It is a flowchart of an intelligent question-answering data processing method provided in Embodiment 2 of the present invention;

[0029] Figure 3 It is a schematic structural diagram of an intelligent question-answering data processing device provided in Embodiment 3 of the present invention;

[0030] Figure 4 It is a schematic structural diagram of an electronic device for implementing the intelligent question-answering data processing method of the embodiments of the present invention. Detailed Description of the Embodiments

[0031] To enable those skilled in the art to better understand the solution of the present invention, the following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0032] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged under appropriate circumstances so that the embodiments of the present invention described herein can be implemented in an order different from those illustrated or described herein. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0033] Embodiment 1

[0034] Figure 1 is a flowchart of an intelligent question-answering data processing method provided in Embodiment 1 of the present invention. This embodiment is applicable to the situation of determining the target processing result corresponding to the question text. This method can be executed by an intelligent question-answering data processing device, which can be implemented in the form of hardware and / or software, and the intelligent question-answering data processing device can be configured in electronic devices such as computers and servers. As Figure 1 shown, the method includes:

[0035] S110. Obtain the text to be processed and determine the initial regulatory inquiry text corresponding to the text to be processed.

[0036] Among them, the text to be processed can be specifically understood as the problem text input by the user. The user can input the corresponding problem text according to different business requirements. Exemplarily, the user can input the problem text or problem voice information through an external input device. If the input is problem voice information, a pre-set voice conversion module is called to process the input problem voice information to obtain the corresponding problem text, and the obtained problem text is used as the text to be processed. If the input is problem text, the input problem text is used as the text to be processed. The initial regulatory query text can be specifically understood as the regulatory-compliant problem text corresponding to the text to be processed. Exemplarily, the text to be processed can be reconstructed through a pre-constructed text reconstruction method to obtain the initial regulatory query text corresponding to the text to be processed; alternatively, the corresponding pre-constructed regulatory knowledge base can be called, and by retrieving in the regulatory knowledge base, the initial regulatory query text corresponding to the text to be processed can be obtained.

[0037] Specifically, the problem text or problem voice information can be input through an external input device, and the text corresponding to the input problem text or problem voice information is used as the text to be processed. A pre-constructed text reconstruction model is called to process the text to be processed to obtain the initial regulatory query text corresponding to the text to be processed.

[0038] In this embodiment, by determining the initial regulatory query text corresponding to the text to be processed, it is used to subsequently assist in determining the processing result corresponding to the text to be processed, so as to ensure that the processing result of the obtained text to be processed complies with the corresponding regulatory requirements, which helps to improve the accuracy of determining the processing result corresponding to the text to be processed.

[0039] Optionally, determining the initial regulatory query text corresponding to the text to be processed includes: reconstructing the text to be processed based on a pre-constructed text reconstruction model to obtain the initial regulatory query text corresponding to the text to be processed, where the initial regulatory query text is a text that meets the regulatory query terms.

[0040] Among them, the pre-constructed text reconstruction model can be specifically understood as a text reconstruction model constructed based on a large language model, which is used to reconstruct the text to be processed to adapt to the regulatory requirements in the business scenario to which the text to be processed belongs, so as to obtain a text that meets the regulatory query terms in this business scenario.

[0041] Specifically, when the text to be processed is obtained, a pre-built text reconstruction model is called to reconstruct the text to be processed through the text reconstruction model, so as to obtain an initial regulatory inquiry text corresponding to the text to be processed that meets the regulatory inquiry terms, enabling a subsequent processing result that meets the regulatory requirements corresponding to the text to be processed to be obtained. Preferably, when reconstructing the text to be processed through the pre-built text reconstruction model, a text reconstruction prompt model corresponding to the text to be processed can be pre-built, and the pre-built text reconstruction model is guided to perform text reconstruction processing through the text reconstruction prompt model, converting the text to be processed into a corresponding initial regulatory inquiry text. Among them, for constructing a corresponding text reconstruction prompt model according to the text to be processed, specifically, a text reconstruction prompt template can be called, and the text to be processed is filled into the text reconstruction prompt template to obtain a text reconstruction prompt model corresponding to the text to be processed.

[0042] S120. Perform a business scenario consistency check on the initial regulatory inquiry text to obtain a target regulatory inquiry text that meets the business scenario consistency requirements.

[0043] It should be noted that for the initial regulatory inquiry text obtained by reconstructing the text according to the text to be processed, there may be a problem that the business scenario corresponding to the initial regulatory inquiry text of the text to be processed does not meet the business scenario consistency requirements with the business scenario corresponding to the text to be processed, which may lead to an error in the subsequent processing result corresponding to the text to be processed. Therefore, after obtaining the initial regulatory inquiry text corresponding to the text to be processed, it is necessary to perform a business scenario consistency check on the initial regulatory inquiry text to obtain a target regulatory inquiry text that meets the business scenario consistency requirements. Exemplarily, a business scenario consistency check can be performed on the business scenario corresponding to the initial regulatory inquiry text of the text to be processed and the business scenario corresponding to the text to be processed by calling a business scenario consistency check algorithm, or the similarity data between the business scenario corresponding to the text to be processed and the business scenario corresponding to the initial regulatory inquiry text can be determined, and then the initial regulatory inquiry text can be checked for business scenario consistency according to the similarity data to obtain a target regulatory inquiry text that meets the business scenario consistency requirements.

[0044] Specifically, when obtaining the initial regulatory inquiry text, the business scenario similarity determination model can be called to process the text to be processed and the initial regulatory inquiry text corresponding to the text to be processed, so as to obtain the similarity data between the business scenario corresponding to the text to be processed and the business scenario corresponding to the initial regulatory inquiry text. According to the similarity data, it is determined whether the initial regulatory inquiry text meets the requirements of business consistency verification. Exemplarily, a preset similarity threshold can be obtained, and the similarity data is compared with the preset similarity threshold. If the similarity data meets the preset similarity threshold, it is determined that the initial regulatory inquiry text meets the requirements of business consistency verification. If the similarity data does not meet the preset similarity threshold, it is determined that the initial regulatory inquiry text does not meet the requirements of business consistency verification. In the case where it is determined that the initial regulatory inquiry text does not meet the requirements of business consistency verification, it is necessary to re-obtain the text to be processed, determine the initial regulatory inquiry text corresponding to the text to be processed, and perform business scenario consistency verification until the target regulatory inquiry text that meets the requirements of business scenario consistency is obtained.

[0045] In this embodiment, after obtaining the initial regulatory inquiry text, business scenario consistency verification is performed on the initial regulatory inquiry text to obtain a target regulatory inquiry text that meets the requirements of business scenario consistency, so as to ensure that a text that meets the regulatory inquiry terms and is consistent with the business scenario corresponding to the text to be processed is obtained, thereby making the subsequent processing result corresponding to the text to be processed comply with the regulatory requirements.

[0046] S130. Determine the reference business rule data corresponding to the text to be processed and the reference regulatory data corresponding to the target regulatory inquiry text.

[0047] Among them, the reference business rule data specifically represents the rule reference data that needs to be relied on at the business level corresponding to the text to be processed. It should be noted that the reference business rule data can be business rule data set by an enterprise or institution internally according to the business scenario. Exemplarily, the business scenario includes but is not limited to the company reimbursement Q&A assistant and the company recruitment Q&A assistant. Correspondingly, the business rule data includes but is not limited to the company reimbursement system document and the company recruitment system document. The reference business rule data of the text to be processed can specifically be understood as the reference data related to the text to be processed extracted from the business rule data. The reference regulatory data specifically represents the reference data extracted from the corresponding regulatory document according to the target regulatory inquiry text, which is used to assist in determining the processing result of the text to be processed subsequently, so as to ensure that the processing result corresponding to the text to be processed can comply with the regulatory requirements.

[0048] Specifically, in order to ensure that the processing result corresponding to the text to be processed meets both the business rule requirements and the regulatory requirements, after obtaining the target regulatory query text corresponding to the text to be processed, the corresponding reference data can be determined based on the text to be processed and the target regulatory query text respectively, that is, the reference business rule data corresponding to the text to be processed and the reference regulatory data corresponding to the target regulatory query text are obtained.

[0049] In this embodiment, by determining the reference business rule data corresponding to the text to be processed and the reference regulatory data corresponding to the target regulatory query text, reference data is provided for subsequent determination of the processing result corresponding to the text to be processed, which helps to improve the accuracy and compliance of the processing result corresponding to the text to be processed.

[0050] Optionally, determining the reference business rule data corresponding to the text to be processed and the reference regulatory data corresponding to the target regulatory query text includes: obtaining a business rule knowledge base and a regulatory knowledge base; retrieving in the business rule knowledge base based on the text to be processed to determine the reference business rule data corresponding to the text to be processed; and retrieving in the regulatory knowledge base based on the target regulatory query text to obtain the reference regulatory data corresponding to the target regulatory query text.

[0051] Among them, the business rule knowledge base specifically refers to a knowledge base constructed based on business data corresponding to different business scenarios. The business data includes rule files corresponding to different business scenarios and data generated during the execution of business processes in different business scenarios. By using a text parsing library to parse rule files of multiple file types, performing chunking processing and text vectorization processing on the parsing results, and storing the obtained text chunk vectors and text chunk detail data into a pre-set business rule knowledge base. The text types include but are not limited to DOC, DOCX, and PDF. The text parsing library includes but is not limited to PyPDF2 and Python-docx. The corresponding text parsing library is selected according to the file type of the rule file to complete the parsing of the rule file of the corresponding file type. The regulatory knowledge base specifically refers to a knowledge base constructed from different regulatory documents. Preferably, the latest released regulatory documents can be obtained through an external interface or imported through an external device. When it is detected that the latest released regulatory documents have been obtained, a text parsing library is called to perform parsing processing on the latest released regulatory documents, and chunking processing and text vectorization processing are performed on the parsing results, and the obtained text chunk vectors and text chunk detail data are stored into the corresponding regulatory knowledge base to ensure that the regulatory data stored in the regulatory knowledge base are all the latest released regulations.

[0052] Specifically, the text to be processed is vectorized to obtain the text vector corresponding to the text to be processed. Vector retrieval is performed in the business rule knowledge base based on the text vector to obtain the text block vector with the highest vector similarity. Then, the text block detail data corresponding to the text block vector is determined as the reference business rule data corresponding to the text to be processed. The target regulation query text is vectorized to obtain the text vector corresponding to the target regulation query text. Vector retrieval is performed in the regulation knowledge base based on the text vector corresponding to the target regulation query text to obtain the text block vector with the highest vector similarity. Then, the text block detail data corresponding to the text block vector is determined as the reference regulation data corresponding to the target regulation query text.

[0053] In a specific embodiment, taking the application scenario of the reimbursement assistant as an example, the business data under the business scenario is collected from the database corresponding to the internal reimbursement assistant of the enterprise. The obtained business data includes but is not limited to the company's reimbursement and system process data, including reimbursement scope and standards, reimbursement application process, approval authority and process, financial accounting and control, and reimbursement payment methods. In addition, the latest relevant regulatory documents are obtained from the relevant regulations publishing platform by calling the external interface. Exemplarily, the relevant regulatory documents can be the "General Principles of Enterprise Finance". After obtaining the company's reimbursement and system process related documents and regulatory documents, PyPDF2 is used to perform text parsing on the company's reimbursement and system process related documents and regulatory documents respectively, and the parsing results are divided into blocks to obtain the text segmentation results corresponding to the company's reimbursement and system process related documents, and the text segmentation results of the corresponding regulatory documents. Exemplarily, the text segmentation data of the company's reimbursement and system process related documents can be expressed as: [{"data":"Company reimbursement system and process...Reasonably control expenses, specially formulate this system","ID":id_1},{"data":"Company reimbursement system and process...Reasonably control expenses, specially formulate this system","ID":id_2},{"data":"Company reimbursement system and process...Reasonably control expenses, specially formulate this system","ID":id_3},{"data":"Company reimbursement system and process...Reasonably control expenses, specially formulate this system","ID":id_4},{"data":"Company reimbursement system and process...Reasonably control expenses, specially formulate this system","ID":id_5},{"data":"Company reimbursement system and process...Reasonably control expenses, specially formulate this system","ID":id_6},{"data":"Company reimbursement system and process a":"Article 2 This system divides financial reimbursement into daily office expenses... specific financial reimbursement systems and reimbursement processes for each expenditure according to relevant financial systems and the actual situation of the company. \nArticle 3 This system applies to all employees of the company. ", "ID": id_2}, ..., {"data": "Part 3 Daily expense reimbursement system and process ... Approval in accordance with the prescribed approval procedures. ", "ID": id_m}], where data represents the text block details, ID represents the unique identifier corresponding to the text block, i represents the i-th text block after splitting, i∈[1,m]. The block data of the reimbursement-related regulatory documents are represented as follows: [{“data”:“Order of the Ministry of Finance of the People’s Republic of China\n\nNo. 41\nAccording to the “State Council’s Decision on the General Rules of Enterprise Finance” and the “Enterprise Accounting Standards”…which will be implemented on January 1, 2007.\n\nMinister\tJin Renqing\nDecember 4, 2006”,“ID”:id_1},{“data”:“General Rules of Enterprise Finance\n\nChapter 1…Except for financial enterprises.\nOther enterprises shall refer to it for implementation.”,“ID”:id_2},……,{“data”:“Chapter 10\tSupplementary Provisions\n\nArticle 77\tInstitutions that implement enterprise management shall apply these general rules by analogy.\nArticle 78\tThese general rules shall be implemented on January 1, 2007.”,“ID”:id_n}], where data here represents the text block details of the reimbursement-related regulatory documents, ID represents the unique identifier corresponding to the text block, i represents the i-th text block after the regulatory document is split, i∈[1,n].For the text chunking results corresponding to the company's reimbursement and system process-related documents, as well as the text chunking results corresponding to the reimbursement-related regulatory documents, vectorization processing is performed respectively. The text chunking model can be used to vectorize each chunk of text, obtaining the chunk vector data of the reimbursement and system process and the chunk vector data of the reimbursement-related regulatory documents. The text chunking model includes but is not limited to the M3E text embedding model, the BGE vector model, and the BCE vector model. The text chunking vectorization results can be expressed as: chunk vector data of the reimbursement and system process: [{"data": text vector chunk_1, "ID": id_1}, {"data": text vector chunk_2, "ID": id_2},... {"data": text vector chunk_n, "ID": id_m}]; chunk vector data of the reimbursement-related regulatory documents: [{"data": text vector chunk_1, "ID": id_1}, {"data": text vector chunk_2, "ID": id_2}, ……, {"data": text vector chunk_n, "ID": id_n}]. Further, the vectorized chunk data is stored in the vector database. Exemplarily, the vector databases that can be used include but are not limited to the milvus vector database and the faiss vector database, thereby obtaining the vector knowledge base composed of the chunk vector data of the reimbursement and system process and the vector knowledge base composed of the chunk vector data of the reimbursement-related regulatory documents. Thus, a business rule knowledge base corresponding to the reimbursement and system process is constructed based on the vector knowledge base and the text knowledge base corresponding to the reimbursement and system process, and a regulatory knowledge base corresponding to the reimbursement-related regulatory documents is constructed based on the vector knowledge base and the text knowledge base corresponding to the reimbursement-related regulatory documents.

[0054] Exemplarily, the user input text is: Borrowing money from the affiliated company, what procedures need to be followed? Then, taking the user input text as the text to be processed, the text reconstruction model is used to reconstruct the text to be processed, and the initial regulatory query text obtained is: What legal terms need to be noted when borrowing money from one's own company? Further, a business scenario consistency check is performed on the text to be processed and the corresponding initial regulatory query text. Specifically, the business scenario similarity determination model is used to determine the business scenario similarity between the text to be processed and the corresponding initial regulatory query text. If the obtained business scenario similarity is greater than or equal to 0.8, it can be determined that the text to be processed and the corresponding initial regulatory query text meet the business scenario consistency requirement, and the initial regulatory query text can be determined as the target regulatory query text. The text vectorization model is used to perform vectorization processing on the text to be processed and the target regulatory query text respectively, to obtain the vector data corresponding to the text to be processed and the target regulatory query text respectively. Through each vector data, retrieval is performed in the corresponding vector knowledge base respectively, to obtain the K text vector blocks with the highest vector similarity corresponding to the text to be processed and the k text vector blocks with the highest vector similarity corresponding to the target regulatory query text. Furthermore, retrieval is performed in the corresponding text knowledge base respectively through the identifiers corresponding to the text vector blocks, to obtain the reference business rule data corresponding to the text to be processed, and the reference regulatory data corresponding to the target regulatory query text, where the value of k can be set according to actual needs and is not limited here.

[0055] S140. Based on the text to be processed, the target regulatory query text, the reference business rule data, and the reference regulatory data, determine the target processing result corresponding to the text to be processed through a pre-constructed first processing model.

[0056] Among them, the first processing model specifically refers to a data processing model constructed based on a large language model, which is used to determine the processing result corresponding to the text to be processed, that is, to determine the Q&A result corresponding to the text to be processed. The first processing model can be pre-constructed according to actual needs and can be directly called when determining the processing result of the text to be processed.

[0057] Specifically, input the obtained text to be processed, the target regulatory query text, the reference business rule data, and the reference regulatory data into the pre-constructed first processing model, and use the target regulatory query text, the reference business rule data, and the reference regulatory data as prompt information to guide the pre-constructed first processing model to process the text to be processed, so as to obtain the target processing result corresponding to the text to be processed.

[0058] Optionally, determine the target processing result corresponding to the text to be processed based on the text to be processed, the target regulation query text, the reference business rule data, and the reference regulation data through a pre-built first processing model, including: obtaining a pre-built answer generation prompt template, and constructing an answer generation prompt model based on the text to be processed, the target regulation query text, the reference business rule data, the reference regulation data, and the pre-built answer generation prompt template; inputting the answer generation prompt model into the pre-built first processing model, and outputting the target processing result of the text to be processed by the pre-built first processing model.

[0059] Specifically, obtain the pre-built answer generation prompt template in the target storage space, and fill the text to be processed, the target regulation query text, the reference business rule data, and the reference regulation data into the pre-built answer generation prompt template respectively to obtain the answer generation prompt model corresponding to the text to be processed. Input the answer generation prompt model into the pre-built first processing model to guide the first processing model to process the text to be processed, and then output the target processing result of the text to be processed.

[0060] Among them, the information in the answer generation prompt template includes but is not limited to defining roles and tasks, input data, and output data. Exemplarily, (1) Defining roles and tasks: You are a professional intelligent question-answering assistant. Please do the following operations: Operation 1: Generate an answer based on the user question and the text blocks similar to the user question; Operation 2: Generate an answer based on the regulation query question and the text blocks similar to the regulation query question; Operation 3: Use the answer of Operation 2 to correct the answer of Operation 1 and return the final answer. (2) Input data: User question: {}, Text blocks similar to the user question: {}, Regulation query question: {}, Text blocks similar to the regulation query question: {}; (3) Output data: Output the final answer: {}. The example of the answer generation prompt template can be expressed as follows:

[0061] """

[0062] You are a professional intelligent question-answering assistant. Please do the following operations:

[0063] Operation 1: Generate an answer based on the user question and the text blocks similar to the user question;

[0064] Operation 2: Generate an answer based on the regulation query question and the text blocks similar to the regulation query question;

[0065] Operation 3: Use the answer of Operation 2 to correct the answer of Operation 1 and return the final answer.

[0066] User question: {}

[0067] Text blocks similar to the user question: {}

[0068] Regulation query question: {}

[0069] Regulatory inquiry question similar text block: {}

[0070] Output the final answer: {}

[0071] """

[0072] In the case of constructing an answer generation prompt model, the text to be processed, the reference business rule data, the target regulatory inquiry text, and the reference regulatory data are filled into the corresponding data storage locations of the user question, the user question similar text block, the regulatory inquiry question, and the regulatory inquiry question similar text block, that is, filled inside the curly braces, to obtain the answer generation prompt model. Furthermore, the answer generation prompt model is input into a pre-constructed first processing model to obtain the target processing result of the text to be processed.

[0073] The technical solution of this embodiment obtains the text to be processed and determines the corresponding initial regulatory inquiry text for the text to be processed; performs a business scenario consistency check on the initial regulatory inquiry text to obtain a target regulatory inquiry text that meets the business scenario consistency requirements; determines the reference business rule data corresponding to the text to be processed and the reference regulatory data corresponding to the target regulatory inquiry text, and determines the target processing result corresponding to the text to be processed through a pre-constructed first processing model based on the text to be processed, the target regulatory inquiry text, the reference business rule data, and the reference regulatory data. This solution determines the corresponding initial regulatory inquiry text for the text to be processed and performs a business scenario consistency check on the initial regulatory inquiry text to obtain a target regulatory inquiry text that meets the business scenario consistency check requirements. Further, it determines the reference business rule data corresponding to the text to be processed and the reference regulatory data corresponding to the target regulatory inquiry text, and then determines the target processing result corresponding to the text to be processed according to the text to be processed, the target regulatory inquiry text, the reference business rule data, and the reference regulatory data, solving the problems of inconsistent business scenarios and non-compliance with regulations existing in the data processing process, making the processing result of the text to be processed meet the business scenario consistency and improving the accuracy and compliance of the processing result.

[0074] Embodiment Two

[0075] Figure 2It is a flowchart of an intelligent Q&A data processing method provided in the second embodiment of the present invention. This embodiment is a further optimization of the method in the above embodiment. Optionally, determine the business scenario similarity determination result between the text to be processed and the initial regulatory inquiry text, and determine the business scenario consistency verification result of the initial regulatory inquiry text based on the business scenario similarity determination result and the preset similarity threshold; if the business scenario consistency verification result of the initial regulatory inquiry text meets the business scenario consistency requirement, then determine the initial regulatory inquiry text as the target regulatory inquiry text; if the business scenario consistency verification result of the initial regulatory inquiry text does not meet the business scenario consistency requirement, then perform correction processing on the initial regulatory inquiry text to obtain the target regulatory inquiry text that meets the business scenario consistency requirement. As Figure 2 shown, the method includes:

[0076] S210. Obtain the text to be processed and determine the initial regulatory inquiry text corresponding to the text to be processed.

[0077] S220. Determine the business scenario similarity determination result between the text to be processed and the initial regulatory inquiry text, and determine the business scenario consistency verification result of the initial regulatory inquiry text based on the business scenario similarity determination result and the preset similarity threshold.

[0078] Specifically, call the pre-constructed business scenario similarity determination model to process the text to be processed and the initial regulatory inquiry text to obtain the business scenario similarity determination result between the text to be processed and the initial regulatory inquiry text. Among them, the business scenario similarity determination model specifically refers to a model constructed according to the large language model. Furthermore, compare the business scenario similarity determination result with the preset similarity threshold. If the business scenario similarity determination result is greater than or equal to the preset similarity threshold, then determine that the business scenario consistency verification result of the initial regulatory inquiry text is that the initial regulatory inquiry text meets the business scenario consistency requirement. If the business scenario similarity determination result is less than the preset similarity threshold, then determine that the business scenario consistency verification result of the initial regulatory inquiry text is that the initial regulatory inquiry text does not meet the business scenario consistency requirement.

[0079] Exemplarily, the business scenario similarity determination model specifically refers to a model constructed according to the large language model. In the actual application process, in order to obtain the business scenario similarity determination result more accurately, a similarity determination prompt model can be determined first to guide the business scenario similarity determination model to determine the business scenario similarity determination result. Specifically, obtain the pre-set similarity determination prompt template, and the information in the similarity determination prompt template includes but is not limited to defining roles and tasks, example content, input data, and output data. Exemplarily, an example of the similarity determination prompt template is as follows:

[0080] """

[0081] Define roles and tasks:

[0082] You are a professional scenario judgment assistant. Please determine whether the user's question and the regulatory inquiry question belong to the same business scenario and give a similarity score; the similarity score is a decimal,

[0083] The value range is 0 - 1, and 1 significant digit is retained.

[0084] Example content:

[0085] User question: I want to borrow 10,000 yuan from the company. What procedures do I need to follow?

[0086] Regulatory inquiry question: What legal provisions should I pay attention to when borrowing from my own company?

[0087] Output similarity score: 0.8

[0088] Input data:

[0089] User question: {}

[0090] Regulatory inquiry question: {}

[0091] Output data:

[0092] Output similarity score:

[0093] """

[0094] In the case of constructing a similarity determination prompt model, the text to be processed and the initial regulatory inquiry text are respectively filled into the corresponding data storage positions of the user question and the regulatory inquiry question in the template, that is, filled inside the curly braces, to obtain a similarity determination prompt model. Then, the similarity determination prompt model is input into the business scenario similarity determination model, and the business scenario similarity determination model outputs a similarity score, that is, the business scenario similarity determination result of the text to be processed and the initial regulatory inquiry text is obtained.

[0095] S230. If the business scenario consistency verification result of the initial regulatory inquiry text meets the business scenario consistency requirements, then the initial regulatory inquiry text is determined as the target regulatory inquiry text.

[0096] Specifically, if it is recognized that the business scenario consistency verification result of the initial regulatory inquiry text meets the business scenario consistency requirements, there is no need to modify the initial regulatory inquiry text, and the initial regulatory inquiry text can be directly determined as the target regulatory inquiry text.

[0097] S240. If the business scenario consistency check result of the initial regulatory inquiry text does not meet the business scenario consistency requirements, the initial regulatory inquiry text is corrected to obtain a target regulatory inquiry text that meets the business scenario consistency requirements.

[0098] Specifically, if the business scenario consistency check result of the initial regulatory inquiry text does not meet the business scenario consistency requirements, it is necessary to call the text correction algorithm and correct the initial regulatory inquiry text according to the text correction algorithm until a target regulatory inquiry text that meets the business scenario consistency requirements is obtained.

[0099] Optionally, correcting the initial regulatory inquiry text to obtain a target regulatory inquiry text that meets the business scenario consistency requirements includes: constructing a text correction prompt model corresponding to the initial regulatory inquiry text, determining a corrected text corresponding to the initial regulatory inquiry text based on the text correction prompt model and a pre-constructed second processing model; re-checking the business scenario consistency of the corrected text corresponding to the initial regulatory inquiry text, and when the corrected text does not meet the business scenario consistency requirements, re-correcting the corrected text until a corrected text that meets the business scenario consistency requirements is obtained.

[0100] Among them, the pre-constructed second processing model specifically refers to a model constructed based on a large language model. In this embodiment, the second processing model is a trained model that can be directly called. The second processing model is used to determine the corrected text, and the corrected text can be the text of the initial regulatory inquiry text after one or more corrections. It should be noted that the process of determining the corrected text is an iterative correction process, that is, the text to be corrected in the current iteration is the corrected text obtained after the previous correction process.

[0101] Specifically, in order to quickly and accurately obtain the revised text, a pre-built text revision prompt template can be called, and the text to be revised can be filled into the text revision prompt template to obtain a text revision prompt model, wherein, for the first revision process, the text to be revised is the initial regulatory inquiry text, and the text to be revised in the subsequent revision process is the revised text obtained by the previous text revision process. The text revision prompt model is input into the pre-built second processing model, and the pre-built second processing model performs text revision processing, outputs the revised text corresponding to the initial regulatory inquiry text, and then re-checks the business scenario consistency of the revised text corresponding to the initial regulatory inquiry text to obtain the business scenario consistency verification result. If the revised text corresponding to the initial regulatory inquiry text meets the business scenario consistency verification requirements, the revised text corresponding to the initial regulatory inquiry text is determined as the target regulatory inquiry text, and the text revision process ends; if the revised text corresponding to the initial regulatory inquiry text does not meet the business scenario consistency verification requirements, the revised text corresponding to the initial regulatory inquiry text continues to be revised until the revised text that meets the business scenario consistency requirements is obtained. In the case of obtaining the revised text that meets the business scenario consistency requirements, the corresponding revised text is determined as the target regulatory inquiry text.

[0102] Optionally, the method also includes: in the process of revising the initial regulatory inquiry text, determining the cumulative number of revisions corresponding to the revised text; when the cumulative number of revisions corresponding to the revised text exceeds a preset revision number threshold and the revised text does not meet the business scenario consistency requirements, the revised text with the highest similarity score corresponding to the business scenario similarity judgment result in the generated revised text is determined as the target regulatory inquiry text.

[0103] Specifically, in order to avoid wasting computing resources, the number of text corrections needs to be limited, and text corrections cannot be performed endlessly. Therefore, in the process of correcting the initial regulatory inquiry text, the number of text corrections is counted in real time to determine the cumulative number of corrections corresponding to the corrected text. The corrected text obtained after each text correction process is checked for business scenario consistency. If the business scenario consistency check result corresponding to the corrected text does not meet the business scenario consistency requirements, it is determined whether the cumulative number of corrections exceeds the preset correction number threshold. If the cumulative number of corrections exceeds the preset correction number threshold, the text correction process is stopped, and the business scenario similarity determination results in the generated corrected text are arranged in descending order to obtain the corrected text corresponding to the highest similarity score in the business scenario similarity determination results, and the corrected text corresponding to the highest similarity score is determined as the target regulatory inquiry text.

[0104] S250: Determine reference business rule data corresponding to the text to be processed and reference regulatory data corresponding to the target regulatory query text.

[0105] S260. Determine the target processing result corresponding to the text to be processed through a pre-constructed first processing model based on the text to be processed, the target regulation query text, the reference business rule data, and the reference regulation data.

[0106] In the technical solution of this embodiment, by obtaining the text to be processed, the initial regulation query text corresponding to the text to be processed is determined; the business scenario similarity determination result between the text to be processed and the initial regulation query text is determined, and based on the business scenario similarity determination result and the preset similarity threshold, the business scenario consistency verification result of the initial regulation query text is determined; if the business scenario consistency verification result of the initial regulation query text meets the business scenario consistency requirement, the initial regulation query text is determined as the target regulation query text; if the business scenario consistency verification result of the initial regulation query text does not meet the business scenario consistency requirement, the initial regulation query text is corrected to obtain the target regulation query text that meets the business scenario consistency requirement; the reference business rule data corresponding to the text to be processed and the reference regulation data corresponding to the target regulation query text are determined; the target processing result corresponding to the text to be processed is determined through a pre-constructed first processing model based on the text to be processed, the target regulation query text, the reference business rule data, and the reference regulation data. This solution determines the initial regulation query text corresponding to the text to be processed and performs a business scenario consistency verification on the initial regulation query text to obtain the target regulation query text that meets the business scenario consistency verification requirement. Further, the reference business rule data corresponding to the text to be processed and the reference regulation data corresponding to the target regulation query text are determined, and then the target processing result corresponding to the text to be processed is determined according to the text to be processed, the target regulation query text, the reference business rule data, and the reference regulation data, solving the problems of inconsistent business scenarios and non-compliance with regulations in the data processing process, making the processing result of the text to be processed meet the business scenario consistency and improving the accuracy and compliance of the processing result.

[0107] Embodiment III

[0108] Figure 3 It is a schematic structural diagram of an intelligent question-answering data processing device provided in Embodiment III of the present invention. As Figure 3 shown, the device includes:

[0109] An initial regulation query text determination module 310, configured to obtain the text to be processed and determine the initial regulation query text corresponding to the text to be processed;

[0110] A target regulation query text determination module 320, configured to perform a business scenario consistency verification on the initial regulation query text to obtain the target regulation query text that meets the business scenario consistency requirement;

[0111] A reference data determination module 330, configured to determine reference business rule data corresponding to the text to be processed and reference regulation data corresponding to the target regulation query text;

[0112] A target processing result determination module 340, configured to determine a target processing result corresponding to the text to be processed based on the text to be processed, the target regulation query text, the reference business rule data, and the reference regulation data through a pre-constructed first processing model.

[0113] In the technical solution of this embodiment, the text to be processed is obtained through the initial regulation query text determination module, and the initial regulation query text corresponding to the text to be processed is determined; the target regulation query text determination module performs a business scenario consistency check on the initial regulation query text to obtain a target regulation query text that meets the business scenario consistency requirements; the reference data determination module determines the reference business rule data corresponding to the text to be processed and the reference regulation data corresponding to the target regulation query text; the target processing result determination module determines the target processing result corresponding to the text to be processed based on the text to be processed, the target regulation query text, the reference business rule data, and the reference regulation data through a pre-constructed first processing model. This solution determines the initial regulation query text corresponding to the text to be processed and performs a business scenario consistency check on the initial regulation query text to obtain a target regulation query text that meets the business scenario consistency check requirements. Further, the reference business rule data corresponding to the text to be processed and the reference regulation data corresponding to the target regulation query text are determined, and then the target processing result corresponding to the text to be processed is determined according to the text to be processed, the target regulation query text, the reference business rule data, and the reference regulation data, solving the problems of inconsistent business scenarios and non-compliance with regulations existing in the data processing process, making the processing result of the text to be processed meet the business scenario consistency and improving the accuracy and compliance of the processing result.

[0114] Based on the above embodiment, optionally, the initial regulation query text determination module 310 is specifically configured to perform text reconstruction on the text to be processed based on a pre-constructed text reconstruction model to obtain an initial regulation query text corresponding to the text to be processed, where the initial regulation query text is a text that meets the regulation query terms.

[0115] Optionally, the target regulation query text determination module 320 is specifically configured to determine the business scenario similarity determination result between the text to be processed and the initial regulation query text, and determine the business scenario consistency verification result of the initial regulation query text based on the business scenario similarity determination result and a preset similarity threshold; if the business scenario consistency verification result of the initial regulation query text meets the business scenario consistency requirement, the initial regulation query text is determined as the target regulation query text; if the business scenario consistency verification result of the initial regulation query text does not meet the business scenario consistency requirement, the initial regulation query text is corrected to obtain a target regulation query text that meets the business scenario consistency requirement.

[0116] Optionally, the target regulation query text determination module 320 is further specifically configured to construct a text correction prompt model corresponding to the initial regulation query text, and determine the corrected text corresponding to the initial regulation query text based on the text correction prompt model and a pre-constructed second processing model; re-perform the business scenario consistency verification on the corrected text corresponding to the initial regulation query text, and when the corrected text does not meet the business scenario consistency requirement, re-correct the corrected text until a corrected text that meets the business scenario consistency requirement is obtained.

[0117] Optionally, the target regulation query text determination module 320 is further specifically configured to determine the cumulative correction times corresponding to the corrected text during the correction process of the initial regulation query text; when the cumulative correction times corresponding to the corrected text exceed the preset correction times threshold and the corrected text does not meet the business scenario consistency requirement, the corrected text with the highest similarity score corresponding to the business scenario similarity determination result among the generated corrected texts is determined as the target regulation query text.

[0118] Optionally, the reference data determination module 330 is specifically configured to obtain a business rule knowledge base and a regulation knowledge base; retrieve in the business rule knowledge base based on the text to be processed to determine the reference business rule data corresponding to the text to be processed; and retrieve in the regulation knowledge base based on the target regulation query text to obtain the reference regulation data corresponding to the target regulation query text.

[0119] Optionally, the target processing result determination module 340 is specifically configured to obtain a pre-constructed answer generation prompt template, and construct an answer generation prompt model based on the text to be processed, the target regulation query text, the reference business rule data, the reference regulation data, and the pre-constructed answer generation prompt template; input the answer generation prompt model into the pre-constructed first processing model, and output the target processing result of the text to be processed by the pre-constructed first processing model.

[0120] The intelligent question-answering data processing device provided by the embodiments of the present invention can execute the intelligent question-answering data processing method provided by any embodiment of the present invention, and has the corresponding functional modules and beneficial effects for executing the method.

[0121] Embodiment 4

[0122] Figure 4 FIG. 7 is a schematic structural diagram of an electronic device provided by Embodiment 4 of the present invention. The electronic device 10 is intended to represent various forms of digital computers, such as, for example, laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as, for example, personal digital processors, cellular telephones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely exemplary and are not intended to limit the implementation of the present invention described and / or claimed herein.

[0123] As Figure 4 shown, the electronic device 10 includes at least one processor 11, and a memory communicatively connected to the at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc. The memory stores a computer program executable by the at least one processor. The processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 into the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other through a bus 14. The input / output (I / O) interface 15 is also connected to the bus 14.

[0124] A plurality of components in the electronic device 10 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a magnetic disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0125] The processor 11 may be various general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as the intelligent question-and-answer data processing method.

[0126] In some embodiments, the intelligent question-and-answer data processing method may be implemented as a computer program tangibly embodied in a computer-readable storage medium, such as the storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed onto the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the intelligent question-and-answer data processing method described above may be executed. Alternatively, in other embodiments, the processor 11 may be configured to execute the intelligent question-and-answer data processing method by any other suitable means (e.g., by means of firmware).

[0127] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SOCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special-purpose or general-purpose programmable processor that receives data and instructions from a storage system, at least one input device, and at least one output device, and transmits the data and instructions to the storage system, the at least one input device, and the at least one output device.

[0128] The computer program for implementing the intelligent question-and-answer data processing method of the present invention can be written in any combination of one or more programming languages. These computer programs can be provided to the processor of a general-purpose computer, a special-purpose computer, or other programmable intelligent question-and-answer data processing devices, such that when the computer programs are executed by the processor, the functions / operations specified in the flowcharts and / or block diagrams are implemented. The computer programs can be executed entirely on the machine, partially on the machine, executed partially on the machine and partially on a remote machine as an independent software package, or executed entirely on a remote machine or server.

[0129] Example 5

[0130] Example 5 of the present invention also provides a computer-readable storage medium storing computer instructions for causing a processor to execute an intelligent question-and-answer data processing method, the method comprising:

[0131] Obtain the text to be processed, and determine the initial regulatory inquiry text corresponding to the text to be processed;

[0132] Perform a business scenario consistency check on the initial regulatory inquiry text to obtain a target regulatory inquiry text that meets the business scenario consistency requirements;

[0133] Determine the reference business rule data corresponding to the text to be processed and the reference regulatory data corresponding to the target regulatory inquiry text;

[0134] Determine the target processing result corresponding to the text to be processed based on the text to be processed, the target regulatory inquiry text, the reference business rule data, and the reference regulatory data through a pre-constructed first processing model.

[0135] In the context of the present invention, a computer-readable storage medium may be a tangible medium that can contain or store a computer program for use by or in connection with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0136] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or a trackball) by which the user can provide input to the electronic device. Other kinds of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0137] The systems and techniques described herein can be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system can be interconnected with each other by digital data communication in any form or medium (e.g., a communication network). Examples of communication networks include: local area network (LAN), wide area network (WAN), blockchain network, and the Internet.

[0138] A computing system can include a client and a server. The client and the server are generally remote from each other and typically interact via a communication network. The client-server relationship is created by computer programs that run on respective computers and have a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system, and solves the defects of difficult management and weak business scalability existing in traditional physical hosts and VPS services.

[0139] It should be understood that various forms of the processes shown above can be used, steps can be reordered, added, or deleted. For example, the steps recited in the present invention can be executed in parallel, sequentially, or in a different order, as long as the desired results of the technical solution of the present invention can be achieved, and no limitation is imposed herein.

[0140] The above specific embodiments do not constitute a limitation on the protection scope of the present invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention shall be included within the protection scope of the present invention.

Claims

1. A method for processing intelligent question and answer data, characterized in that: include: Acquire the text to be processed, and determine the initial regulatory inquiry text corresponding to the text to be processed; Performing a business scenario consistency check on the initial regulatory query text to obtain a target regulatory query text that meets the business scenario consistency requirements; Determine reference business rule data corresponding to the text to be processed and reference regulatory data corresponding to the target regulatory inquiry text; A target processing result corresponding to the text to be processed is determined through a pre-constructed first processing model based on the text to be processed, the target regulatory inquiry text, the reference business rule data and the reference regulatory data.

2. The method according to claim 1, characterized in that The determining of the initial regulatory inquiry text corresponding to the text to be processed includes: The text to be processed is reconstructed based on a pre-built text reconstruction model to obtain an initial regulatory inquiry text corresponding to the text to be processed, wherein the initial regulatory inquiry text is a text that satisfies the regulatory inquiry terms.

3. The method according to claim 1, characterized in that The performing of a business scenario consistency check on the initial regulatory query text to obtain a target regulatory query text that meets the business scenario consistency requirements includes: Determine a business scenario similarity determination result between the to-be-processed text and the initial regulatory inquiry text, and determine a business scenario consistency verification result of the initial regulatory inquiry text based on the business scenario similarity determination result and a preset similarity threshold; If the business scenario consistency check result of the initial regulatory inquiry text satisfies the business scenario consistency requirement, the initial regulatory inquiry text is determined as the target regulatory inquiry text; If the business scenario consistency check result of the initial regulatory inquiry text is that it does not meet the business scenario consistency requirement, the initial regulatory inquiry text is corrected to obtain a target regulatory inquiry text that meets the business scenario consistency requirement.

4. The method according to claim 3, characterized in that The modifying process of the initial regulatory inquiry text to obtain a target regulatory inquiry text that meets the consistency requirements of the business scenario includes: Constructing a text correction prompt model corresponding to the initial regulatory inquiry text, and determining a correction text corresponding to the initial regulatory inquiry text based on the text correction prompt model and a pre-constructed second processing model; The revised text corresponding to the initial regulatory inquiry text is re-checked for business scenario consistency, and when the revised text does not meet the business scenario consistency requirements, the revised text is revised again until a revised text that meets the business scenario consistency requirements is obtained.

5. The method according to claim 4, characterized in that The method further comprises: During the revision process of the initial regulatory inquiry text, determining the cumulative number of revisions corresponding to the revised text; When the cumulative number of revisions corresponding to the revised text exceeds a preset revision number threshold and the revised text does not meet the business scenario consistency requirement, the revised text with the highest similarity score corresponding to the business scenario similarity determination result in the generated revised text is determined as the target regulatory inquiry text.

6. The method according to claim 1, characterized in that The determining of the reference business rule data corresponding to the to-be-processed text and the reference regulatory data corresponding to the target regulatory inquiry text includes: Obtain business rules knowledge base and regulatory knowledge base; Based on the text to be processed, the business rule knowledge base is searched to determine the reference business rule data corresponding to the text to be processed; and based on the target regulatory query text, the regulatory knowledge base is searched to obtain the reference regulatory data corresponding to the target regulatory query text.

7. The method according to claim 1, characterized in that The determining a target processing result corresponding to the text to be processed by a pre-built first processing model based on the text to be processed, the target regulatory inquiry text, the reference business rule data and the reference regulatory data includes: Obtain a pre-built answer generation prompt template, and build an answer generation prompt model based on the text to be processed, the target regulatory inquiry text, the reference business rule data, the reference regulatory data, and the pre-built answer generation prompt template; The answer generation prompt model is input into the pre-constructed first processing model, and the pre-constructed first processing model outputs the target processing result of the text to be processed.

8. An intelligent question-answering data processing device, characterized in that: include: An initial regulatory inquiry text determination module is used to obtain a text to be processed and determine an initial regulatory inquiry text corresponding to the text to be processed; A target regulatory inquiry text determination module is used to perform a business scenario consistency check on the initial regulatory inquiry text to obtain a target regulatory inquiry text that meets the business scenario consistency requirements; A reference data determination module, used to determine reference business rule data corresponding to the text to be processed and reference regulatory data corresponding to the target regulatory inquiry text; The target processing result determination module is used to determine the target processing result corresponding to the text to be processed through a pre-built first processing model based on the text to be processed, the target regulatory inquiry text, the reference business rule data and the reference regulatory data.

9. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the intelligent question and answer data processing method described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the intelligent question and answer data processing method according to any one of claims 1 to 7 when executed.

Images