Data auditing method, device and equipment, storage medium and computer program product

Abstract

The invention discloses a data auditing method, device and equipment, a storage medium and a computer program product, and relates to the technical field of data processing.The method comprises the steps that in response to input data input by a user, an auditing rule of the input data is determined based on the input data; based on the auditing rule, generating an auditing scheme of the input data through an intelligent language model; performing feature extraction operation and text vectorization operation on the input data to obtain a keyword vector of the input data; and performing auditing operation on the keyword vector based on the auditing scheme to obtain an auditing report of the input data. The auditing scheme of the input data is generated through the intelligent language model, the data auditing efficiency can be improved, and meanwhile the auditing accuracy is improved.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a data audit method, device, equipment, storage medium and computer program product. Background Art

[0002] Data review is a crucial part of most industries’ work, which involves the careful review of a large number of contracts and documents. However, since this work is currently mainly done manually, it is not only inefficient but also prone to errors. Despite the rapid development of technology in recent years, traditional automated review systems still have many shortcomings when handling this type of work.

[0003] Traditional automated review systems are mostly based on rule matching or simple text analysis technology. They often have difficulty accurately understanding and processing complex contexts and unstructured information. In addition, the processing capabilities of these systems are also very limited for documents with long lengths and many variables. In this case, the system cannot achieve deep understanding and intelligent decision-making, which affects the quality and efficiency of review work.

[0004] The above contents are only used to assist in understanding the technical solution of the present application and do not constitute an admission that the above contents are prior art. Summary of the invention

[0005] The main purpose of this application is to provide a data audit method, device, equipment, storage medium and computer program product, aiming to solve the technical problem of low efficiency in text data entry and audit operations.

[0006] To achieve the above purpose, the present application proposes a data audit method, which includes:

[0007] In response to input data input by a user, determining a review rule for the input data based on the input data;

[0008] Based on the audit rules, an audit plan for the input data is generated through an intelligent language model;

[0009] Performing feature extraction and text vectorization operations on the input data to obtain a keyword vector of the input data;

[0010] An audit operation is performed on the keyword vector based on the audit scheme to obtain an audit report of the input data.

[0011] In one embodiment, the step of generating an audit scheme for the input data through an intelligent language model based on the audit rules includes:

[0012] Determine the task instruction corresponding to the audit rule based on the audit rule;

[0013] Classify and mark the task instructions by using marking symbols to obtain marking instructions, and input the marking instructions into an intelligent language model;

[0014] Based on the audit rules and the marking instructions, an audit plan for the input data is generated through the intelligent language model.

[0015] In one embodiment, the step of performing a feature extraction operation and a text vectorization operation on the input data to obtain a keyword vector of the input data includes:

[0016] Performing a word segmentation operation on the input data to obtain a word segmentation sequence of the input data;

[0017] Extract keywords from the word segmentation sequence using the intelligent language model to obtain a keyword sequence;

[0018] The keyword sequence is text-vectorized to obtain a keyword vector of the input data, and the keyword vector is stored in a vector database.

[0019] In one embodiment, the step of performing an audit operation on the keyword vector based on the audit scheme to obtain an audit report of the input data includes:

[0020] Based on the input data and audit rules, configure the audit tool to audit the input data;

[0021] The audit tool is called based on the audit steps in the audit scheme to perform an audit operation on the keyword vector to obtain an audit report of the input data.

[0022] In one embodiment, the step of calling the audit tool to perform an audit operation on the keyword vector based on the audit steps in the audit scheme to obtain an audit report of the input data includes:

[0023] Based on the audit steps in the audit plan, locate the target data and audit tools corresponding to the audit steps;

[0024] Based on the audit steps, the audit tool is called to perform an audit operation on the target data.

[0025] In one embodiment, the method further comprises:

[0026] determining, based on the audit report, instructions for analyzing the audit report;

[0027] The analysis instruction is input into an intelligent language model, and the audit report is analyzed by the intelligent language model to obtain an analysis report of the audit report.

[0028] In addition, to achieve the above purpose, the present application also proposes a data audit device, the data audit device comprising:

[0029] A rule determination module, configured to determine, in response to input data input by a user, a review rule for the input data based on the input data;

[0030] A scheme determination module, used to generate an audit scheme for the input data through an intelligent language model based on the audit rules;

[0031] A data processing module, used for performing feature extraction and text vectorization operations on the input data to obtain a keyword vector of the input data;

[0032] The data audit module is used to perform an audit operation on the keyword vector based on the audit scheme to obtain an audit report on the input data.

[0033] In addition, to achieve the above objectives, the present application also proposes a data audit device, which includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program is configured to implement the steps of the data audit method described above.

[0034] In addition, to achieve the above-mentioned purpose, the present application also proposes a storage medium, which is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, the steps of the data audit method described above are implemented.

[0035] In addition, to achieve the above-mentioned purpose, the present application also provides a computer program product, which includes a computer program, and when the computer program is executed by a processor, the steps of the data audit method described above are implemented.

[0036] One or more technical solutions proposed in this application have at least the following technical effects:

[0037] The embodiments of the present application propose a data audit method, apparatus, device, storage medium and computer program product, which respond to the input data input by the user, determine the audit rules of the input data based on the input data; based on the audit rules, generate an audit plan for the input data through an intelligent language model; perform feature extraction and text vectorization operations on the input data to obtain a keyword vector of the input data; perform an audit operation on the keyword vector based on the audit plan to obtain an audit report on the input data. Generating an audit plan for input data through an intelligent language model can improve the efficiency of data auditing and improve the accuracy of auditing. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0039] In order to more clearly illustrate the embodiments of the present application 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, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0040] Figure 1 A flowchart of the first embodiment of the data review method of the present application is provided;

[0041] Figure 2 This is a schematic diagram of the module structure of the data audit device according to an embodiment of the present application;

[0042] Figure 3 Schematic diagram of the device structure of the hardware operating environment involved in the data audit method in the embodiment of the present application.

[0043] The purpose, features and advantages of this application will be further described in conjunction with the embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0044] It should be understood that the specific embodiments described herein are only used to explain the technical solutions of the present application and are not used to limit the present application.

[0045] In order to better understand the technical solution of the present application, a detailed description will be given below in conjunction with the accompanying drawings and specific implementation methods.

[0046] The main solution of the embodiment of the present application is: in response to input data input by a user, determining the audit rules of the input data based on the input data; based on the audit rules, generating an audit plan for the input data through an intelligent language model; performing feature extraction and text vectorization operations on the input data to obtain a keyword vector of the input data; performing an audit operation on the keyword vector based on the audit plan to obtain an audit report on the input data.

[0047] In this embodiment, for the convenience of description, the following description is made by taking the identification data audit device as the execution subject.

[0048] Data review is a crucial part of most industries’ work, which involves the careful review of a large number of contracts and documents. However, since this work is currently mainly done manually, it is not only inefficient but also prone to errors. Despite the rapid development of technology in recent years, traditional automated review systems still have many shortcomings when handling this type of work.

[0049] Traditional automated review systems are mostly based on rule matching or simple text analysis technology. They often have difficulty accurately understanding and processing complex contexts and unstructured information. In addition, the processing capabilities of these systems are also very limited for documents with long lengths and many variables. In this case, the system cannot achieve deep understanding and intelligent decision-making, which affects the quality and efficiency of review work.

[0050] This application generates an audit plan for input data through an intelligent language model, which can improve the efficiency of data auditing and at the same time improve the accuracy of the audit.

[0051] It should be noted that the execution subject of this embodiment can be a computing service device with data processing, network communication and program running functions, such as a tablet computer, a personal computer, a mobile phone, etc., or an electronic device capable of realizing the above functions, a data review device, etc. The following takes the data review device as an example to illustrate this embodiment and the following embodiments.

[0052] Based on this, the present application embodiment provides a data audit method, referring to Figure 1 , Figure 1 This is a flow chart of the first embodiment of the data review method of this application.

[0053] In this embodiment, the data audit method includes steps S10 to S30:

[0054] Step S10, in response to the input data input by the user, determining the review rules of the input data based on the input data.

[0055] It should be noted that the input data refers to the data input into the system by the user. In one embodiment of the present application, the input data is text data, for example, the input data may be a financial statement, a customer profile, a product description or any other information that needs to be reviewed.

[0056] In addition, it should be noted that the audit rules for input data are a series of standards or criteria developed based on the characteristics of the input data and business needs, and are used to determine the expected goals and requirements of the input data audit, which includes the scope, standards, key indicators, etc. of the audit.

[0057] Specifically, clarify the audit objectives and audit requirements for input data, including the audit scope, standards, key indicators, etc., and determine the audit rules for input data based on the audit objectives and audit requirements.

[0058] Step S20: Based on the audit rules, an audit plan for the input data is generated through an intelligent language model.

[0059] It should be noted that the intelligent language model refers to the large language model LLM (Large Language Model), an artificial intelligence algorithm based on deep learning, which can model natural language text by training a large amount of text data, learning the grammar, semantics and contextual information of the language. LLM has a wide range of applications in the field of natural language processing, including but not limited to text generation, text classification, machine translation, sentiment analysis, etc. Its core advantage lies in its ability to understand and generate natural language, and its strong context perception and generalization capabilities. During the reasoning audit process, LLM can automatically determine the applicable audit rules based on the input data entered by the user, and generate the corresponding audit plan. Then, the input data is audited according to the audit plan to check whether the data meets the requirements of the audit rules. Finally, a detailed audit report is generated to record the audit results, problems found, and possible suggestions or solutions.

[0060] In addition, it should be noted that the audit plan refers to a specific, step-by-step audit plan or strategy generated by an intelligent language model based on the audit rules determined by the input data entered by the user. This plan details how to audit the input data, including which audit steps to perform, which audit tools or methods to use, and how to generate an audit report based on the audit results. The audit plan details how to perform the audit operation based on the audit rules and audit tools preset by the user.

[0061] More specifically, the audit plan is a step-by-step audit plan automatically generated by the data audit device based on the task instructions preset by the user through the planning capability of the intelligent language model. The audit plan includes but is not limited to

[0062] Specifically, based on the audit rules, the task instructions corresponding to the audit rules are determined; the task instructions are classified and marked by marking symbols to obtain marking instructions, and the marking instructions are input into the intelligent language model; based on the audit rules and the marking instructions, an audit plan for the input data is generated by the intelligent language model.

[0063] Step S30: performing an audit operation on the keyword vector based on the audit scheme to obtain an audit report of the input data.

[0064] It should be noted that the audit report refers to the report data obtained after the intelligent language model audits the input data based on the audit plan. After calling the API of the audit tool, a result in JSON format will be returned. The content of the audit report includes but is not limited to the call status of the audit tool, success or failure; audit messages, more detailed information about the results of the audit tool call, including but not limited to successful confirmation, specific description of errors or other related notifications; data generation time, data format, timestamp of the audit tool call, etc.; execution time, execution steps, execution results, etc. of each step of the audit plan.

[0065] Specifically, based on the input data and the audit rules, an audit tool and corresponding tool parameters for auditing the input data are configured; based on the audit steps in the audit scheme, the audit tool is called to audit the keyword vector to obtain an audit report on the input data.

[0066] This embodiment uses the above scheme to determine the audit rules of the input data based on the input data in response to the input data input by the user; based on the audit rules, generate the audit scheme of the input data through the intelligent language model; perform feature extraction and text vectorization operations on the input data to obtain the keyword vector of the input data; perform audit operations on the keyword vector based on the audit scheme to obtain the audit report of the input data. Generating the audit scheme of the input data through the intelligent language model can improve the audit efficiency of the data and improve the accuracy of the audit.

[0067] Based on the above implementation scheme, in a feasible implementation manner, the audit scheme for generating the input data based on the audit rules through the intelligent language model includes steps S21 to S23:

[0068] Step S21, determining the task instruction corresponding to the audit rule based on the audit rule.

[0069] It should be noted that the task instructions refer to the input text or instructions used to guide the intelligent language model to generate the audit plan for input data. The task instructions include but are not limited to the scope, standards, key indicators of the audit; and the configuration information of the required audit tools, such as the interface URL (Uniform Resource Locator), interface input parameters, interface output parameters, etc.

[0070] Among them, interface input parameters refer to the data received by the interface of the audit tool, which is used to perform specific operations or calculations. The values ​​of interface input parameters are required by the called function, and they are usually sent to the interface through HTTP requests (such as GET, POST, etc.). Interface input parameters can include various types of data, such as strings, numbers, objects, etc. Interface output parameters are the data returned after the interface performs an operation, which is used to provide results or information to the caller; the values ​​of output parameters are required by the calling function, and they are usually returned to the caller through HTTP responses; output parameters can include the results of the operation, status code, error information, etc.; in interface design, the output parameters need to be clearly defined and described to ensure that the caller can correctly parse and use the returned data; at the same time, the output parameters also need to be verified and formatted to ensure the accuracy and readability of the data. Interface URL refers to the network endpoint address provided by the audit tool, through which the system communicates with the audit tool; when configuring the audit tool, the user needs to enter the interface URL of the tool so that the system can accurately call the tool in the subsequent execution process; the interface URL usually includes the protocol (such as HTTP or HTTPS), domain name or IP address, port number (optional) and path.

[0071] In one embodiment of the present application, the task instruction uses the prompt word of the intelligent language model as the task instruction, providing the necessary context and guidance for the intelligent language model so that it can accurately understand the audit rules and generate corresponding responses accordingly. The prompt words of the task instruction include but are not limited to the background and role based on which the audit plan is based, the step-by-step decomposition requirements of the audit operation, and the configuration description of the audit tool.

[0072] Specifically, analyze and understand the content and requirements of the audit rules, and determine a series of task instructions that need to be executed based on the audit rules. The prompt words of the task instructions include but are not limited to the background and role based on the audit plan, the step-by-step decomposition requirements of the audit operation, and the configuration description of the audit tool. The prompt words of the task instructions are determined based on the basic knowledge of the business corresponding to the input data and the data reflux of the historical related input data.

[0073] Step S22, classify and mark the task instructions by marking symbols to obtain marked instructions, and input the marked instructions into the intelligent language model.

[0074] It should be noted that the labeling instructions refer to the results obtained after classifying and labeling the task instructions. These labels are usually completed using specific marking symbols. The purpose of the labeling instructions is to provide clearer and more specific instructions to the intelligent language model so that the model can generate an audit plan based on these instructions and audit rules. Classification marking includes but is not limited to classification marking of data such as the description of the audit tool, the input parameters of the interface, and the output parameters of the interface.

[0075] Specifically, determine a set of marking symbols suitable for task instruction classification. These symbols can be numbers, letters, special characters, etc., which are used to distinguish different types of task instructions, and use the selected marking symbols to classify and mark the task instructions. After classification and marking, the task instructions are converted into marked instructions. The marked instructions contain all the information of the task instructions, and also provide additional context or instructions through marking symbols. The generated marked instructions are sorted into an input format acceptable to the intelligent language model. The sorted marked instructions are input into the intelligent language model. The model will generate an audit plan based on these instructions and built-in algorithms, knowledge bases, etc. For example, when marking with letters, the description of the audit tool is marked with "T"; the description of the input parameters of the interface is marked with "I"; and the output parameters of the interface are marked with the letter "O".

[0076] Step S23, based on the audit rules and the marking instructions, generate an audit plan for the input data through the intelligent language model.

[0077] Specifically, the intelligent language model will first parse and understand the input annotation instructions, combine the annotation instructions with the audit rules, determine the audit path planning of the audit plan, and then determine the audit steps of the audit plan, determine the target data and audit tools corresponding to the audit steps, and based on the annotation instructions and audit rules, the intelligent language model will generate a detailed audit plan. This plan may include multiple steps, conditional judgments, data operations, etc., to guide how to audit the input data.

[0078] Through the above-mentioned scheme, this embodiment generates an audit plan through an intelligent language model based on audit rules and marking instructions, which can more accurately match the characteristics and requirements of the input data and reduce the possibility of misjudgment and missed judgment.

[0079] Based on the above implementation scheme, in a feasible implementation manner, the step of performing feature extraction and text vectorization operations on the input data to obtain the keyword vector of the input data includes steps S41 to S43:

[0080] Step S41, performing word segmentation operation on the input data to obtain a word segmentation sequence of the input data.

[0081] It should be noted that word segmentation is the process of dividing a continuous text into independent words or phrases. A word segmentation sequence refers to the sequence of words or phrases obtained after word segmentation. For example, for the sentence "I like to eat apples", the possible word segmentation sequence after word segmentation is "I / like / eat / apples".

[0082] Specifically, a word segmentation algorithm is used to segment the input data text. The word segmentation algorithm can be a rule-based method or a machine learning-based method. Chinese word segmentation usually needs to take into account information such as word boundaries, parts of speech, and context. The text is cut into independent words or phrases to obtain a word segmentation sequence. In one embodiment of the present application, the input data is segmented into paragraphs, and then the segmented paragraphs are stored in a text retrieval engine ElasticSearch, which serves as a document library, and the paragraphs in the document library are segmented to obtain a word segmentation sequence.

[0083] Step S42: extract keywords from the word segmentation sequence using the intelligent language model to obtain a keyword sequence.

[0084] It should be noted that keywords are words or phrases extracted from the word segmentation sequence, which can summarize the subject or content of the input data. The keyword sequence refers to the sequence of keywords extracted from the word segmentation sequence by feature extraction algorithms (such as TF-IDF, TextRank, deep learning-based methods, etc.). For example, the keyword sequence that may be extracted from "I / like / eat / apple" is "like / apple".

[0085] Specifically, an intelligent language model (such as a deep learning model) is used to analyze the word segmentation sequence, extract titles and keywords that can summarize the text theme or content, and obtain a keyword sequence.

[0086] Step S43, performing text vectorization on the keyword sequence to obtain a keyword vector of the input data, and storing the keyword vector in a vector database.

[0087] It should be noted that keyword vector refers to converting each keyword in the keyword sequence into a vector. These vectors are usually obtained through some method (such as word embedding technology Word2Vec, BERT, etc.), which can capture the semantic relationship between words.

[0088] In addition, it should be noted that the vector database refers to a database specifically used to store vector data, which can efficiently process operations such as similarity calculation and indexing between vectors. In the field of natural language processing and machine learning, vector databases are usually used to store word vectors, sentence vectors, etc., for fast text retrieval and matching.

[0089] Specifically, each keyword in the keyword sequence is converted into a vector through word embedding technology to obtain a keyword vector of the input data, and the keyword vector is stored in a vector database. In one embodiment of the present application, the keyword sequence is text-vectorized through the vector model bgem3 to obtain a keyword vector of the input data, and the keyword vector is stored in a vector database, and ElasticSearch can be used as a vector database.

[0090] Through the above scheme, this embodiment converts the keyword sequence into a vector form, which can more intuitively represent text information and facilitate subsequent storage and retrieval; storing the keyword vector in a vector database can achieve efficient indexing and query, improve the speed and accuracy of data retrieval, and thus improve the efficiency of review.

[0091] Based on the above implementation scheme, in a feasible implementation manner, performing an audit operation on the keyword vector based on the audit scheme to obtain an audit report of the input data includes steps S31 to S32:

[0092] Step S31, based on the input data and the audit rules, configure an audit tool for performing an audit operation on the input data.

[0093] It should be noted that the audit tool refers to various tools configured and used to complete the audit task, such as document retrieval tools and equivalence comparison tools. The interface URL, interface input parameters, and interface output parameters of the audit tool need to be configured. The interface input parameters include the fields of the interface input parameters, the field description of the interface input parameters, whether they must be passed in, fixed values, and other information. The interface output parameters include the fields of the interface output parameters and the field description of the interface output parameters, so that the system can correctly parse the results when calling the tool.

[0094] Specifically, according to the audit rules, determine the required audit tools; configure the parameters of the audit tool, which include but are not limited to the tool's URL interface, interface input parameters (such as keywords, file paths, etc.), interface output parameters, output format, etc.

[0095] Step S32: Based on the audit steps in the audit scheme, the audit tool is called to perform an audit operation on the keyword vector to obtain an audit report of the input data.

[0096] Specifically, based on the specific steps and sequence of the audit in the audit plan, and the audit tools that need to be called in each step; call the corresponding audit tools in sequence according to the sequence of steps in the audit plan; for each audit tool, pass the keyword vector obtained based on the input data as an input parameter to the tool according to the configured parameters; wait for the tool to complete execution and obtain the output results. The output results may include the status of audit pass / fail, specific error or problem information, etc. Organize and summarize the execution results of all audit steps; analyze and judge the execution results according to the audit rules, and generate the final audit report; the content of the audit report includes but is not limited to the overall audit status of the input data (such as pass / fail), detailed results of each audit step, existing problems and suggestions, etc.; it also includes the background and role on which the audit plan is based, the input and output parameter descriptions of the audit tool interface, and the execution results of the steps in the audit plan.

[0097] More specifically, the audit plan is parsed through the intelligent language model to obtain the audit tool that needs to be called to enter the data, and then the text of the audit plan is parsed to obtain the interface input result of the audit tool. Then, the device executes the API call of the audit tool, returns the result in json format, and obtains the audit report after parsing.

[0098] For example, when the input data is a market research report, the audit rules are "1. Check whether the report contains the market size, growth rate and forecast data of the target market; 2. Verify whether the market trends and driving factors mentioned in the report are reasonable and supported by data; 3. Check whether there is a detailed analysis of competitors in the report, including market share, competitive advantages and disadvantages; 4. Check whether the report contains a description of the target customer group and an analysis of consumption habits; 5. Verify whether all data sources cited in the report are reliable and clearly marked", and then determine the task instructions based on these audit rules, input the task instructions into the intelligent language model; convert the market research report into a keyword vector and store it in the vector database. The intelligent language model generates an audit plan based on the task instructions, determines the audit tool based on the audit plan, and configures the parameters of the audit tool; the audit tool list includes tool A (data verification tool, used to verify the accuracy and rationality of digital data), tool B (trend analysis tool, used to analyze the rationality of market trends and driving factors), tool C (text analysis tool, used to analyze the integrity and logic of text content), and tool D (data tracing tool, used to track and verify the reliability of data sources). The audit plan is divided into four steps. The first step is to call tool A to use the data verification tool to verify the market size, growth rate and forecast data in the report. The input parameters are "the market size, growth rate and forecast data part in the market research report", and the output temporary result is "data verification result (correct / wrong, reason for error)". The second step is to call tool B to use the trend analysis tool to analyze whether the market trends and driving factors mentioned in the report are reasonable and supported by data. The input parameters are the market trends and driving factors part of the market research report, and the output temporary result is the trend analysis result (reasonable / unreasonable, unreasonable reasons and suggestions). The third step is to call tool C to use the text analysis tool to review whether there is a detailed analysis of competitors in the report, including market share, competitive advantages and disadvantages, as well as the description of the target customer group and consumption habits analysis. The input parameters are the competitor analysis and target customer group analysis parts in the market research report, and the output temporary result is the text analysis result (complete / incomplete, missing information points). The fourth step is to call tool D to use the data tracing tool to track and verify whether all the data sources cited in the report are reliable and clearly marked. The input parameters are the data reference part of the market research report, and the output temporary result is the data tracing result (reliable / unreliable, unreliable data points and suggestions). The four-step audit process is comprehensively analyzed through an intelligent language model, and finally an audit report is obtained.

[0099] Through the above scheme, this embodiment configures the audit tool and performs automated audit according to the audit plan, which significantly reduces the workload of manual audit and improves audit efficiency; the audit tool operates based on predefined audit rules to ensure that the audit process strictly follows established standards and processes, thereby improving the accuracy of the audit.

[0100] Based on the above implementation scheme, in a feasible implementation manner, the auditing step in the auditing scheme is used to call the auditing tool to perform an auditing operation on the keyword vector to obtain an audit report of the input data, including steps S321 to S322:

[0101] Step S321, based on the audit steps in the audit scheme, locate the target data and audit tools corresponding to the audit steps.

[0102] It should be noted that the audit step is the specific implementation link in the audit plan, which describes in detail how to perform the audit task. The specific content of each audit step includes target data, audit tools, and audit operations.

[0103] In addition, it should be noted that the target data refers to the specific information that needs to be reviewed or verified in the review step. This data may be raw data from different sources, or it may be data that has been initially processed or converted. For example, in the contract term review step, the target data is the contract signing entity, term time, and other data.

[0104] Specifically, in the process of auditing the input data based on the audit plan, the audit point is first identified, that is, the specific audit step is identified and located, and the path planning of the audit plan is reviewed through the intelligent language model. When the audit step is located, the target data and audit tools corresponding to the audit step are located. The target data can be located by performing segment polling or text retrieval on the vector database to locate the target data.

[0105] Step S322: Based on the audit step, the audit tool is called to perform an audit operation on the target data.

[0106] Specifically, based on the specific operation of the audit step, the audit tool is called to perform the operation corresponding to the audit step on the target data. During the audit operation execution, the audit process is monitored to ensure that the audit tool can operate normally and record any abnormalities or errors.

[0107] Based on the above implementation scheme, in a feasible implementation scheme, the data audit method further includes steps S50 to S60:

[0108] Step S50: determining analysis instructions for the audit report based on the audit report.

[0109] It should be noted that the analysis instruction refers to a piece of text or instruction designed in advance to guide the intelligent language model to perform self-reflection, which often contains information such as the content that needs to be analyzed by the intelligent language model, the analysis angle or method, and the expected output results.

[0110] Specifically, a structured text or instruction (Prompt) is designed as an analysis instruction to clearly instruct the intelligent language model on the content of the audit report that needs to be analyzed, including previous planning steps, results of executed operations, and any relevant contextual information.

[0111] Step S60: input the analysis instruction into an intelligent language model, analyze the audit report through the intelligent language model, and obtain an analysis report of the audit report.

[0112] Specifically, the analysis instructions are input into the intelligent language model to guide the model to analyze the audit report, determine whether the audit plan needs to update the execution steps, and determine whether the final results of the input and audit are successful. If problems are found, the system will make improvement suggestions to improve the efficiency and quality of the audit the next time it is executed. If it is found in reflection that the audit can be ended, the audit process will be terminated and the final audit results will be returned to the user. The results of the analysis are recorded to generate an analysis report.

[0113] Through the above scheme, this embodiment analyzes the audit report through an intelligent language model based on analysis instructions, which can enhance the scalability of the audit solution. As the audit requirements and rules change, the device can continuously optimize and expand the audit process through self-reflection to meet new audit rules.

[0114] It should be noted that the above examples are only used to understand the present application and do not constitute a limitation on the data review method of the present application. More simple transformations based on this technical concept are all within the scope of protection of the present application.

[0115] This application also provides a data audit device, please refer to Figure 2 , the data audit device comprises:

[0116] A rule determination module 201 is used to determine a review rule for the input data based on the input data in response to the input data input by the user;

[0117] A scheme determination module 202, for generating an audit scheme for the input data through an intelligent language model based on the audit rules;

[0118] The data processing module 203 is used to perform feature extraction and text vectorization operations on the input data to obtain a keyword vector of the input data.

[0119] The data audit module 204 is used to perform an audit operation on the keyword vector based on the audit scheme to obtain an audit report of the input data.

[0120] The data audit device provided by the present application adopts the data audit method in the above embodiment, which can solve the technical problem of low efficiency in the input and audit operation of text data. Compared with the prior art, the beneficial effects of the data audit device provided by the present application are the same as the beneficial effects of the data audit method provided by the above embodiment, and the other technical features in the data audit device are the same as the features disclosed in the above embodiment method, which will not be repeated here.

[0121] The present application provides a data audit device, which includes: at least one processor; and a memory that is communicatively connected to the at least one processor; wherein the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the data audit method in the above-mentioned embodiment one.

[0122] Reference below Figure 3 , which shows a schematic diagram of the structure of a data audit device suitable for implementing an embodiment of the present application. The data audit device in the embodiment of the present application may include but is not limited to mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions), PMPs (Portable Media Players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 3 The data audit device shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0123] like Figure 3As shown, the data audit device may include a processing device 1001 (e.g., a central processing unit, a graphics processor, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM: Read Only Memory) 1002 or a program loaded from a storage device 1003 to a random access memory (RAM: Random Access Memory) 1004. In RAM1004, various programs and data required for the operation of the data audit device are also stored. The processing device 1001, ROM1002, and RAM1004 are connected to each other through a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Generally, the following systems can be connected to the I / O interface 1006: an input device 1007 including, for example, a touch screen, a touch pad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; an output device 1008 including, for example, a liquid crystal display (LCD: Liquid Crystal Display), a speaker, a vibrator, etc.; a storage device 1003 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 1009. The communication device 1009 can allow the data audit device to communicate with other devices wirelessly or wired to exchange data. Although the data audit device with various systems is shown in the figure, it should be understood that it is not required to implement or have all the systems shown. More or fewer systems can be implemented or have alternatively.

[0124] In particular, according to the embodiments disclosed in the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the present application include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication device, or installed from a storage device 1003, or installed from a ROM 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the method of the embodiment disclosed in the present application are executed.

[0125] The data audit device provided by the present application adopts the data audit method in the above embodiment, which can solve the technical problem of low efficiency in the input and audit operation of text data. Compared with the prior art, the beneficial effects of the data audit device provided by the present application are the same as the beneficial effects of the data audit method provided by the above embodiment, and the other technical features in the data audit device are the same as the features disclosed in the method of the previous embodiment, which will not be repeated here.

[0126] It should be understood that the various parts disclosed in this application can be implemented by hardware, software, firmware or a combination thereof. In the description of the above embodiments, specific features, structures, materials or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0127] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0128] The present application provides a computer-readable storage medium having computer-readable program instructions (ie, computer programs) stored thereon, wherein the computer-readable program instructions are used to execute the data audit method in the above-mentioned embodiment.

[0129] The computer-readable storage medium provided in the present application may be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, systems or devices, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, 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 disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, system or device. The program code contained on the computer-readable storage medium may be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination of the above.

[0130] The computer-readable storage medium may be included in the data audit device; or may exist independently without being assembled into the data audit device.

[0131] The above-mentioned computer-readable storage medium carries one or more programs. When the above-mentioned one or more programs are executed by the data review device, the data review device: responds to the input data input by the user, and determines the review rules of the input data based on the input data; generates an audit plan for the input data through an intelligent language model based on the audit rules; performs feature extraction operations and text vectorization operations on the input data to obtain a keyword vector of the input data; and performs an audit operation on the keyword vector based on the audit plan to obtain an audit report on the input data.

[0132] Computer program code for performing the operations of the present application may be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0133] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present application. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0134] The modules involved in the embodiments described in this application may be implemented by software or hardware, wherein the name of the module does not constitute a limitation on the unit itself in some cases.

[0135] The readable storage medium provided by the present application is a computer-readable storage medium, which stores computer-readable program instructions (i.e., computer programs) for executing the above-mentioned data audit method, and can solve the technical problem of low efficiency in the input and audit operations of text data. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided by the present application are the same as the beneficial effects of the data audit method provided by the above-mentioned embodiment, and will not be repeated here.

[0136] The present application also provides a computer program product, including a computer program, which implements the steps of the above-mentioned data audit method when executed by a processor.

[0137] The computer program product provided by the present application can solve the technical problem of low efficiency in the input and review operations of text data. Compared with the prior art, the beneficial effects of the computer program product provided by the present application are the same as the beneficial effects of the data review method provided by the above embodiment, which will not be repeated here.

[0138] The above descriptions are only some embodiments of the present application, and are not intended to limit the patent scope of the present application. All equivalent structural changes made using the contents of the present application specification and drawings under the technical concept of the present application, or direct / indirect applications in other related technical fields are included in the patent protection scope of the present application.

Claims

1. A data audit method, characterized in that: The method comprises: In response to input data input by a user, determining a review rule for the input data based on the input data; Based on the audit rules, an audit plan for the input data is generated through an intelligent language model; Performing feature extraction and text vectorization operations on the input data to obtain a keyword vector of the input data; An audit operation is performed on the keyword vector based on the audit scheme to obtain an audit report of the input data.

2. The method according to claim 1, characterized in that The step of generating an audit scheme for the input data through an intelligent language model based on the audit rules includes: Determine the task instruction corresponding to the audit rule based on the audit rule; Classify and mark the task instructions by using marking symbols to obtain marking instructions, and input the marking instructions into an intelligent language model; Based on the audit rules and the marking instructions, an audit plan for the input data is generated through the intelligent language model.

3. The method according to claim 1, characterized in that The step of performing feature extraction and text vectorization operations on the input data to obtain a keyword vector of the input data includes: Performing a word segmentation operation on the input data to obtain a word segmentation sequence of the input data; Extract keywords from the word segmentation sequence using the intelligent language model to obtain a keyword sequence; The keyword sequence is text-vectorized to obtain a keyword vector of the input data, and the keyword vector is stored in a vector database.

4. The method according to claim 1, characterized in that The step of performing an audit operation on the keyword vector based on the audit scheme to obtain an audit report of the input data includes: Based on the input data and audit rules, configure the audit tool to audit the input data; The audit tool is called based on the audit steps in the audit scheme to perform an audit operation on the keyword vector to obtain an audit report of the input data.

5. The method according to claim 4, characterized in that The step of calling the audit tool to perform an audit operation on the keyword vector based on the audit steps in the audit scheme includes: Based on the audit steps in the audit plan, locate the target data and audit tools corresponding to the audit steps; Based on the audit steps, the audit tool is called to perform an audit operation on the target data.

6. The method according to claim 1, characterized in that The method further comprises: determining, based on the audit report, instructions for analyzing the audit report; The analysis instruction is input into an intelligent language model, and the audit report is analyzed by the intelligent language model to obtain an analysis report of the audit report.

7. A data audit device, characterized in that: The device comprises: A rule determination module, configured to determine, in response to input data input by a user, a review rule for the input data based on the input data; A scheme determination module, used to generate an audit scheme for the input data through an intelligent language model based on the audit rules; A data processing module, used for performing feature extraction and text vectorization operations on the input data to obtain a keyword vector of the input data; The data audit module is used to perform an audit operation on the keyword vector based on the audit scheme to obtain an audit report on the input data.

8. A data audit device, characterized in that: The device comprises: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program is configured to implement the steps of the data audit method according to any one of claims 1 to 6.

9. A storage medium, characterized in that: The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, the steps of the data audit method according to any one of claims 1 to 6 are implemented.

10. A computer program product, characterized in that The computer program product comprises a computer program, and when the computer program is executed by a processor, the steps of the data audit method according to any one of claims 1 to 6 are implemented.

Images