Automatic auditing method, device and equipment and medium

Abstract

The invention relates to the technical field of artificial intelligence, and discloses an automatic auditing method, device and equipment and a medium, and the method comprises the steps: carrying out the preprocessing of obtained financial data, and inputting the preprocessed financial data into a GPT model; performing information extraction on the input financial data by the GPT model to obtain key information; analyzing and recognizing the financial data through a deep learning model and a pattern recognition algorithm to obtain an analysis and recognition result; and automatically generating an audit report according to the analysis and identification result and the key information. According to the method, the semantic comprehension and information extraction capability of the GPT model is utilized, the comprehension and analysis capability of financial data is improved, the detailed audit report is generated through the deep learning model and the mode recognition algorithm, the workload of manually compiling the report is reduced, and the audit efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the fields of artificial intelligence technology and financial technology, and in particular to an automated auditing method, device, equipment and medium. Background Art

[0002] In the field of FinTech, financial auditing is a key link in ensuring the authenticity, accuracy and legality of corporate accounting information. Traditionally, financial auditing relies on the personal experience and professional judgment of auditors. Although this method is effective to a certain extent, it also has obvious limitations. First, it is susceptible to the subjectivity of auditors and may not be able to fully and objectively evaluate all financial data. Second, in the face of the huge financial data of enterprises, it is difficult for manual audits to achieve comprehensive coverage, which increases the risk of missing important information. The automation of financial auditing also faces challenges. Auditors need to adapt to new technologies and learn how to use and interpret data analysis provided by automated tools. At the same time, audit institutions need to invest in technical infrastructure and ensure the security and reliability of automated audit systems. Therefore, how to use advanced technical means to achieve the automation of financial auditing has become a problem to be solved. Summary of the invention

[0003] The present invention provides an automated auditing method, apparatus, computer equipment and medium to solve the problem that the traditional label classification model is sensitive to changes in specific vocabulary and context, resulting in large fluctuations in label classification accuracy and high development costs.

[0004] In a first aspect, an automated audit method is provided, comprising:

[0005] Preprocess the acquired financial data and input the preprocessed financial data into the GPT model;

[0006] The GPT model extracts information from the input financial data to obtain key information;

[0007] Analyze and identify the financial data through a deep learning model and a pattern recognition algorithm to obtain analysis and identification results;

[0008] An audit report is automatically generated based on the analysis identification results and the key information.

[0009] In a second aspect, an automated auditing device is provided, comprising:

[0010] The data collection and preprocessing module is used to preprocess the acquired financial data and input the preprocessed financial data into the GPT model;

[0011] The semantic understanding and information extraction module is used to extract information and understand the semantics of the input financial data through the GPT model to obtain key information;

[0012] An analysis and recognition module, used to analyze and recognize the financial data through a deep learning model and a pattern recognition algorithm to obtain analysis and recognition results;

[0013] An audit report generation module is used to automatically generate an audit report based on the analysis and identification results and the key information.

[0014] In a third aspect, a computer device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above-mentioned automated audit method when executing the computer program.

[0015] In a fourth aspect, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above-mentioned automated audit method are implemented.

[0016] The scheme implemented by the above-mentioned automated audit method, device, computer equipment and storage medium involves a series of automated and optimized steps to improve the efficiency and accuracy of data collection. The specific process is as follows: First, the system needs to obtain the financial data to be audited, which may include the company's financial statements, transaction records, etc. The obtained raw financial data often needs to be preprocessed to facilitate subsequent analysis and calculation. The preprocessed data is input into the GPT (Generate Pre-trained Transformer) model. The GPT model extracts information from the input financial data and identifies and extracts key information from the data. The extracted key information is then analyzed and identified by a deep learning model and a pattern recognition algorithm. Finally, based on the analysis and identification results and key information, the system automatically generates an audit report. This report summarizes the problems and risks found in the audit process and provides decision support for auditors. The degree of automation of the audit process of this method has been significantly improved, reducing the reliance on manual analysis and improving the efficiency and accuracy of the audit. At the same time, this method also helps to reduce audit costs and discover potential financial problems. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments of the present invention will be briefly introduced below. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative labor.

[0018] Figure 1 is a schematic diagram of an application environment of an automated audit method in one embodiment of the present invention;

[0019] Figure 2is a flow chart of an automated audit method in one embodiment of the present invention;

[0020] Figure 3 is a structural diagram of an automated auditing device in one embodiment of the present invention;

[0021] Figure 4 is a schematic diagram of a structure of a computer device in one embodiment of the present invention;

[0022] Figure 5 FIG. 4 is another schematic diagram of the structure of a computer device in one embodiment of the present invention. DETAILED DESCRIPTION

[0023] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by technicians in the technical field of the present invention; the terms used in the specification of the application herein are only for the purpose of describing specific embodiments and are not intended to limit the present invention; the terms "including" and "having" and any variations thereof in the specification and claims of the present invention and the above-mentioned drawings are intended to cover non-exclusive inclusions. The terms "first", "second", etc. in the specification and claims of the present invention or the above-mentioned drawings are used to distinguish different objects, not to describe a specific order.

[0024] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present invention. The appearance of the phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0025] The automated audit method provided by the embodiment of the present invention can be applied in Figure 1In the application environment of, the client communicates with the server through the network. The server obtains the financial data to be audited through the client, which may include the company's financial statements, transaction records, etc. The server preprocesses the obtained raw financial data for subsequent analysis and calculation. The preprocessing steps may include data cleaning, format conversion, data normalization, etc. The preprocessed data is input into the GPT (Generative Pre-trained Transformer) model. The GPT model is a language model based on deep learning that can understand and generate natural language text. Information extraction uses the GPT model to extract information from the input financial data, identify and extract key information in the data, such as amount, date, transaction party, etc. The extracted key information is then analyzed and identified by the deep learning model and pattern recognition algorithm. Finally, based on the analysis and identification results and key information, the server automatically generates an audit report. This report summarizes the problems and risks found in the audit process and provides decision support for auditors. The server transmits the audit report to the client and displays it on the human-computer interaction screen. The automation level of the audit process of this method has been significantly improved, reducing the reliance on manual analysis and improving the efficiency and accuracy of the audit. At the same time, this method also helps to reduce audit costs and discover potential financial problems. Among them, the client can be but is not limited to various personal computers, laptops, smart phones, tablet computers and portable wearable devices. The server can be implemented with an independent server or a server cluster composed of multiple servers. The present invention is described in detail below through specific embodiments.

[0026] See also Figure 2 As shown, Figure 2 A flowchart of an automated audit method provided by an embodiment of the present invention includes the following steps:

[0027] S10, preprocessing the acquired financial data, and inputting the preprocessed financial data into the GPT model;

[0028] S20, the GPT model extracts information from the input financial data to obtain key information;

[0029] S30, analyzing and identifying the financial data through a deep learning model and a pattern recognition algorithm to obtain an analysis and recognition result;

[0030] S40: Automatically generate an audit report based on the analysis and identification results and the key information.

[0031] In this embodiment, the financial data of the enterprise is automatically collected and sorted through the Internet and the enterprise system interface. These data include financial statement data and transaction data. Using the semantic understanding ability of the GPT model, these financial data are deeply analyzed and key information such as accounting subjects, transaction amounts, time, location, etc. are extracted. By using deep learning and pattern recognition algorithms, the system can identify anomalies and potential risks in financial data, detect violations such as financial fraud and insufficient internal control, and obtain analysis and recognition results.

[0032] Based on these analysis and identification results and key information, the system automatically generates a detailed audit report. This process not only improves the ability to understand and analyze financial data, but also greatly reduces the workload of manual report writing and improves audit efficiency by automating the generation of audit reports. In addition, automated audit reports can also provide specific suggestions and improvement measures for the problems found, providing enterprises with more comprehensive risk management and decision-making support.

[0033] In order to improve the security and flexibility of the system, the embodiment also includes setting up data permission management, which is a mechanism for managing the access rights and operation rights of different users. In addition, the system also supports multilingual applications and can support the analysis and identification of financial statements and transaction data in different languages, which enables the system to serve a wider user group. Through these comprehensive measures, the system not only improves the automation and intelligence level of financial artificial intelligence, but also enhances the security and multilingual support capabilities of the system, thereby meeting the needs of enterprises for financial auditing in the context of globalization.

[0034] Specific implementation methods of multi-language support, the system is able to handle the input and analysis of financial statements and transaction data in multiple languages. According to the characteristics and rules of each language, the system will adjust its models and algorithms accordingly to ensure the accuracy and effectiveness of the analysis.

[0035] In terms of rights management, the implementation methods include designing a user rights management system to distinguish the access rights and operation rights of different users. The system implements an identity authentication mechanism to ensure that only authorized users can access the system and data. There are various means of identity authentication, including but not limited to fingerprints, passwords, face and pupil biometrics. The use of these biometrics provides additional security for the system.

[0036] Authentication mechanisms include both traditional password input and more advanced biometric technologies. For example, fingerprint recognition is performed through a dedicated fingerprint sensor, while face recognition uses a camera to capture the user's facial features for verification. Iris recognition uses a dedicated device to scan the user's pupils, providing a higher level of security verification.

[0037] When authentication fails, the system implements limits on the number of incorrect attempts, including account lockout and time limits. Account lockout means setting a maximum number of times a user can enter incorrect credentials in a row, such as 5 times. Once this limit is exceeded, the account will be locked and access must be restored through additional verification or administrator intervention. Time limits mean that after multiple authentication errors, the system will require the user to wait for a specified period of time (such as 10 minutes) before trying to log in again.

[0038] The system also provides immediate feedback when users enter incorrect credentials, such as "wrong username or password", and sets up warning and notification mechanisms. When multiple failed login attempts are detected, the system automatically sends warning notifications to users, such as via email or SMS, to remind users to pay attention to account security.

[0039] To further protect account security, if a user fails to verify multiple times, the system will provide a secure process to reset the password or verify the identity, such as through a verification code or security questions. After multiple failures, the system can require the user to perform additional verification, such as sending a one-time password (OTP) to the user's registered email or mobile phone number, or requiring the user to answer security questions to verify the identity.

[0040] In addition, the system monitors the user's login location, device, and time in real time, and identifies anomalies based on the user's historical behavior. By implementing these authentication mechanisms and error handling strategies, the system can significantly improve its security, ensuring that only authorized users can access the system and its data. At the same time, by promptly reminding users to pay attention to account security, the system helps prevent potential security threats.

[0041] In one embodiment, the obtaining of financial data, preprocessing the financial data, and inputting the preprocessed financial data into the GPT model includes:

[0042] Automatically collect transaction data and financial statement data through data collection interface;

[0043] The transaction data and financial statement data are preprocessed, and the preprocessed transaction data and financial statement data are input into the GPT model.

[0044] In this embodiment, the system will design and develop special data collection interfaces before data collection. These interfaces enable the system to automatically obtain the required transaction data and financial statement data from different data sources through the Internet and enterprise system interfaces. The collected data will undergo a series of preprocessing operations to ensure the quality and availability of the data. The preprocessing steps include data cleaning, standardization, removal of errors and redundant data, etc. Data cleaning refers to identifying and correcting errors and inconsistencies in data, such as removing duplicate records, correcting obvious input errors, etc. Standardization is to convert data into a unified format for subsequent processing and analysis. Removing errors and redundant data helps to improve the accuracy and consistency of data.

[0045] After preprocessing, the cleaned and standardized financial data is input into the GPT (referring to GPT or other related models) model for further audit analysis. Through automated data collection and preprocessing, the system can quickly acquire and process a large amount of financial statements and transaction data, thereby improving the efficiency and effectiveness of audit work. This automated process not only speeds up artificial intelligence, but also helps reduce human errors and ensure the reliability of audit results.

[0046] Specifically, designing and developing data collection interfaces to automatically collect corporate financial statements and transaction data can use a variety of technologies and tools:

[0047] 1. API (Application Programming Interface)

[0048] RESTful API: Using the RESTful architecture, data exchange is achieved through the HTTP protocol, which is suitable for obtaining financial data and reports. If the company's financial system has a RESTful API, data can be obtained with simple HTTP requests.

[0049] SOAP API: Communicate with financial management systems that support SOAP through the SOAP protocol to obtain financial and transaction data.

[0050] 2. Database connection

[0051] SQL Database: Directly connect to relational databases via ODBC (Open Database Connectivity) or JDBC (Java Database Connectivity) for data query and extraction.

[0052] NoSQL database: If the enterprise uses a document-based or key-value NoSQL database, use the corresponding database driver (such as MongoDB driver) to query data.

[0053] 3. Crawler technology

[0054] Web crawler: By writing a crawler program, you can automatically access the company's public webpage and crawl financial data. If the company does not provide an API, this method can effectively obtain public information.

[0055] Data parsing: Use HTML parsing libraries such as Beautiful Soup or Scrapy to extract specific tabular data or information.

[0056] 4. ETL Tools

[0057] Extract-Transform-Load (ETL) tools: Use ETL tools (such as Apache NiFi, Talend, Informatica) to automatically extract data from different data sources, perform data transformation, and load it into the data warehouse for analysis.

[0058] 5. Data file extraction

[0059] CSV / Excel import: Design interfaces to support data extraction from files in CSV, Excel and other formats, and perform data parsing and loading.

[0060] File monitoring: Establish a file monitoring mechanism to detect changes in financial statement files generated within the company and capture new data in a timely manner.

[0061] 6. Middleware Technology

[0062] Message queue: Use message queue systems (such as RabbitMQ, Kafka) to send data from enterprise systems to ensure the real-time and reliability of data collection.

[0063] Integration Platform: Use an integration platform (such as MuleSoft, Apache Camel) to integrate multiple data sources and protocols.

[0064] 7. Data synchronization tools

[0065] Data mirroring and synchronization tools: Use data mirroring or synchronization tools (such as Fivetran, Stitch) to achieve real-time or periodic synchronization between source data and target storage.

[0066] 8. Safety Technology

[0067] Authentication and authorization: Implement security mechanisms in the data collection interface to ensure limited access to sensitive data, such as using authentication methods such as OAuth2.0.

[0068] Data encryption: Encrypt data during transmission to protect data security.

[0069] Through the above technical means, the designed and developed automated data acquisition interface can effectively and reliably collect and process corporate financial statements and transaction data, and support subsequent data analysis and decision-making.

[0070] In one embodiment, the GPT model extracts information from the input financial data to obtain key information, including:

[0071] Use the pre-trained GPT model to perform semantic understanding and information extraction on the input financial data and output the extraction results;

[0072] The extraction results are analyzed, and key information is extracted according to the extraction results.

[0073] In this example, the system uses a pre-trained GPT model to perform in-depth semantic understanding and key information extraction on financial data. The first step is to collect and clean the financial data to ensure the uniformity and integrity of the data format. Next, prepare the text content to be extracted, such as financial statements or transaction records, and format it into a format suitable for model input.

[0074] In order to improve the accuracy of information extraction, the system may fine-tune the GPT model and use annotated data sets to train the model so that it can better understand the specific language in the financial field. The model is guided to extract information by designing prompts. For example, prompts in the question format can be used to clarify the type of information to be extracted, such as "accounting subject", "transaction amount", "time", etc.

[0075] After model training and fine-tuning, the formatted text is input into the GPT model for inference. The model will generate output containing key information based on its understanding of the context. The system will then parse the model output results to extract key information, and regular expressions or other text processing methods can be used to further refine the results.

[0076] Finally, the system compares the model output with the results of manual annotation to evaluate the accuracy. It can further optimize the input prompts or fine-tune the model parameters based on the feedback to improve the extraction effect. Through this series of steps, the system can effectively use the GPT model to accurately extract key information from financial data, ensuring comprehensive analysis of the text and accurate acquisition of specific information.

[0077] In one embodiment, the analyzing and identifying the financial data by using a deep learning model and a pattern recognition algorithm to obtain an analysis and recognition result includes:

[0078] Collecting historical financial data and extracting financial features from the historical financial data;

[0079] Training the deep learning model using the financial features until the model converges;

[0080] Inputting the financial data into the trained deep learning model for analysis to obtain analysis results;

[0081] Inputting the financial data into a pattern recognition algorithm for recognition to obtain a recognition result;

[0082] The analysis result and the recognition result are fused to obtain the analysis and recognition result.

[0083] In this embodiment, the system first collects historical financial data, including normal and abnormal cases, to ensure that the data sample fully covers a variety of possible abnormal situations. Then, the historical financial data is cleaned, missing values ​​are processed, redundant information is removed, and standardized to keep the data within the same range, thereby ensuring the accuracy and consistency of the data.

[0084] Next, the system extracts industry-related features from the financial data, such as financial ratios, transaction frequencies, abnormal transaction amounts, etc. To improve the efficiency of the model, the system may use principal component analysis (PCA) to reduce the data dimension, or use normalization to increase the convergence speed of the model.

[0085] Depending on the complexity of the problem, the system selects the appropriate deep learning model architecture. For example, for simpler classification problems, a multi-layer perceptron (MLP) may be used. If the data contains time series features, a convolutional neural network (CNN) may be considered. For sequential data, a recurrent neural network (RNN) or a long short-term memory network (LSTM) may be used to capture dependencies in the time dimension.

[0086] After that, the system divides the historical financial data into training sets, validation sets, and test sets, and selects appropriate loss functions (such as cross entropy) and optimization algorithms (such as Adam or SGD) according to the task. When training the model using the training set, the system optimizes the model performance by adjusting hyperparameters (such as learning rate, batch size), and evaluates the model performance based on indicators such as precision, recall, and F1 score.

[0087] The system adjusts model hyperparameters based on the evaluation results, or tries different model architectures to improve performance. Over time, the system continues to collect more historical financial data and uses this data to train and optimize the model, regularly evaluate and update the model to improve the accuracy and efficiency of financial data analysis and risk identification.

[0088] Finally, the system inputs the financial data into the trained deep learning model for analysis and obtains analysis results, including anomalies and potential risks in the financial data. At the same time, the system inputs the financial data into the pattern recognition algorithm for identification, using pattern recognition algorithms and statistical methods to detect financial fraud, inadequate internal control and other violations.

[0089] Furthermore, the system processes financial data through carefully selected pattern recognition algorithms, which may include clustering algorithms (such as K-means or DBSCAN), classification algorithms (such as support vector machines SVM, decision trees) or neural networks. In addition, the system uses statistical tests (such as t-tests, chi-square tests) to evaluate whether the anomalies found are statistically significant. Control charts (such as X-bar charts) are used to monitor data trends in order to identify points that exceed normal limits.

[0090] The system aggregates and analyzes financial ratios and daily transaction consumption to identify time periods of abnormal fluctuations. At the same time, it defines what is abnormal behavior in context, such as a single transaction above a certain amount or frequent abnormal account activity. Through information fusion technology, the system combines the analysis results and the recognition results to obtain the final analysis and recognition results.

[0091] In order to effectively design deep learning models and combine pattern recognition and statistical methods, the system collects financial data from multiple channels. Data sources include internal enterprise data, such as bills, income, expenditures, balance sheets, cash flow statements, etc. extracted from the ERP system; transaction records, including customer purchase records, supplier invoices, and financial audit results; and related data in the human resources management system involving salaries, costs, employee expenses, etc.

[0092] External market data is also an important data source, including average financial indicators and trends of relevant industries obtained from industry databases, reports released by market research companies, and public financial data, such as public company annual reports and financial reports provided by government agencies and regulatory agencies.

[0093] In addition, the system also obtains financial data of other companies in the industry through data exchange and cooperation, and conducts industry comparison and analysis. Surveys and questionnaires are also means of collecting user purchasing behavior and preference data, which can affect the financial decision of the product. Participating in industry conferences and exchanges can collect relevant market and financial information.

[0094] The system also uses API interfaces provided by third parties to obtain the latest financial data, such as those provided by financial information service companies. Social media and network data are also one of the collection channels, through which we can obtain the public opinion response of enterprises and industries and understand the potential impact of market dynamics on financial data.

[0095] Through comprehensive data collection from the above-mentioned multiple channels, the system can provide a rich source of financial data for model training and optimization, thereby improving the accuracy and effectiveness of the model.

[0096] In one embodiment, the automatically generating an audit report according to the analysis and identification results and the key information includes:

[0097] Generate an audit report based on the key information;

[0098] Integrate the analysis and identification results into the audit report and update the audit report.

[0099] In this embodiment, detailed audit results are generated based on the extracted key information, and the analysis results are integrated into the audit report through automatic system analysis and in-depth analysis and judgment using other technical means, providing relevant analysis and judgment basis to generate the audit report. In a specific embodiment, a timed task or event trigger mechanism can be designed to generate audit reports and suggestions regularly or when a specific event occurs, or according to the set generation rules and requirements, the corresponding report is automatically generated, and the necessary formatting and typesetting are performed.

[0100] Furthermore, specific methods for integrating analytical identification results into audit reports include:

[0101] 1. Automatic analysis by the system

[0102] Data integration: Automatically summarize the data results of identified risks and violations and integrate them into the corresponding parts of the audit report.

[0103] Pattern recognition: The model identifies patterns of abnormal behavior, and the system can provide detailed explanations of these patterns, such as time patterns, transaction frequency, abnormal amounts, etc.

[0104] Risk Rating: Automatically generate a rating (high, medium, low) for each identified risk and embed the rating results directly into the report.

[0105] 2. Statistical methods and data analysis

[0106] Descriptive statistics: Use statistical tools to analyze data characteristics and provide information such as frequency analysis of violations and amount distribution to support conclusions.

[0107] Comparative analysis: Compare current data with historical data to identify trend changes and analyze the risks they may pose to the company's financial situation.

[0108] 3. Visualization Analysis

[0109] Dashboards and Charts: Use data visualization tools, such as charts and dashboards, to present risk analysis results and help readers more intuitively understand the risks and violations found.

[0110] Heatmaps and cluster analysis: Use heatmaps to show concentrated areas of risk, and cluster analysis to show clustering of data points, thus helping to identify high-risk areas in a specific time period or department.

[0111] 4. Analysis by Domain Experts

[0112] Expert opinion: Combined with the analysis and judgment of domain experts, manual review and supplementation of automatic analysis results are carried out to provide more in-depth background information and business context.

[0113] Case study: Select representative cases, conduct in-depth analysis and combine them with the results of automatic detection for comparison and verification to enhance the accuracy of analysis.

[0114] 5. Rules Engine

[0115] Compliance check rules: Use the preset rule engine to check whether financial data complies with industry standards and laws and regulations, and identify potential compliance risks in a timely manner.

[0116] 6. Feedback and Iteration

[0117] Result feedback mechanism: Through feedback from manual audits, the system analysis model is regularly updated to ensure the accuracy and relevance of automatic analysis results.

[0118] It can ensure that the analysis and judgment basis of risk identification and violation detection results in the audit report are sufficient and reliable, thereby providing effective support for decision makers.

[0119] The audit report specifically includes the following contents:

[0120] 1. Report basic information:

[0121] Report title: such as "Enterprise Financial Audit Report".

[0122] Company Name: The name of the audited company.

[0123] Audit Date: The date the report was prepared.

[0124] Auditor information: the person responsible for the audit and his / her position.

[0125] 2. Audit objectives and scope;

[0126] Audit purpose: Clarify the purpose of the audit, such as verifying the authenticity and compliance of financial data.

[0127] Audit Scope: Describes the time period and range of financial data covered by the audit.

[0128] 3. Data Overview:

[0129] Financial Data Summary: A brief summary of the main financial data, which may include key indicators such as revenue, expenditure, assets and liabilities.

[0130] Data source: Explain the data source and collection method.

[0131] 4. Key findings;

[0132] Abnormal Transactions: Lists the identified abnormal transactions, including: transaction amount, date and time, and the accounts or departments involved.

[0133] Risk points: Identified potential risks, such as financial fraud and inadequate internal control.

[0134] 5. Risk Assessment:

[0135] Nature of risk: Classify the identified risks, such as high risk, medium risk, and low risk.

[0136] Impact assessment: Explain the possible impact these risks may have on the company's financial reporting.

[0137] 6. Analysis of illegal behaviors;

[0138] Violations found: Detailed description of the violations detected, such as financial fraud, lack of internal control procedures.

[0139] Case Examples: Give examples to support the violations found.

[0140] 7. Suggestions for improvement:

[0141] Specific recommendations for each risk and violation: Propose improvement measures, including: strengthening internal controls, introducing new audit procedures, training relevant personnel,

[0142] 8. Audit conclusion:

[0143] Audit Summary: A summary of the overall audit work, including an overview of key findings and risks.

[0144] Recommended Actions: Recommended actions based on the audit findings.

[0145] 9. Appendix:

[0146] Detailed Data and Analysis: Provides detailed analysis results and additional charts of the main data.

[0147] References: List all references and regulations used in the audit process.

[0148] 10. Auditor’s signature:

[0149] Signature and Date: The auditor signs to ensure the authenticity and authority of the report.

[0150] The structure and content of the audit report generated based on the analysis and identification results and key information should ensure its comprehensiveness, accuracy and readability, and help corporate management make effective decisions and improvements.

[0151] In one embodiment, the obtaining of financial data, after preprocessing the financial data, further includes:

[0152] Acquiring financial data, preprocessing the financial data, and storing the preprocessed financial data in a database;

[0153] The financial data is monitored in real time, and when an abnormality occurs in the monitored financial data, a recovery strategy is implemented.

[0154] In this embodiment, before acquiring data, a database is designed and established to store and manage the collected financial statements and transaction data. Financial data is acquired, pre-processed, and stored in the database. Data stored in the database is backed up, and financial data is monitored in real time. When an abnormality occurs in the monitored financial data, a recovery strategy is implemented to ensure the security and integrity of the data. The database can also store the collected historical financial data, and the data management and storage system is used to store and manage the large amount of collected financial data, which is convenient for subsequent analysis and model training.

[0155] In one embodiment, after automatically generating an audit report according to the analysis and identification results and the key information, the method further includes:

[0156] Displaying the audit report on a human-computer interaction interface;

[0157] When there are risky behaviors in the audit report, an alarm prompt message is output through the human-computer interaction interface.

[0158] In this embodiment, the system designs and develops a user-friendly human-computer interaction interface, which enables users to interact with the system intuitively. The interface is mainly used to display audit reports, and provides the function of viewing audit reports and suggestions, while supporting users to query and export reports.

[0159] The design of the human-computer interaction interface includes a main navigation bar with links to various modules, such as data entry, report generation, risk analysis, user settings, etc., allowing users to quickly access the required functions. The dashboard serves as a central display panel that provides an overview of the system's main indicators and key data, including the latest audit results, risk status, etc.

[0160] The interface also includes an analysis button, which allows users to start the data analysis and audit process with one click. The data visualization function displays the data analysis results in a graphical way through charts and dashboards, such as bar charts, line charts, pie charts, etc., to help users intuitively understand trends and anomalies. Heat maps are used to display the concentration of high-risk areas or abnormal transactions, so that users can quickly identify problems.

[0161] The human-computer interaction interface also provides a data upload function, allowing users to upload financial statements and transaction data files, supporting multiple file formats (such as CSV, Excel). In addition, a form is provided for users to manually enter financial data to ensure the flexibility of data entry.

[0162] The data analysis and results display area clearly displays the analysis results, including risk identification, abnormal transactions, etc. The results can be displayed in the form of tables, lists or charts. The report generation function allows users to select specific data analysis results, generate detailed audit reports, and provide generation options, such as selecting the report scope and format (PDF, Word, etc.).

[0163] The report preview function allows users to preview the generated report online to confirm the content. The risk overview provides real-time monitoring of identified risks, including risk level, transactions involved, recommended improvement measures, etc. The history function displays past audit records and processing status, which is convenient for users to compare and investigate.

[0164] User settings and permission management functions include user account management, allowing users to register, log in, modify personal information and passwords. Permission management ensures that users with different roles are assigned different access rights according to their functions to ensure the security of the system.

[0165] The human-computer interaction interface also includes alarm settings, designing and implementing alarm and reminder mechanisms to promptly notify users of potential risks and violations found. Users can customize threshold settings, and the system notifies users of potential risks and violations found through pop-ups or emails. It also provides records of past alarms to help users understand the changing trends of risks.

[0166] The help document of this method provides guidance on the use of system functions and FAQs to help users use the system better. The system also provides online chat or customer support channels to solve problems encountered by users during use. Designing an interface with good user experience can enhance the usability of the system and help users conduct financial audits and risk management efficiently and conveniently.

[0167] It can be seen that the above scheme is suitable for scenarios that require a large amount of data collection and analysis, especially in financial fields such as digital marketing, user behavior analysis, and product optimization. First, the system needs to obtain the financial data to be audited, which may include the company's financial statements, transaction records, etc. The obtained raw financial data often needs to be preprocessed to facilitate subsequent analysis and calculation. The preprocessing steps may include data cleaning, format conversion, data normalization, etc. The preprocessed data is input into the GPT (Generative Pre-trained Transformer) model. The GPT model is a language model based on deep learning that can understand and generate natural language text. Information extraction: The GPT model extracts information from the input financial data, identifies and extracts key information from the data, such as amount, date, transaction party, etc. The extracted key information is then analyzed and identified by a deep learning model and pattern recognition algorithm. Finally, based on the analysis and identification results and key information, the system automatically generates an audit report. This report summarizes the problems and risks found in the audit process and provides decision support for auditors. The automation of the audit process of this method has been significantly improved, reducing the reliance on manual analysis and improving the efficiency and accuracy of the audit. At the same time, this approach also helps reduce audit costs and detect potential financial problems.

[0168] It should be understood that the order of execution of the steps in the above embodiment does not necessarily mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present invention.

[0169] In one embodiment, an automated auditing device is provided, which corresponds one-to-one to the automated auditing method in the above embodiment. Figure 3 As shown, the automated audit device includes a data collection and preprocessing module 101, a semantic understanding and information extraction module 102, an analysis and identification module 103, and an audit report generation module 104. The functional modules are described in detail as follows:

[0170] The data collection and preprocessing module 101 is used to preprocess the acquired financial data and input the preprocessed financial data into the GPT model;

[0171] The semantic understanding and information extraction module 102 is used to extract information and perform semantic understanding on the input financial data by the GPT model to obtain key information;

[0172] The analysis and identification module 103 is used to analyze and identify the financial data through a deep learning model and a pattern recognition algorithm to obtain an analysis and identification result;

[0173] The audit report generation module 104 is used to automatically generate an audit report based on the analysis and identification results and the key information.

[0174] In one embodiment, the data acquisition and preprocessing module 101 includes:

[0175] An interface collection unit, used to automatically collect transaction data and financial statement data through a data collection interface;

[0176] The artificial intelligence unit is used to preprocess the transaction data and financial statement data, and input the preprocessed transaction data and financial statement data into the GPT model.

[0177] In one embodiment, the semantic understanding and information extraction module 102 includes:

[0178] The extraction result unit is used to perform semantic understanding and information extraction on the input financial data through the pre-trained GPT model and output the extraction results;

[0179] The key information unit is used to parse the extraction result and extract key information according to the extraction result.

[0180] In one embodiment, the analysis and identification module 103 includes:

[0181] A financial feature unit, used to collect historical financial data and extract financial features from the historical financial data;

[0182] A model training unit, used for inputting the financial data into the trained deep learning model for analysis to obtain analysis results;

[0183] A recognition result unit, used for inputting the financial data into a pattern recognition algorithm for recognition to obtain a recognition result;

[0184] The analysis and recognition result unit is used to fuse the analysis result and the recognition result to obtain the analysis and recognition result.

[0185] In one embodiment, the audit report generation module 104 includes:

[0186] An audit result unit, used to generate an audit report based on the key information;

[0187] The audit report unit is used to integrate the analysis and identification results into the audit report and update the audit report.

[0188] In one embodiment, the automated auditing device includes a database module, specifically including:

[0189] A database unit, used to obtain financial data, pre-process the financial data, and store the pre-processed financial data in a database;

[0190] The data monitoring unit is used to monitor the financial data in real time and implement a recovery strategy when an abnormality occurs in the monitored financial data.

[0191] In one embodiment, the automated auditing device includes a human-computer interaction module, specifically including:

[0192] A report display unit, used to display the audit report on a human-computer interaction interface;

[0193] The alarm unit is used to output an alarm prompt message through the human-computer interaction interface when there is risk behavior in the audit report.

[0194] For the specific definition of the automated audit device, please refer to the definition of the automated audit method above, which will not be repeated here. Each module in the above-mentioned automated audit device can be implemented in whole or in part by software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0195] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Figure 4 As shown. The computer device includes a processor, a memory, a network interface and a database connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile and / or volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external client via a network connection. When the computer program is executed by the processor, it realizes the functions or steps on the service side of an automated audit method.

[0196] In one embodiment, a computer device is provided. The computer device may be a client, and its internal structure diagram may be as follows: Figure 5 As shown. The computer device includes a processor, a memory, a network interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external server through a network connection. When the computer program is executed by the processor, it realizes the functions or steps of the client side of an automated audit method

[0197] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the following steps when executing the computer program:

[0198] Preprocess the acquired financial data and input the preprocessed financial data into the GPT model;

[0199] The GPT model extracts information from the input financial data to obtain key information;

[0200] Analyze and identify the financial data through a deep learning model and a pattern recognition algorithm to obtain analysis and identification results;

[0201] An audit report is automatically generated based on the analysis identification results and the key information.

[0202] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:

[0203] Preprocess the acquired financial data and input the preprocessed financial data into the GPT model;

[0204] The GPT model extracts information from the input financial data to obtain key information;

[0205] Analyze and identify the financial data through a deep learning model and a pattern recognition algorithm to obtain analysis and identification results;

[0206] An audit report is automatically generated based on the analysis identification results and the key information.

[0207] It should be noted that the above functions or steps that can be implemented by the computer-readable storage medium or computer device can refer to the relevant descriptions on the server side and the client side in the aforementioned method embodiment. To avoid repetition, they will not be described one by one here.

[0208] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in the present application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0209] Those skilled in the art can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0210] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, a magnetic disk, or an optical disk), and includes a number of instructions for enabling a terminal device (which can be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to execute the methods described in each embodiment of the present invention.

[0211] It should be noted that if software tools or components other than those of the Company appear in the embodiments of the present application, they are only used for illustration and do not represent actual use. The above-described embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the above-mentioned embodiments, a person of ordinary skill in the art should understand that the technical solutions described in the above-mentioned embodiments can still be modified, or some of the technical features thereof can be replaced by equivalents; and these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. An automated audit method, characterized in that: include: Preprocess the acquired financial data and input the preprocessed financial data into the GPT model; The GPT model extracts information from the input financial data to obtain key information; Analyze and identify the financial data through a deep learning model and a pattern recognition algorithm to obtain analysis and identification results; An audit report is automatically generated based on the analysis identification results and the key information.

2. The automated audit method according to claim 1, characterized in that: The step of preprocessing the acquired financial data and inputting the preprocessed financial data into the GPT model includes: Automatically collect transaction data and financial statement data through data collection interface; The transaction data and financial statement data are preprocessed, and the preprocessed transaction data and financial statement data are input into the GPT model.

3. The automated audit method according to claim 1, characterized in that: The GPT model extracts information from the input financial data to obtain key information, including: Use the pre-trained GPT model to perform semantic understanding and information extraction on the input financial data and output the extraction results; The extraction results are analyzed, and key information is extracted according to the extraction results.

4. The automated audit method according to claim 1, characterized in that: The analyzing and identifying the financial data by using a deep learning model and a pattern recognition algorithm to obtain analysis and recognition results includes: Collecting historical financial data and extracting financial features from the historical financial data; Training the deep learning model using the financial features until the model converges; Inputting the financial data into the trained deep learning model for analysis to obtain analysis results; Inputting the financial data into a pattern recognition algorithm for recognition to obtain a recognition result; The analysis result and the recognition result are fused to obtain the analysis and recognition result.

5. The automated audit method according to claim 1, characterized in that: The automatically generating an audit report according to the analysis and identification results and the key information includes: Generate an audit report based on the key information; Integrate the analysis and identification results into the audit report and update the audit report.

6. The automated audit method according to claim 5, characterized in that: The preprocessing of the acquired financial data includes: Acquiring financial data, preprocessing the financial data, and storing the preprocessed financial data in a database; The financial data is monitored in real time, and when an abnormality occurs in the monitored financial data, a recovery strategy is implemented.

7. The automated audit method according to claim 1, characterized in that: After the audit report is automatically generated according to the analysis and identification results and the key information, the method further includes: Displaying the audit report on a human-computer interaction interface; When there are risky behaviors in the audit report, an alarm prompt message is output through the human-computer interaction interface.

8. An automated auditing device, characterized in that: include: The data collection and preprocessing module is used to preprocess the acquired financial data and input the preprocessed financial data into the GPT model; The semantic understanding and information extraction module is used to extract information and understand the semantics of the input financial data through the GPT model to obtain key information; An analysis and identification module, used to analyze and identify the financial data through a deep learning model and a pattern recognition algorithm to obtain an analysis and identification result; An audit report generation module is used to automatically generate an audit report based on the analysis and identification results and the key information.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the automated audit method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the automated audit method according to any one of claims 1 to 7 are implemented.

Images