Tax risk display method, system, equipment and medium

Abstract

The invention provides a tax risk display method, system and device and a medium, and belongs to the field of tax risk analysis. The method comprises the following steps: extracting multi-modal data contained in cargo flow, contract flow, fund flow and invoice flow in a business process, constructing a transaction relation knowledge graph, and associating the multi-modal data with entities in the graph through a mapping rule base; performing consistency detection on the cargo flow, the contract flow, the fund flow and the invoice flow, determining a risk level in combination with historical risk events in the risk knowledge base, matching countermeasures from the risk knowledge base, associating tax regulations from the tax knowledge base, and generating a risk report; and generating video clips by adopting a text video technology, and splicing and merging the video clips to generate a target video. According to the method, the transaction relation knowledge graph is constructed through multi-modal data extraction, the risk indexes are quantified, the risk report is generated, and finally the result is converted into the video through the text video technology, so that the visual display of the tax risk is realized, and the accuracy and response efficiency of risk identification are improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of tax risk analysis, and particularly relates to a method, system, device and medium for displaying tax risks. Background Art

[0002] With the increasing number of invoices issued and received by enterprises, the potential tax-related risks are also increasing. For example, in some cases, it may not be known that the invoice is a false invoice and is considered a normal business, but as a result, administrative and criminal penalties may be faced due to suspected false invoicing. In order to detect and respond to tax risks, analysis and processing are carried out in the form of risk reports.

[0003] In the prior art, through machine learning algorithms, a large amount of data can be trained to learn complex business models, so as to train a risk report model. According to the input prompt words and combined with natural language processing, valuable information can be automatically extracted from a large amount of data and a text-based intelligent risk report can be quickly generated, greatly simplifying the risk analysis process, improving the availability of data, and significantly enhancing the accuracy of risk report generation.

[0004] However, most risk reports are presented in the form of text. When company leaders or business personnel receive the report to understand the risk situation and analyze the risks, it is easy to cause visual fatigue, and it is impossible to quickly and accurately grasp the business nodes where risks occur, and it is impossible to well understand the causes of risks. Summary of the Invention

[0005] The purpose of the embodiments of the present invention is to provide a method, system, device and medium for displaying tax risks, which are used to wholly or at least partially solve the technical problems in the prior art that the risk report presented in the form of text is easy to cause visual fatigue, and it is impossible to quickly and accurately grasp the business nodes where risks occur, and it is impossible to well understand the causes of risks.

[0006] In a first aspect, an embodiment of the present application provides a method for displaying tax risks, including: Extracting multi-modal data included in the physical flow of goods, contract flow, capital flow and invoice flow in the business process and constructing multi-modal data; Based on the association relationship, constructing a transaction relationship knowledge graph, and associating the multi-modal data with entities in the transaction relationship knowledge graph through a mapping rule library; Based on the transaction relationship knowledge graph, performing consistency detection on the physical flow of goods, contract flow, capital flow and invoice flow to obtain a detection result; Based on the detection result, determining a risk level in combination with historical risk events in the risk knowledge base, matching countermeasures from the risk knowledge base, and associating tax regulations from the tax knowledge base to generate a risk report; Use text-to-video technology to generate video clips from risk reports, splice and merge the video clips, and generate a target video for tax risk display.

[0007] Optionally, based on the association relationship, construct a transaction relationship knowledge graph, and associate the multimodal data with the entities in the transaction relationship knowledge graph through a mapping rule library, including: For the goods flow, identify information from logistics documents, and extract key information such as the shipper, consignee, goods description, and waybill number; For the contract flow, analyze the contract text through natural language processing technology, and extract key terms, contract parties, goods description, and amount information; For the funds flow, identify information from bank statements, and extract key information such as the transaction amount, transaction date, and counterparty; For the invoice flow, identify information from invoices, and extract key information such as the invoice number, amount, goods description, and invoicing date; Unify the extracted key information of the shipper, consignee, goods description, waybill number, key terms, contract parties, goods description, amount information, transaction amount, transaction date, counterparty key information, invoice number, amount, goods description, and invoicing date multimodal data into text format, create a mapping rule library for the multimodal data and the entities in the transaction relationship knowledge graph, and associate the extracted multimodal information with the entities in the knowledge graph.

[0008] Optionally, determine the risk level based on the detection result in combination with historical risk events in the risk knowledge base, including: Based on the transaction relationship knowledge graph, perform consistency detection on the goods flow, contract flow, funds flow, and invoice flow to determine the business nodes with inconsistent information in the goods flow, contract flow, funds flow, and invoice flow; Define abnormal indicators related to the risk of taxpayers issuing false invoices; Collect data related to tax risks in the business nodes, and divide the collected data related to tax risks into a training set and a test set. Among them, divide the test set into multiple samples corresponding to different risk levels; Set the number of decision trees and the maximum depth parameters in the initial random forest model, and use the training set for training. Use the trained random forest model to predict the test set to obtain the prediction category, prediction probability, risk impact degree, and index weight value; Construct a false invoice risk model based on the prediction category, prediction probability, risk impact degree, and index weight value, and perform a scan and evaluation on the false invoice risk model through the risk scanning function in combination with historical risk events to determine the risk level.

[0009] Optionally, the abnormal indicators at least include abnormal changes in taxpayers' tax burdens, abnormal changes in the usage of special invoices, abnormal dispersion of customers for sales invoices, abnormally high rates of voided special invoices, abnormal differences in the magnitude of ending inventory greater than paid-in capital, and abnormal differences in the magnitude of ending inventory and cumulative current income.

[0010] Optionally, after determining the risk level based on the detection results in combination with historical risk events in the risk knowledge base, matching countermeasures from the risk knowledge base, and associating tax regulations from the tax knowledge base to generate a risk report, the method for presenting the tax risk further includes: Define evaluation indicators for evaluating the generation results, and continuously adjust the parameter information according to these indicators; Divide the countermeasures retrieved from the risk knowledge base and the context information related to historical risk events into multiple paragraphs, evaluate the relevance for each paragraph, and then use the relevant information as input to continue the retrieval to obtain the optimal countermeasures and the context information related to historical risk events; Use the retrieved event description, risk level, optimal countermeasures, and context information as input to find relevant tax regulations in the tax knowledge base and generate a target risk report containing risk analysis and countermeasure plans; Among them, the risk knowledge base records event descriptions, countermeasures, context information related to events, and risk levels.

[0011] Optionally, use text-to-video technology to generate video clips from the risk report, and splice and merge the video clips to generate a target video for presenting tax risks, including: Use a pre-trained language model combined with a retrieval algorithm to extract key information from the generated risk report to form a summary, and based on the summary, use the pre-trained language model to generate the text content for video narration; Perform sentiment analysis on the text content for video narration through a sentiment analysis model to determine the sentiment tendency of the text; Based on the risk report, use a data visualization tool to generate risk analysis charts for display in the video; Convert the text content for video narration into voice narration, use text-to-video technology to integrate the text content for video narration, voice narration, and risk analysis charts to generate corresponding video clips, and then call the video synthesis function to splice and merge the video clips in sequence to generate a target video.

[0012] Optionally, converting the text content for video narration into voice narration includes: Convert the text content for video narration into a.txt file and provide an audio sample with a speech rate and intonation that match the style of tax risk content for a preset time period; Use text - to - speech synthesis to generate speech from the text in a.txt file, and use audio samples for voice cloning to obtain a voice narration.

[0013] In a second aspect, an embodiment of the present application further provides a tax risk display system, including: An extraction unit, configured to extract multimodal data included in the goods flow, contract flow, funds flow, and invoice flow in a business process and construct an association relationship between the multimodal data; A determination unit, configured to construct a transaction relationship knowledge graph based on the association relationship, and associate the multimodal data with entities in the transaction relationship knowledge graph through a mapping rule library; A detection unit, configured to perform consistency detection on the goods flow, contract flow, funds flow, and invoice flow based on the transaction relationship knowledge graph to obtain a detection result; A generation unit, configured to determine a risk level based on the detection result in combination with historical risk events in a risk knowledge base, match countermeasures from the risk knowledge base, and associate tax regulations from a tax knowledge base to generate a risk report; A display unit, configured to use text - to - video technology to generate video segments from the risk report, splice and merge the video segments, and generate a target video for tax risk display.

[0014] In a third aspect, an embodiment of the present application further provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, the steps of the above - mentioned tax risk display method are implemented.

[0015] In a fourth aspect, an embodiment of the present application further provides a storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the above - mentioned tax risk display method are implemented.

[0016] From the above technical solutions, it can be seen that the present invention has the following advantages: In the tax risk display method, system, device, and medium provided by the present application, in the above - mentioned embodiment, a transaction relationship knowledge graph is constructed through multimodal data extraction, risk indicators are quantified in combination with a machine learning model, and a risk report is generated. Finally, the analysis result is converted into a video containing voice, charts, and dynamic scenes through text - to - video technology, realizing the visual intelligent display of tax risks and improving the accuracy of risk identification and response efficiency. Description of the Drawings

[0017] To more clearly illustrate the technical solution of the present invention, the accompanying drawings required for description will be briefly introduced below. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other accompanying drawings can be obtained based on these drawings without creative efforts.

[0018] Figure 1 It is a flowchart of a method for displaying tax risks provided by an embodiment of the present invention; Figure 2 It is a flowchart of the association relationship of multimodal data provided by an embodiment of the present invention; Figure 3 It is a flowchart of determining a risk level provided by an embodiment of the present invention; Figure 4 It is the synthesis process of a tax risk display video provided by an embodiment of the present invention; Figure 5 It is a schematic structural diagram of a tax risk display system provided by an embodiment of the present invention; Figure 6 It is a schematic hardware structure diagram of an electronic device provided by an embodiment of the present invention. Detailed implementation manners

[0019] In the following detailed description, various embodiments of the present disclosure will be described more fully. The present disclosure may have various embodiments and adjustments and changes may be made therein. However, it should be understood that there is no intention to limit the various embodiments of the present disclosure to the specific embodiments disclosed herein, but the present disclosure should be understood to cover all adjustments, equivalents, and / or alternative solutions falling within the spirit and scope of the various embodiments of the present disclosure.

[0020] In the following, the term "include" or "may include" that may be used in various embodiments of the present disclosure indicates the existence of the disclosed function or operation and does not limit the addition of one or more functions or operations. In addition, as used in various embodiments of the present disclosure, the terms "include", "have" and their cognates are only intended to represent a specific feature, number, step, operation or combination of the foregoing items, and should not be understood to first exclude the existence or addition of one or more other features, numbers, steps, operations or combinations of the foregoing items.

[0021] In various embodiments of the present disclosure, the expression "or" or "at least one of A or / and B" includes any combination or all combinations of the recited words. For example, the expression "A or B" or "at least one of A or / and B" may include A, may include B, or may include both A and B.

[0022] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0023] Refer to Figure 1 The figure shown is a flowchart of a method for displaying tax risks in a specific embodiment, including the following execution steps: Step 100: Extract multimodal data included in the physical flow of goods, contract flow, capital flow, and invoice flow in the business process and construct the association relationship between the multimodal data.

[0024] Step 101: Based on the association relationship, construct a transaction relationship knowledge graph and associate the multimodal data with the entities in the transaction relationship knowledge graph through a mapping rule library.

[0025] Specifically, refer to Figure 2 As shown, when executing steps 100 and 102, the following steps can be specifically executed: S1000: For the physical flow of goods, identify the information on the logistics documents and extract the key information such as the shipper, consignee, description of goods, and waybill number.

[0026] S1001: For the contract flow, analyze the contract text through natural language processing technology and extract the key terms, contract parties, description of goods, and amount information.

[0027] S1002: For the capital flow, identify the information on the bank statement and extract the key information such as the transaction amount, transaction date, and counterparty.

[0028] S1003: For the invoice flow, identify the information on the invoice and extract the key information such as the invoice number, amount, description of goods, and invoice date.

[0029] S1004: Unify the multimodal data such as the extracted key information of the shipper, consignee, description of goods, waybill number, key terms, contract parties, description of goods, amount information, transaction amount, transaction date, counterparty key information, invoice number, amount, description of goods, and invoice date key information into text format, create a mapping rule library between the multimodal data and the entities in the transaction relationship knowledge graph, and associate the extracted multimodal information with the entities in the knowledge graph.

[0030] Step 102: Based on the transaction relationship knowledge graph, perform consistency detection on the physical flow of goods, contract flow, capital flow, and invoice flow to obtain the detection result.

[0031] It should be understood that each record in the risk knowledge base should include an event description (detailed description of the risk event), response measures, and context information (other relevant information, such as the time of occurrence of the event, the degree of impact, etc.) for the analysis of tax risks.

[0032] Step 103: Determine the risk level based on the detection results in combination with historical risk events in the risk knowledge base, match response measures from the risk knowledge base, and associate tax regulations from the tax knowledge base to generate a risk report.

[0033] Specifically, refer to Figure 3 As shown, when performing Step 103, the following steps can be specifically executed: S1030: Based on the transaction relationship knowledge graph, perform consistency detection on the physical flow, contract flow, fund flow, and invoice flow to determine the business nodes with inconsistent information in the physical flow, contract flow, fund flow, and invoice flow.

[0034] S1031: Define abnormal indicators related to the risk of taxpayers issuing false invoices.

[0035] Exemplarily, the abnormal indicators at least include abnormal changes in taxpayers' tax burdens, abnormal changes in the usage of special invoices, abnormal dispersion of customers of sales invoices, abnormally high rates of voided special invoices, abnormal differences in the magnitude of ending inventory greater than paid-in capital, and abnormal differences in the magnitude of ending inventory and current cumulative income.

[0036] S1032: Collect data related to tax risks in the business nodes, and divide the collected data related to tax risks into a training set and a test set. Among them, the test set is further divided into multiple samples corresponding to different risk levels.

[0037] S1033: Set the number of decision trees and the maximum depth parameter in the initial random forest model, and use the training set for training. Use the trained random forest model to predict the test set to obtain the predicted category, predicted probability, risk impact degree, and index weight value.

[0038] Exemplarily, collect data related to tax risks (such as tax regulations, policy documents, tax risk cases, tax-related statements, etc., or collect database data from internal systems, established knowledge graphs, etc.), and preprocess the data, including missing value filling, data standardization, etc. Divide the collected data into a training set and a test set, with 70% of the data as the training set and 30% of the data as the test set. Further divide the test set into 5 samples, corresponding to 5 risk levels respectively (Level 1 - no risk, Level 2 - low risk, Level 3 - medium risk, Level 4 - relatively high risk, Level 5 - high risk). Use random forest to set model parameters such as the number of decision trees and the maximum depth (the number of decision trees n_estimators is set to 100, the maximum depth max_depth is set to 10, and the minimum number of samples required for further division of internal nodes min_samples_split is set to 5). Train the training set data, use the trained random forest model to predict the test set data, obtain the predicted category and predicted probability, draw the ROC curve and PR curve based on the predicted probability, and then calculate the AUC and AUPR values, which can intuitively evaluate the classification ability of the model and determine the optimal threshold. Similarly, the likelihood of risk occurrence (extremely low, unlikely, possible, likely, almost certain) can be obtained using random forest, and the degree of risk impact (extremely low, low, medium, high, extremely high) and the index weight value can be obtained using decision trees.

[0039] S1034: Construct a risk model for false invoicing based on the predicted category, predicted probability, degree of risk impact, and index weight value, and evaluate the risk model for false invoicing through the risk scanning function in combination with historical risk events to determine the risk level.

[0040] Exemplarily, a risk model for false invoicing is created based on the above indicators, which includes information such as indicators, thresholds, warning values, likelihood of risk occurrence, degree of risk impact, weights, etc. By setting the indicator categories (positive indicators, inverse indicators, moderate indicators), the warning value is calculated according to the indicator category and the threshold. The warning value is the upper limit of the threshold when the indicator category is an inverse indicator, otherwise it is the lower limit of the threshold. The risk model for false invoicing is scanned and evaluated through the risk scanning function in combination with the following calculation relationships to obtain the risk scanning result. The deviation rate is calculated based on the indicator and the warning value. The deviation rate = |(indicator value ÷ warning value - 1)| × 100%. The risk score is calculated based on the likelihood of risk occurrence and the degree of risk impact. The risk score = likelihood of risk occurrence × degree of risk impact. The risk weighted score is calculated based on the risk score and the weight. The risk weighted score = risk score × weight. The risk score rating is calculated based on the risk weighted score and the deviation rate. The risk score rating = risk weighted score × deviation rate. The total risk score rating of the model is calculated based on the score ratings of each risk indicator. The total risk score rating = sum of the risk indicator risk score ratings. The risk level is determined by comparing the total risk score rating with the threshold.

[0041] Exemplarily, the risk event description and countermeasures are combined into structured training data similar to instruction - input - output, and the data is organized into a format acceptable to the DeepSeek model, including input_ids, attention_mask, and labels. The pre - trained model is fine - tuned using the prepared training data so that it can generate risk response plans that meet business requirements.

[0042] Specifically, after step 103 is executed, the following steps are also executed: S1: Define evaluation indicators for evaluating the generation results, and continuously adjust the parameter information according to this indicator.

[0043] S2: Divide the countermeasures retrieved from the risk knowledge base and the context information related to historical risk events into multiple paragraphs to evaluate the relevance for each paragraph, and then use the relevant information as input to continue the retrieval to obtain the optimal countermeasures and the context information related to historical risk events; S3: Use the retrieved event description, risk level, optimal countermeasures, and context information as input to find relevant tax regulations in the tax knowledge base and generate a target risk report containing risk analysis and response plans.

[0044] Among them, the risk knowledge base records event descriptions, countermeasures, context information related to the events, and risk levels.

[0045] Step 104: Use text-to-video technology to generate video clips from the risk report, and splice and merge the video clips to generate a target video for tax risk display.

[0046] Specifically, referring to Figure 4 As shown, when executing Step 104, the following steps can be specifically executed: S1040: Use a pre-trained language model combined with a retrieval algorithm to extract key information from the generated risk report to form a summary, and based on the summary, use the pre-trained language model to generate the text content for video narration.

[0047] S1041: Perform sentiment analysis on the text content for video narration through a sentiment analysis model to determine the sentiment tendency of the text.

[0048] Exemplarily, the sentiment analysis model can adopt the sentiment analysis model of Hugging Face.

[0049] S1042: Based on the risk report, use a data visualization tool to generate a risk analysis chart for display in the video.

[0050] S1043: Convert the text content for video narration into speech narration, use text-to-video technology to integrate the text content for video narration, speech narration, and risk analysis chart to generate corresponding video clips, and then call the video synthesis function to splice and merge the video clips in sequence to generate a target video.

[0051] Exemplarily, TTS technology can be used to convert the text content for video narration into speech narration, and the video synthesis function can adopt the concatenate_videoclips function in the MoviePy library of Python.

[0052] Specifically, converting the text content for video narration into speech narration includes: converting the text content for video narration into a.txt file, providing an audio sample with a speech rate and intonation that match the style of tax risk content for a preset time period; using speech synthesis to generate speech from the text in the.txt file, and using the audio sample for voice cloning to obtain speech narration.

[0053] It should be noted that the preset time period can be set according to specific application scenarios, such as 10 - 30s, and there is no limitation here.

[0054] It should be understood that the magnitudes of the sequence numbers of the steps in the above embodiments do not mean the order of execution. The order of execution of each process should be determined by its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present invention.

[0055] In the above embodiments, a transaction relationship knowledge graph is constructed by extracting multi-modal data, and a machine learning model is combined to quantify risk indicators and generate a risk report. Finally, through the text-to-video technology, the analysis results are converted into a video containing voice, charts, and dynamic scenarios, realizing the visual intelligent display of tax risks and improving the accuracy and response efficiency of risk identification.

[0056] As Figure 5 shown, the following is an embodiment of a tax risk display system provided by the embodiments of the present disclosure, which belongs to the same inventive concept as the tax risk display methods of the above embodiments. Details not described in detail in the embodiments of the tax risk display system can refer to the embodiments of the above tax risk display methods.

[0057] An extraction unit, configured to extract multi-modal data included in the goods flow, contract flow, funds flow, and invoice flow in the business process and construct an association relationship between the multi-modal data; A determination unit, configured to construct a transaction relationship knowledge graph based on the association relationship and associate the multi-modal data with entities in the transaction relationship knowledge graph through a mapping rule library; A detection unit, configured to perform consistency detection on the goods flow, contract flow, funds flow, and invoice flow based on the transaction relationship knowledge graph to obtain a detection result; A generation unit, configured to determine a risk level based on the detection result in combination with historical risk events in a risk knowledge base, match countermeasures from the risk knowledge base, and associate tax regulations from a tax knowledge base to generate a risk report; A display unit, configured to generate video segments from the risk report by using text-to-video technology and splice and combine the video segments to generate a target video for tax risk display.

[0058] Figure 6 It is a schematic hardware structure diagram of an electronic device for implementing various embodiments of the present invention.

[0059] The method for displaying tax risks provided by the embodiments of the present application can be applied to an electronic device. Those skilled in the art can understand that the structure of the electronic device involved in the embodiments of the present invention does not constitute a limitation on the electronic device. The electronic device may include more or fewer components than shown in the figure, or combine certain components, or have different component arrangements. In the embodiments of the present invention, the electronic device includes, but is not limited to, a laptop computer, a desktop computer, a workbench, a personal digital assistant, a server, a blade server, a mainframe computer, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as, a personal digital processor, a cellular phone, a smart phone, a wearable device, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the embodiments of the present application described herein and / or claimed.

[0060] The electronic device may include a processor, an external memory interface, an internal memory, a universal serial bus (USB) interface, a charging management module, a power management module, a battery, a wireless communication module, an audio module, a speaker, a microphone, a sensor module, a key, a camera, a display screen, and a SIM card interface, etc.

[0061] It can be understood that the structure schematically shown in the embodiments of the present application does not constitute a specific limitation on the electronic device. In other embodiments of the present application, the electronic device may include more or fewer components than shown in the figure, or combine certain components, or split certain components, or have different component arrangements. The components shown in the figure can be implemented in hardware, software, or a combination of software and hardware.

[0062] The processor may include one or more processing units. For example, the processor may include a central processing unit (CPU), etc., an application processor (AP), a modem processor, a graphics processing unit (GPU), an image signal processor (ISP), a controller, a memory, a video codec, a digital signal processor (DSP), a baseband processor, and / or a neural-network processing unit (NPU), etc. Among them, different processing units may be independent devices or integrated in one or more processors.

[0063] Among them, the processor can be the nerve center and command center of the electronic device. The controller can generate operation control signals according to the instruction operation code and timing signal to complete the control of fetching and executing instructions.

[0064] A memory can also be set in the processor for storing instructions and data. In some embodiments, the memory in the processor is a cache memory. This memory can save the instructions or data that the processor has just used or recycled. If the processor needs to use the instruction or data again, it can directly call it from this memory. This avoids repeated accesses, reduces the waiting time of the processor, and thus improves the system efficiency.

[0065] The external memory interface can be used to connect to an external memory card, such as a MicroSD card, to expand the storage capacity of the electronic device. The external memory card communicates with the processor through the external memory interface to achieve the data storage function. For example, files such as music and videos are saved in the external memory card.

[0066] The internal memory can be used to store computer-executable program code, and the computer-executable program code includes instructions. The processor executes various functional applications and data processing of the electronic device by running the instructions stored in the internal memory. The internal memory can include a program storage area and a data storage area. The internal memory can include a high-speed random access memory and can also include a non-volatile memory, such as at least one disk storage device, a flash memory device, a universal flash storage (UFS), etc.

[0067] The wireless communication function of the electronic device can be implemented through an antenna, a wireless communication module, a modulation and demodulation processor, a baseband processor, etc.

[0068] The wireless communication module can provide wireless communication solutions applied to the electronic device, including wireless local area networks (WLANs) (such as wireless fidelity (Wi-Fi) networks), Bluetooth (BT), global navigation satellite systems (GNSSs), frequency modulation (FM), near field communication (NFC), infrared technology (IR), etc.

[0069] The electronic device can implement audio functions, etc. through an audio module, a speaker, a receiver, a microphone, a headphone interface, an application processor, etc.

[0070] An electronic device can implement a shooting function through an ISP, a camera, a video codec, a GPU, a display screen, an application processor, etc.

[0071] An electronic device can implement a display function through a GPU, a display screen, an application processor, etc.

[0072] The GPU is a microprocessor for image processing, connecting the display screen and the application processor. The GPU is used to perform mathematical and geometric calculations for graphics rendering. The processor may include one or more GPUs, which execute program instructions to generate or change display information.

[0073] The display screen is used to display images, videos, etc. The display screen includes a display panel.

[0074] In the storage medium provided by this application, there is a program product that can implement the display method of tax risks.

[0075] The display method of tax risks includes: extracting multimodal data included in the goods flow, contract flow, capital flow, and invoice flow in the business process and constructing the association relationship between the multimodal data; based on the association relationship, constructing a transaction relationship knowledge graph and associating the multimodal data with the entities in the transaction relationship knowledge graph through a mapping rule library; based on the transaction relationship knowledge graph, performing consistency detection on the goods flow, contract flow, capital flow, and invoice flow to obtain a detection result; determining a risk level based on the detection result in combination with historical risk events in the risk knowledge base, matching countermeasures from the risk knowledge base, and associating tax regulations from the tax knowledge base to generate a risk report; using text-to-video technology to generate video segments from the risk report and splicing and combining the video segments to generate a target video for tax risk display.

[0076] In some possible implementation manners, the subject name of the present disclosure, the display method and system of tax risks, can be implemented in the form of a program product, which includes program code. When the program product runs on a terminal device, the program code is used to cause the terminal device to execute the steps according to various exemplary embodiments of the present disclosure described in the "Exemplary Method" section above of this specification.

[0077] The storage medium of the present disclosure may adopt any combination of one or more readable media. The readable media may be a readable signal medium or a readable storage medium. The readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples (a non-exhaustive list) of the readable storage medium include: an electrical connection having one or more wires, a portable 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.

[0078] The foregoing description of the disclosed embodiments enables those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for displaying tax risks, characterized in that: include: Extract multimodal data contained in the goods flow, contract flow, capital flow and invoice flow in the business process and build the association relationship between multimodal data; Based on the association relationship, a transaction relationship knowledge graph is constructed, and the multimodal data is associated with entities in the transaction relationship knowledge graph through a mapping rule base; Based on the transaction relationship knowledge graph, consistency detection is performed on the goods flow, contract flow, capital flow and invoice flow to obtain detection results; Determine the risk level based on the detection results combined with historical risk events in the risk knowledge base, match response measures from the risk knowledge base, and associate tax regulations from the tax knowledge base to generate a risk report; Vincent video technology is used to generate video clips from risk reports, and the video clips are spliced ​​and merged to generate target videos for tax risk presentation.

2. The method for displaying tax risks according to claim 1, characterized in that: Based on the association relationship, a transaction relationship knowledge graph is constructed, and the multimodal data is associated with entities in the transaction relationship knowledge graph through a mapping rule base, including: For cargo flow, identify the logistics documents and extract key information such as consignor, consignee, cargo description and waybill number; For the contract flow, natural language processing technology is used to analyze the contract text and extract key terms, contract subjects, goods description, and amount information; Regarding the flow of funds, we can identify the information on bank statements and extract the transaction amount, transaction date, and key information of the counterparty. For the invoice flow, identify the invoice information and extract key information such as invoice number, amount, goods description, and invoicing date; The extracted multimodal data of shipper, consignee, cargo description, waybill number key information, key terms, contract subject, cargo description, amount information, transaction amount, transaction date, counterparty key information, invoice number, amount, cargo description, and invoicing date key information are unified into text format, a mapping rule library between multimodal data and transaction relationship knowledge graph entities is created, and the extracted multimodal information is associated with the entities in the knowledge graph.

3. The method for displaying tax risks according to claim 1, characterized in that: Determine the risk level based on the detection results combined with historical risk events in the risk knowledge base, including: Based on the transaction relationship knowledge graph, consistency detection is performed on the goods flow, contract flow, capital flow and invoice flow to determine the business nodes with inconsistent information in the goods flow, contract flow, capital flow and invoice flow; Define abnormal indicators related to taxpayers’ risk of false invoicing; Collecting data related to tax risks in business nodes, and dividing the collected data related to tax risks into a training set and a test set, wherein the test set is further divided into multiple samples corresponding to different risk levels; Set the number of decision trees and maximum depth parameters in the initial random forest model, and use the training set for training. Use the trained random forest model to predict the test set and obtain the predicted category, predicted probability, risk impact degree, and indicator weight value. A false invoice risk model is constructed based on prediction category, prediction probability, risk impact degree and indicator weight value. The false invoice risk model is scanned and evaluated through the risk scanning function combined with historical risk events to determine the risk level.

4. The method for displaying tax risks according to claim 3, characterized in that: The abnormal indicators include at least abnormal changes in taxpayers' tax burden, abnormal changes in the usage of special invoices, abnormal customer dispersion of sales invoices, abnormally high cancellation rate of special invoices, abnormal difference between the end-of-period inventory and paid-in capital, and abnormal difference between the end-of-period inventory and the current period's accumulated income.

5. The method for displaying tax risks according to claim 1, characterized in that: After determining the risk level based on the detection result combined with historical risk events in the risk knowledge base, matching countermeasures from the risk knowledge base, and associating tax regulations from the tax knowledge base to generate a risk report, the tax risk display method further includes: Define the evaluation index for evaluating the generated results and continuously adjust the parameter information according to the index; The countermeasures and contextual information related to historical risk events retrieved from the risk knowledge base are divided into multiple sections to evaluate the relevance of each section, and then the relevant information is used as input to continue the search to obtain the optimal countermeasures and contextual information related to historical risk events; Taking the retrieved event description, risk level, optimal response measures and context information as input, finding relevant tax regulations in the tax knowledge base, and generating a target risk report including risk analysis and response plans; The risk knowledge base records event descriptions, response measures, contextual information related to the event, and risk levels.

6. The method for displaying tax risks according to claim 1, characterized in that: The risk report is generated into video clips using Vincent video technology, and the video clips are spliced ​​and merged to generate a target video for tax risk presentation, including: Using a pre-trained language model in combination with a retrieval algorithm, extract key information from the generated risk report to form a summary, and based on the summary, use the pre-trained language model to generate text content narrated in the video; Use sentiment analysis models to analyze the text content of the video to determine the emotional tendency of the text; Based on the risk report, use data visualization tools to generate risk analysis charts for presentation in the video; Convert the text content of the video narration into voice narration, use Vincent video technology to integrate the video narration text content, voice narration, and risk analysis charts to generate corresponding video clips, and then call the video synthesis function to splice and merge the video clips in sequence to generate the target video.

7. The method for displaying tax risks according to claim 6, characterized in that: Convert the text content of the video narration into voice narration, including: Convert the text content of the video into a .txt file, providing an audio sample of a preset time period with a speed and tone that matches the style of the tax risk content; Use speech synthesis to generate speech from the text in the .txt file, and use audio samples for voice cloning to obtain voice narration.

8. A tax risk display system, characterized in that: include: An extraction unit, used to extract multimodal data contained in the goods flow, contract flow, capital flow and invoice flow in the business process and to build associations between the multimodal data; A determination unit, configured to construct a transaction relationship knowledge graph based on the association relationship, and associate the multimodal data with entities in the transaction relationship knowledge graph through a mapping rule base; A detection unit, configured to perform consistency detection on the goods flow, contract flow, capital flow and invoice flow based on the transaction relationship knowledge graph to obtain a detection result; A generating unit, configured to determine the risk level based on the detection result in combination with historical risk events in the risk knowledge base, match the response measures from the risk knowledge base, and associate the tax regulations from the tax knowledge base to generate a risk report; The display unit is used to generate video clips from the risk report using Vincent video technology, and to splice and merge the video clips to generate a target video for tax risk display.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the steps of the tax risk display method according to any one of claims 1 to 7 are implemented.

10. A storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the tax risk display method according to any one of claims 1 to 7 are implemented.

Images