Account identification method and device, storage medium and electronic equipment

Abstract

The invention discloses an account identification method and device, a storage medium and electronic equipment. The method comprises the steps that a target prompt word is set according to a target business scene, and the target prompt word is used for indicating a target large language model to process business data based on the target business scene; obtaining business data corresponding to the target account, matching a target image displayed by a target application and a business process template through skipping to the target application according to the business data, and performing character recognition determination on the target image under the conditions that the matching is successful and the target application always runs in a system foreground; and inputting the target cue word and the business data into the target large language model to obtain a target identification result, the target identification result comprising a risk tag, and the risk tag being used for indicating whether the target account is a risk account. According to the method and the device, the technical problem of relatively low account identification accuracy caused by one-sided business data processing modes corresponding to the accounts is solved.

Description

Technical Field

[0001] The present application relates to the field of computers, and more specifically, to an account identification method and device, a storage medium, and an electronic device. Background Art

[0002] The existing technology often uses a single or limited business data processing method to assess account risk. For example, it only relies on traditional business records or simple account status information, which often fails to fully capture the multi-dimensional information that affects account risk, especially those non-direct financial indicators, such as users' living habits, professional background, marital status, etc. In general, due to the one-sided business data processing method corresponding to the account, the account identification accuracy is low and cannot accurately reflect the user's true account status.

[0003] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention

[0004] The embodiments of the present application provide an account identification method and device, a storage medium, and an electronic device to at least solve the technical problem of low account identification accuracy due to one-sided processing of business data corresponding to an account.

[0005] According to one aspect of an embodiment of the present application, an account identification method is provided, comprising: setting a target prompt word according to a target business scenario, wherein the target prompt word is used to instruct a target large language model to process business data based on the target business scenario; obtaining the business data corresponding to the target account, wherein the business data is matched with a target image and a business process template displayed by the target application by jumping to the target application, and when the match is successful and the target application is always running in the foreground of the system, performing text recognition and determination on the target image; inputting the target prompt word and the business data into the target large language model to obtain a target identification result, wherein the target identification result includes a risk label, and the risk label is used to indicate whether the target account is a risky account.

[0006] According to another aspect of an embodiment of the present application, an account identification device is also provided, including: a setting module, used to set a target prompt word according to a target business scenario, wherein the target prompt word is used to instruct a target large language model to process business data based on the target business scenario; an acquisition module, used to obtain the business data corresponding to the target account, wherein the business data is matched with a target image and a business process template displayed by the target application by jumping to the target application, and when the match is successful and the target application is always running in the foreground of the system, the target image is subjected to text recognition and determination; an identification module, used to input the target prompt word and the business data into the target large language model to obtain a target identification result, wherein the target identification result includes a risk label, and the risk label is used to indicate whether the target account is a risky account.

[0007] Optionally, the device is used to input the target prompt word and the business data into the target large language model to obtain a target recognition result in the following manner: input the target prompt word and the business data into the target large language model to determine the risk type, risk level and the reason why the target account is at risk; determine the risk label based on the risk type and risk level; and generate the target recognition result based on the risk label.

[0008] Optionally, the device is used to input the target prompt word and the business data into the target large language model to obtain a target recognition result in the following manner: input the target prompt word and the business data into the target large language model to determine the risk type, risk level and the reason why the target account is at risk; determine the risk label based on the risk type and risk level, and determine the account risk information based on the risk label and the reason why the target account is at risk; generate the target recognition result based on the risk label and the account risk information.

[0009] Optionally, the device is also used to: obtain a preset target business field, wherein the target business field represents a field that can affect the risk label; extract target business data from the business data according to the target business field, wherein the target business data represents the business data corresponding to the target business field; concatenate the target prompt word and the business data to obtain target input data, and input the target input data into the target large language model to obtain the target recognition result.

[0010] Optionally, the device is also used to: embed the target prompt word in the initial sample data to obtain target sample data, wherein the target sample data represents the initial sample data associated with the target business scenario; and use the target sample data to train an initial large language model to obtain the target large language model.

[0011] Optionally, the device is used to obtain business data corresponding to the target account in the following manner: based on the target image, query whether the target template exists in the business process template and whether the target application is running in the system foreground, wherein the image features of the target template and the target image satisfy a preset matching condition; when the target template exists and the target application is always running in the system foreground, perform a text recognition operation on the target image to obtain the business data, and store the business data and the target image in the target database.

[0012] Optionally, the device is also used for at least one of the following: generating a second prompt message when the target template does not exist, wherein the second prompt message is used to indicate that the business data needs to be reacquired; and generating a third prompt message when the target application switches from running in the system foreground to the system background, wherein the third prompt message is used to indicate that the business data needs to be reacquired.

[0013] According to another aspect of the embodiments of the present application, a computer-readable storage medium is provided, in which a computer program is stored, wherein the computer program is configured to execute the above-mentioned account identification method when running.

[0014] According to another aspect of the embodiments of the present application, a computer program product or a computer program is provided, the computer program product or the computer program including computer instructions, the computer instructions being stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device performs the above account identification method.

[0015] According to another aspect of the embodiments of the present application, there is further provided an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor is configured to execute the account identification method through the computer program.

[0016] In an embodiment of the present application, a target prompt word is set according to a target business scenario, wherein the target prompt word is used to instruct the target large language model to process business data based on the target business scenario; the business data corresponding to the target account is obtained, wherein the business data is matched with the target image and the business process template displayed by the target application by jumping to the target application, and when the match is successful and the target application is always running in the foreground of the system, the target image is recognized and determined by text; the target prompt word and the business data are input into the target large language model to obtain a target recognition result, wherein the target recognition result includes a risk label, and the risk label is used to indicate whether the target account is a risky account. Through the comprehensive analysis of multi-dimensional business data and the intelligent judgment of the large language model, the purpose of improving the accuracy of account risk identification is achieved, thereby achieving the technical effect of accurate account identification, and then solving the technical problem of low account identification accuracy due to the one-sided processing method of the business data corresponding to the account. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0018] Figure 1 is a schematic diagram of an application environment of an optional account identification method according to an embodiment of the present application;

[0019] Figure 2 is a flowchart of an optional account identification method according to an embodiment of the present application;

[0020] Figure 3 is a schematic diagram of an optional account identification method according to an embodiment of the present application;

[0021] Figure 4 is a schematic diagram of the structure of an optional account identification device according to an embodiment of the present application;

[0022] Figure 5 is a schematic diagram of the structure of an optional account identification product according to an embodiment of the present application;

[0023] Figure 6 It is a schematic diagram of the structure of an optional electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0024] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.

[0025] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0026] The present application is described below in conjunction with embodiments:

[0027] According to one aspect of an embodiment of the present application, an account identification method is provided. Optionally, in this embodiment, the account identification method can be applied to Figure 1 In the hardware environment composed of the server 101 and the terminal device 103 shown in FIG. Figure 1As shown, the server 101 is connected to the terminal device 103 via a network, and can be used to provide services for the terminal device or an application 107 installed on the terminal device. The application can be a video application, an instant messaging application, a browser application, an educational application, a game application, etc. A database 105 may be set up on the server or independently of the server to provide data storage services for the server 101, for example, a game data storage server. The above-mentioned network may include, but is not limited to, a wired network and a wireless network, wherein the wired network includes, a local area network, a metropolitan area network and a wide area network; the wireless network includes, Bluetooth, WIFI and other networks that implement wireless communication; the terminal device 103 may be a terminal configured with an application, and may include, but is not limited to, at least one of the following: a mobile phone (such as an Android phone, an iOS phone, etc.), a laptop computer, a tablet computer, a PDA, a MID (Mobile Internet Devices), a PAD, a desktop computer, a smart TV, an intelligent voice interaction device, a smart home appliance, a vehicle-mounted terminal, an aircraft, a virtual reality (VR) terminal, an augmented reality (AR) terminal, a mixed reality (MR) terminal and other computer devices; the above-mentioned server may be a single server, a server cluster consisting of multiple servers, or a cloud server.

[0028] Combination Figure 1 As shown, the above account identification method can be executed by an electronic device, which can be a terminal device or a server. The above account identification method can be implemented by the terminal device or the server respectively, or by the terminal device and the server together.

[0029] The above is only an example and is not specifically limited in this embodiment.

[0030] Optionally, as an optional implementation, as Figure 2 As shown, the above account identification method includes:

[0031] S202, setting a target prompt word according to a target business scenario, wherein the target prompt word is used to instruct a target large language model to process business data based on the target business scenario;

[0032] Optionally, in an embodiment of the present application, the above-mentioned target large language model may include but is not limited to a pre-trained model based on the Transformer architecture, a BERT model, a GPT series model, etc.; the above-mentioned target prompt words refer to a set of keywords or phrases that can guide the target large language model to focus on specific risk factors, including but not limited to "overdue records", "financial stability", "credit scores", etc., which are closely related to specific business scenarios and are used to clarify the direction and focus of model analysis; the above-mentioned business data refers to various types of information related to the target account in the target business scenario, including but not limited to financial statements, transaction records, credit reports, user behavior logs, etc.

[0033] It should be noted that the setting of the above target prompt words can be flexibly adjusted according to different business scenarios. For example, in personal credit analysis, the target prompt words may involve overdue records, marital status, income status, etc.; when reviewing loan applications, the target prompt words may focus on loan history, repayment ability, guarantee status, etc. This application does not limit this, and the selection and application of target prompt words should fully consider the particularity of the scenario and the relevance of the data.

[0034] Exemplarily, first, target prompt words are determined according to the target business scenario. These prompt words will serve as a guide for the model to process business data. Then, multi-dimensional business data of the target account in the business scenario is obtained, including but not limited to financial information, behavioral data, etc.

[0035] S204, obtaining business data corresponding to the target account, wherein the business data is redirected to the target application, and the target image displayed by the target application is matched with the business process template. If the match is successful and the target application is always running in the system foreground, the target image is recognized and determined by text;

[0036] Optionally, in an embodiment of the present application, the above-mentioned target account refers to a specific user account whose risk needs to be assessed by a financial institution or related service provider, including but not limited to a personal credit account, a corporate financial service account, a credit card account, etc.

[0037] It should be noted that the design and matching process of the business process template can be flexible and diverse to adapt to the interface features and operation processes of different target applications. The template may be based on the key page elements, operation steps or business logic definitions of the target application. This application does not limit this. The key is that the template must accurately reflect the requirements of the target business scenario to ensure the accuracy and effectiveness of the matching process.

[0038] Exemplarily, in the technical solution of the embodiment of the present application, first, the user account needs to enter the target application, such as a personal credit inquiry platform, through a preset jump mechanism; then, when the target application is running in the system foreground, the system matches the key page image displayed by the target application with the business process template to ensure that the user operation complies with the expected process; once the match is successful, the system will perform text recognition on the target image and extract business data related to risk assessment; finally, combined with the target prompt words set for the target business scenario, the business data is input into the target large language model, and through model analysis and processing, a risk assessment result closely related to the business scenario is obtained, effectively improving the accuracy and comprehensiveness of account risk identification.

[0039] S206, input the target prompt word and business data into the target large language model to obtain a target recognition result, wherein the target recognition result includes a risk label, and the risk label is used to indicate whether the target account is a risky account.

[0040] Optionally, in an embodiment of the present application, the above-mentioned target prompt words refer to keywords or phrases used to guide the analysis direction of the target large language model, including but not limited to guidance information in specific business scenarios such as "risk assessment", "credit check", and "fraud detection".

[0041] It should be noted that the selection of the target large language model and the setting of the target prompt words can be adjusted according to different business scenarios and needs. For example, in a credit approval scenario, the model may be required to have the ability to process complex financial documents, while in insurance claims analysis, more emphasis may be placed on the model's understanding of medical terminology and document structure. The types and sources of business data are also diverse, and can come from internal databases, third-party services, or information directly uploaded by users. This application does not limit this. The key is that the data must be associated with the target business scenario and can be effectively processed by the large language model through the target prompt words.

[0042] For example, first, the corresponding target prompt words are defined according to the target business scenario to guide the analysis focus of the target large language model; then, the business data related to the scenario is collected, which may include the user's behavior records in financial transactions, credit reports, financial statements, etc. Next, the target prompt words and business data are combined and input into the target large language model. The model conducts in-depth analysis and understanding of the data according to the guidance of the prompt words, and outputs the target recognition results related to the business scenario. In the output results, the risk label is a key component used to clearly indicate whether the target account involves risks, so that financial institutions or service providers can make quick and accurate business decisions.

[0043] In an exemplary embodiment, Figure 3is a schematic diagram of an optional method for processing business data according to an embodiment of the present application. The implementation process of the present application is as follows Figure 3 shown.

[0044] Through the embodiments of the present application, a target prompt word is set according to the target business scenario, wherein the target prompt word is used to instruct the target large language model to process business data based on the target business scenario; the business data corresponding to the target account is obtained, wherein the business data is matched with the target image and the business process template displayed by the target application by jumping to the target application, and when the match is successful and the target application is always running in the foreground of the system, the target image is recognized and determined by text; the target prompt word and the business data are input into the target large language model to obtain a target recognition result, wherein the target recognition result includes a risk label, and the risk label is used to indicate whether the target account is a risk account. Through the comprehensive analysis of multi-dimensional business data and the intelligent judgment of the large language model, the purpose of improving the accuracy of account risk identification is achieved, thereby achieving the technical effect of accurate account identification, and then solving the technical problem of low account identification accuracy due to the one-sided processing method of the business data corresponding to the account.

[0045] As an optional solution, the target prompt word and the business data are input into the target big language model to obtain a target recognition result, including: inputting the target prompt word and the business data into the target big language model to determine the risk type, risk level and the reason why the target account is at risk; determining the risk label based on the risk type and risk level; and generating the target recognition result according to the risk label.

[0046] It should be noted that the determination of the above-mentioned risk types and risk levels, as well as the analysis of the reasons that lead to the risk of the target account, may be customized according to different business scenarios and model capabilities. For example, in the credit field, risk types may include credit risk, fraud risk, etc., and risk levels can be a series of quantitative indicators, such as low risk, medium risk, high risk, etc. This application does not limit this.

[0047] In an exemplary embodiment, taking the application scenario of personal loan review as an example, when a financial institution receives a loan application, it first sets prompt words related to the loan review according to the target business scenario, such as "credit history", "income stability", "debt ratio", etc. These prompt words will serve as a guide for model analysis; then, the system automatically obtains corresponding records from the business data of the target account, including credit reports, bank statements, tax information, etc.; then, the above prompt words and business data are input into the target large language model, and the model determines the risk type (such as poor credit history) and risk level (such as high risk) of the target account through in-depth analysis, and elaborates on the specific reasons for the risk (such as multiple overdue repayment records); based on the analysis of the above risk type and risk level, the system automatically determines the risk label (such as "high credit risk"), and finally generates a target recognition result based on the risk label, informing the loan auditor whether to approve the loan application and the information points that need to be further verified, such as the need for additional income proof or a detailed explanation of the reasons for overdue repayment.

[0048] Through the embodiments of the present application, the method of combining target prompt words and business data input into the target large language model is adopted to achieve the technical effect of intelligent risk assessment, achieve the purpose of quickly and accurately identifying the risk status of the target account, and assist financial institutions in making efficient decisions, thereby improving customer experience and business processing efficiency while ensuring risk control.

[0049] As an optional solution, the above-mentioned target prompt words and the above-mentioned business data are input into the above-mentioned target big language model to obtain the target recognition result, including: inputting the above-mentioned target prompt words and the above-mentioned business data into the above-mentioned target big language model to determine the risk type, risk level and the reason why the above-mentioned target account is at risk; determining the above-mentioned risk label based on the above-mentioned risk type and risk level, and determining the account risk information based on the above-mentioned risk label and the above-mentioned reason why the above-mentioned target account is at risk; generating the above-mentioned target recognition result according to the above-mentioned risk label and the above-mentioned account risk information.

[0050] It should be noted that the risk cause analysis involves details such as overdue records and abnormal trading behaviors, which are not limited in this application.

[0051] In an exemplary embodiment, taking the application scenario of credit card application review as an example, first, the system sets target prompt words based on the business needs of credit card application review, such as "past credit status", "repayment ability", etc.; then, collect relevant business data from the target account, such as personal credit report, income certificate, bank flow record, etc.; then, input the target prompt words and business data into the target large language model, the model analyzes the data and determines the risk type (such as credit risk), risk level (such as medium risk), and identifies the specific reasons that cause the account to be at risk (such as a recent record of continuous overdue repayments); based on the risk type and risk level, the system determines the risk label (such as "medium credit risk"), and generates detailed account risk information in combination with the risk reasons; finally, based on the risk label and account risk information, generates a target recognition result to provide a reference for the credit card approver, such as suggesting raising the application threshold or requiring additional income confirmation.

[0052] Through the embodiments of the present application, a method of inputting target prompt words and business data into a target large language model is adopted to achieve an intelligent and customized risk assessment technical effect, achieve the purpose of quickly and accurately identifying and analyzing the risk status of target accounts in different business scenarios, effectively support the decision-making process of financial institutions, and ensure the security and compliance of customer information processing.

[0053] As an optional solution, the method further includes: obtaining a preset target business field, wherein the target business field represents a field that can affect the risk label; extracting target business data from the business data according to the target business field, wherein the target business data represents the business data corresponding to the target business field; concatenating the target prompt word and the business data to obtain target input data, and inputting the target input data into the target large language model to obtain the target recognition result.

[0054] Optionally, in an embodiment of the present application, the above-mentioned target business field refers to key data points that can affect the risk assessment results, including but not limited to "historical overdue times", "income stability", "debt ratio", etc., and the above-mentioned target business data represents specific information filtered out from the business data according to the target business field. For example, if the target business field is "historical overdue times", then the target business data is the information related to the overdue times in the business data, and the above-mentioned target prompt words are used to guide the target large language model to pay attention to these specific fields when analyzing business data, ensuring that the model's analysis is highly relevant to risk assessment.

[0055] It should be noted that the selection of the above target business fields and the setting of the target prompt words can be adjusted according to the specific business scenario and risk assessment requirements. For example, when assessing investment risk, the target business fields may include "asset allocation" and "market volatility", while when assessing loan risk, more attention may be paid to "income level" and "credit history". This application does not limit this.

[0056] In an exemplary embodiment, taking the application scenario of personal consumer loan approval as an example, the system first determines the target business fields, such as "credit score", "proof of income" and "historical loan records". Subsequently, information related to these fields is extracted from the business data, such as the credit score in the credit report, bank statements as proof of income, and past loan records. Next, the keywords related to risk assessment (i.e., target prompt words) are spliced ​​with the extracted target business data to form a model input, such as splicing "credit risk analysis" with specific credit score data. Finally, this input data is sent to the target large language model, and the model outputs the target recognition results based on the input prompt words and business data, which may include risk labels (such as high risk), risk types (such as credit risk), risk levels (such as extremely high levels) and specific risk cause analysis, such as "the applicant has a frequent record of overdue repayments recently."

[0057] Through the embodiments of the present application, a strategy of combining the target business field with the target prompt word and inputting the target large language model is adopted to achieve accurate and efficient risk assessment technical effects, achieve the purpose of quickly identifying and quantifying the risk status of the target account in complex business scenarios, effectively support the decision-making process of financial institutions, and ensure the accuracy and reliability of risk assessment.

[0058] As an optional solution, the above method also includes: embedding the above target prompt words in the initial sample data to obtain target sample data, wherein the above target sample data represents the above initial sample data associated with the above target business scenario; using the above target sample data to train the initial large language model to obtain the above target large language model.

[0059] Optionally, in an embodiment of the present application, the above-mentioned initial sample data represents the original data set without specific processing, which is used to train the model, including but not limited to historical transaction records, credit report samples, user behavior data, etc., which are the basis for building the target large language model.

[0060] It should be noted that the process of generating target sample data and using it to train the target large language model can be adjusted for different business scenarios and target prompt words. For example, in the field of financial risk management, the target prompt words may be more related to the identification of financial risks, while in the field of social media analysis, they may focus on sentiment analysis and content understanding. This application does not limit this, the key is that the embedding of the target prompt words needs to be closely integrated with the business scenario to improve the pertinence and effectiveness of model training.

[0061] In an exemplary embodiment, taking the application scenario of financial credit risk assessment as an example, first, the system analyzes business needs and determines target prompt words, such as "credit score", "repayment record", "financial status", etc. Next, collect initial sample data from historical credit approval cases, including applicants' credit reports, financial statements, repayment records and other information. Then, embed the target prompt words into these initial sample data to generate target sample data to enhance the model's understanding of risk assessment-related factors. Finally, use the target sample data to train the initial large language model, and by adjusting the model parameters and optimizing the training process, obtain a target large language model that can more accurately identify risk signals in credit business.

[0062] Through the embodiments of the present application, a technology of embedding target prompt words in the initial sample data and performing customized training on the large language model is adopted, thereby achieving a precise match between the model analysis capability and business needs, and achieving the purpose of improving the accuracy and efficiency of risk assessment.

[0063] As an optional solution, the above-mentioned obtaining of business data corresponding to the target account includes: based on the above-mentioned target image, querying whether there is a target template in the above-mentioned business process template and whether the above-mentioned target application is running in the system foreground, wherein the image features of the above-mentioned target template and the above-mentioned target image satisfy a preset matching condition; when the above-mentioned target template exists and the above-mentioned target application is always running in the system foreground, performing a text recognition operation on the above-mentioned target image to obtain the above-mentioned business data, and storing the above-mentioned business data and the above-mentioned target image in the target database.

[0064] Optionally, in an embodiment of the present application, the target image refers to an image frame captured from the user screen sharing, which is used for subsequent image feature analysis and business data extraction; the business process template refers to a predefined set of template images used to match and identify business operation scenarios in the target image, including but not limited to a credit report query interface, a financial statement submission interface, etc. The template contains key image feature information; the target template refers to a specific template that successfully matches the target image feature and meets the preset matching conditions; the target application refers to the application that the user is using during the user screen sharing process, which is used to determine whether the application is in the system foreground running state. The target database is a database used to store business data and original images extracted from the target image for subsequent analysis.

[0065] It should be noted that the generation of business process templates and the setting of matching conditions can be diversified. For example, the template can be automatically generated through machine learning technology, or designed manually; the matching condition can be an exact match based on image features, or a fuzzy match or a probability-based match, which is not limited in this application. The running status of the target application can also be determined in a variety of ways, such as using the operating system API to detect the active window of the application, or determining whether the application is in the screen area where the user is currently interacting through screen area analysis.

[0066] For example, the target image is first captured from the screen sharing stream, and then the image feature matching is performed using the business process template to determine whether there is a target template that matches the target image feature; at the same time, it is detected whether the target application is always running in the system foreground. If the target template exists and the target application is indeed running in the foreground, the system will perform text recognition operations, extract business data from the target image, and store this data together with the target image itself in the target database to provide a basis for subsequent risk assessment and business decisions.

[0067] In an exemplary embodiment, taking the personal credit report upload function in the bank's credit approval process as an example, the user uses the screen sharing function to display the credit report page he is viewing. The system first captures the image frames of these pages as target images, and then uses the predefined credit report query interface template to match the image features to confirm that the user is viewing the credit report page. At the same time, the system detects whether the bank's credit service application is always running in the system foreground of the user's device to prevent the user from switching to other applications to tamper with the data. If the match is successful and the application is running in the foreground, the system will perform text recognition on the target image, extract key business data, such as overdue records, credit scores, etc., and then store these data and images in the target database for subsequent credit risk assessment.

[0068] Through the embodiments of the present application, feature matching of real-time screen sharing images and foreground running status detection of business applications are adopted, combined with text recognition technology to extract business data, thereby realizing the authenticity and integrity verification of user uploaded information, achieving the purpose of improving the accuracy of risk identification in credit business, and optimizing the user experience at the same time, avoiding unnecessary data queries.

[0069] As an optional solution, the above method also includes at least one of the following: when the above target template does not exist, generating a second prompt information, wherein the above second prompt information is used to indicate that the above business data needs to be re-acquired; when the above target application switches from running in the system foreground to the system background, generating a third prompt information, wherein the above third prompt information is used to indicate that the above business data needs to be re-acquired.

[0070] Optionally, in an embodiment of the present application, the second prompt information refers to information generated by the system when the target template does not match the target image, prompting the user to re-operate to obtain valid business data, including but not limited to prompt statements such as "Please return to the credit report viewing interface" and "Please make sure the current interface is the correct business operation interface"; the third prompt information refers to information generated by the system when the target application switches from the system foreground to the background, indicating that the user needs to re-enter the application and display the correct interface, such as "Do not switch applications, stay in the credit report viewing interface" or "Application switching is detected, please start again."

[0071] It should be noted that the second prompt information and the third prompt information generated by the system can be in a variety of forms, such as text prompts, voice broadcasts or pop-up warnings. The specific forms can be customized according to the actual application scenarios and user interaction interfaces to ensure that users can receive the re-operation instructions clearly and promptly. This application does not limit this.

[0072] For example, when the target image captured by the system cannot match any target template in the business process template, the user's current interface is considered to be inconsistent with the requirements for obtaining business data, and the system will generate a second prompt message to guide the user to return to the correct business operation interface to reacquire business data. Similarly, when it is detected that the target application is switched from the system foreground to the background, that is, the user may have switched to other applications, the system generates a third prompt message to remind the user to switch back to the target application and remain in the business operation interface to ensure the integrity and authenticity of the business data.

[0073] Through the embodiments of the present application, the matching situation of the screen sharing image and the business process template and the running status of the target application are monitored in real time to generate the second prompt information and the third prompt information, thereby realizing real-time guidance of user operations and accuracy control of data collection, achieving the purpose of improving the efficiency and quality of information collection, effectively avoiding user's human or non-human erroneous operations, ensuring the integrity and authenticity of business data, and improving the user experience and the smoothness of business processes through timely user guidance.

[0074] In an exemplary embodiment, the embodiment of the present application inputs the corresponding information of the output result of screen recognition (the above-mentioned business data) into the target large language model, and the target large language model analyzes and parses the information and gives judgments and reference bases.

[0075] Specifically, taking the financial industry as an example, in the current To C (consumer-oriented) services, it is usually necessary to introduce external customer information to enrich user portraits as an important basis for business handling decisions. The embodiment of the present application can enable users to actively share their desktops and personally conduct corresponding information queries and "upload" relevant personal information under the premise of user authorization. In the process of users sharing their screens, real-time prompts are given to users, and technical means are used to ensure that users do not cheat in the process. Specifically:

[0076] 1) The client (mobile phone or other terminal device) front desk jumps to the corresponding function interface of other APPs (the above-mentioned target application) through external links, and performs step-by-step operations according to the voice prompts or pop-up window guidance of the client background. If each step exceeds the time limit, an error will be prompted. In addition, other APPs record whether the user is always running in the system foreground during the operation, and the corresponding time start and end points.

[0077] 2) The client background starts sharing the screen from the jump action, pushes the video stream of the current client shared screen to the server, pushes the video stream of the shared screen to the server through the corresponding screen sharing function, and provides voice or pop-up prompts at each node.

[0078] 3) On the server side, the received video stream is captured periodically / frame by frame and sent to the interface judgment module. If it is detected that the shared screenshot matches the defined template, the current user is determined to be at the corresponding process node and a voice operation prompt is given to the user; if the shared screenshot does not match any defined template, a new screenshot is taken and a voice prompt is given.

[0079] Among them, for the interface judgment function, the template can be defined in advance by recording the existence and relative position of key elements in the interface of the business process. During the judgment process, the match between the screen interface and the template is judged based on the comparison of key elements. After the entire business process is executed, a pop-up window and voice prompt can be used to jump back to the original APP and perform the next step.

[0080] More specifically, on the server side, when the interface is judged as mismatched, it is considered that the user did not operate according to the regulations and may be suspected of cheating. The corresponding pictures are stored as feedback data, which can be used for subsequent model optimization. The user is prompted that the current operation has failed and needs to be re-recorded; after the user's business processing flow is completed, the records and time periods of whether other APPs of the original APP have always been running in the foreground during the business processing are queried. If after comparing the query results, it is judged that the user has switched to the background during the corresponding process, it is considered that the user did not operate according to the regulations and may be suspected of cheating, and the user is prompted that the current operation has failed and needs to be re-recorded; for the judgment of interface matching, the judgment starts from the key frame of the first page in the business process to reduce the configuration workload of the service and configuration templates; for the key screenshots that are identified for text recognition, they are stored after deduplication, that is, the interface matching at the image level is performed first, and then text recognition is performed after passing, which reduces the pressure on the text recognition service.

[0081] It should also be noted that in the embodiments of the present application, when achieving the goals of user screen sharing and information recognition, it is not necessary to use screen recording, because the present application adopts different technical paths, such as real-time screenshots, image recognition, application status monitoring, etc., to achieve the purpose of ensuring the authenticity and integrity of user operations and preventing cheating, without the need to record the video of the user's screen operations throughout the process, so as to reduce the amount of data transmission, reduce the demand for computing resources, improve processing efficiency, and avoid recording unnecessary video content.

[0082] Optionally, in the embodiment of the present application, the above-mentioned target large language model can be any large language model combined with the corresponding prompt word engineering, and the key fields to be focused on by the large model are defined according to the business needs of the business scenario. For example, the business scenario targeted by the embodiment of the present application is whether there is a risk in customer credit investigation, such as:

[0083] 1) "You are a credit report analysis specialist. Parse the input personal credit report file {customer credit report} and identify the key points that need attention and the possible risks. Find the credit card part of the personal credit report. If there is "details of overdue credit card accounts", identify the number of entries. Only output numbers. If there is no "details of overdue credit card accounts", output 0."

[0084] 2) "You are a credit report analysis specialist. Parse the input personal credit report file {customer credit report} and identify the key points that need attention and the possible risks. Find the loan part of the personal credit report. If there is "delayed account details", identify the number of entries. Output only numbers. If there is no "delayed account details", output 0."

[0085] 3) "You are a credit report analysis specialist. Parse the input personal credit report file {customer credit report} and identify the key points that need attention and the possible risks. Find the marital status in the personal credit report. Possible statuses include "unmarried", "single", "married", "widowed", and "divorced". Output and only output the identified marital status. Do not make any associations."

[0086] Furthermore, the module outputs of 1)-3) above are used as the input of the target large language model defined by the prompt word project:

[0087] "As a credit report analysis specialist, we analyze the input personal credit report file {customer credit report} to identify the key points that need attention and the possible risks. The final output is the interpretation report summary. Based on the identified customer {credit card overdue number}, {loan overdue number}, and referring to the customer's {marital status}, we output an analysis of the customer's overall credit situation."

[0088] In an exemplary embodiment, the identified customer information results (the above-mentioned business data) can be input into the target large language model with pre-programmed prompt words to further analyze the input information in a scenario-oriented manner and output the analysis results. The judgment result can be a specific label such as whether the customer is risky or a paragraph of analysis reference text.

[0089] For example, after entering the key fields of credit investigation, the large language model outputs the following analysis content:

[0090] “1. Specific label: Credit risk assessment of this customer: Risk exists;

[0091] 2. Analyze reference texts:

[0092] The customer credit report is interpreted as follows. Through an in-depth analysis of the customer Zhang San's personal credit report, we can draw the following key points and potential risks:

[0093] 2-1. Basic personal information: Customer Zhang San holds multiple identification documents, including his ID card, passport and military ID card, which indicates that he has extensive social activities. The customer is married, and this information may affect his family financial planning and credit assessment.

[0094] 2-2, Credit record overview: Zhang San has a complex credit background, including credit cards, loans and other businesses. In particular, the information on advances and asset disposals indicates that he has had serious financial problems or defaults in the past; a total of 7 accounts have overdue records, of which 5 accounts have been seriously overdue for more than 90 days, which is a major negative factor in the credit rating.

[0095] 2-3. Credit card records: In terms of credit cards, Zhang San has multiple accounts that have been overdue, and two of the cards have been listed as bad debts, indicating that he has not managed his credit card debts well in the past; although some accounts have never been overdue, overall, his credit card usage habits need to be improved.

[0096] 2-4. Loan records: In terms of loans, Zhang San has one loan that has become a bad debt, and several other loans have been overdue, which further highlights his challenges in financial management. However, he also has some loans that have not been overdue, indicating that he is responsible in the management of some loans.

[0097] 2-5. Non-credit transaction records and public records: In the non-credit transaction records, Zhang San has a fixed-line telephone bill in arrears, which shows the insufficiency of financial management in daily life; in the public records, Zhang San has a record of unpaid taxes, civil judgments and compulsory executions, which have a negative impact on his credit status, especially in legal disputes involving property, he was sentenced to compensate others for a large amount of money, and there is a record of completion of execution, which shows that he has the ability to fulfill his past economic responsibilities.

[0098] 2-6, Inquiry records: In the past two years, several financial institutions have conducted inquiries on Zhang San. This may be the result of his attempt to apply for a new loan or credit service. It also means that his credit history is frequently reviewed.

[0099] In summary, the credit report of customer Zhang San shows complex and diverse credit activities, including a lot of negative information, such as advances, asset disposals, overdue payments, and legal disputes. These not only reflect his difficulties in financial management in the past, but may also have an adverse impact on obtaining loans or credit services in the future. Therefore, when considering providing new credit to Zhang San, financial institutions should carefully assess his credit risk, focus on his repayment ability and willingness, and take into account his married background and possible joint financial responsibilities. "

[0100] It can be understood that in the analysis results output by the above-mentioned large language model, the specific label fields can be input into the subsequent business system model for analysis and modeling, or directly used as the basis for business system evaluation; the analysis reference text can be used as the basis for manual verification. Through the summary function of the large language model, the workload of manual reading of large amounts of information can be greatly reduced.

[0101] Through the embodiment of the present application, the method of "uploading" data by users sharing the screen is adopted to avoid the impact on user information during the public-to-public information query process, which is more "friendly" to some customers; avoid the relevant costs of public-to-public inquiries; technically guarantee the authenticity of user-reported data to avoid user tampering and cheating; can be applied to various businesses, such as after the function of the credit report is developed, only the corresponding template or identification field needs to be changed, and it can be reused in the query of other information such as provident fund and social security. In addition, the analysis and summary function based on the large language model can assist other business systems in making business decisions and judgments, and help reduce labor costs in the manual review process.

[0102] It is understandable that in the specific implementation of this application, related data such as user information is involved. When the above embodiments of this application are applied to specific products or technologies, user permission or consent is required, and the collection, use and processing of relevant data need to comply with relevant laws, regulations and standards of relevant countries and regions.

[0103] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present application.

[0104] According to another aspect of the embodiments of the present application, an account identification device for implementing the above account identification method is also provided. Figure 4 As shown, the device comprises:

[0105] A setting module 402, configured to set a target prompt word according to a target business scenario, wherein the target prompt word is used to instruct the target large language model to process business data based on the target business scenario;

[0106] The acquisition module 404 is used to acquire the business data corresponding to the target account, wherein the business data is matched with the target image displayed by the target application and the business process template by jumping to the target application. If the match is successful and the target application is always running in the system foreground, the target image is recognized and determined by text recognition;

[0107] The recognition module 406 is used to input the target prompt word and business data into the target large language model to obtain a target recognition result, wherein the target recognition result includes a risk label, and the risk label is used to indicate whether the target account is a risky account.

[0108] As an optional solution, the above-mentioned device is used to input the target prompt words and business data into the target large language model in the following manner to obtain the target recognition result: input the target prompt words and business data into the target large language model to determine the risk type, risk level and the reason why the target account is at risk; determine the risk label based on the risk type and risk level; generate the target recognition result according to the risk label.

[0109] As an optional solution, the above-mentioned device is used to input the target prompt words and business data into the target large language model in the following manner to obtain the target recognition result: input the target prompt words and business data into the target large language model to determine the risk type, risk level and the reason why the target account is at risk; determine the risk label based on the risk type and risk level, and determine the account risk information based on the risk label and the reason why the target account is at risk; generate the target recognition result based on the risk label and the account risk information.

[0110] As an optional solution, the above-mentioned device is also used to: obtain a preset target business field, wherein the target business field represents a field that can affect the risk label; extract target business data from the business data according to the target business field, wherein the target business data represents the business data corresponding to the target business field; splice the target prompt word and the business data to obtain target input data, and input the target input data into the target large language model to obtain a target recognition result.

[0111] As an optional solution, the above-mentioned device is also used to: embed target prompt words in the initial sample data to obtain target sample data, wherein the target sample data represents the initial sample data associated with the target business scenario; use the target sample data to train the initial large language model to obtain the target large language model.

[0112] As an optional solution, the above-mentioned device is used to obtain the business data corresponding to the target account in the following manner: based on the target image, query whether the target template exists in the business process template and whether the target application is running in the system foreground, wherein the image features of the target template and the target image meet the preset matching conditions; when the target template exists and the target application is always running in the system foreground, perform text recognition operation on the target image to obtain business data, and store the business data and the target image in the target database.

[0113] As an optional solution, the above-mentioned device is also used for at least one of the following: when the target template does not exist, generating a second prompt message, wherein the second prompt message is used to indicate that the business data needs to be reacquired; when the target application switches from running in the system foreground to the system background, generating a third prompt message, wherein the third prompt message is used to indicate that the business data needs to be reacquired.

[0114] In the embodiments of the present application, the term "module" or "unit" refers to a computer program or a part of a computer program with a predetermined function, and works together with other related parts to achieve a predetermined goal, and can be implemented in whole or in part by using software, hardware (such as processing circuits or memories) or a combination thereof. Similarly, a processor (or multiple processors or memories) can be used to implement one or more modules or units. In addition, each module or unit can be part of an overall module or unit that includes the function of the module or unit.

[0115] Regarding the device in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated here.

[0116] According to one aspect of the present application, a computer program product is provided. The computer program product includes a computer program.

[0117] The serial numbers of the above-mentioned embodiments of the present application are for description only and do not represent the advantages or disadvantages of the embodiments.

[0118] Figure 5 The structure block diagram of a computer system for implementing an electronic device according to an embodiment of the present application is schematically shown.

[0119] It should be noted that Figure 5 The computer system 500 of the electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0120] like Figure 5As shown, the computer system 500 includes a central processing unit 501 (CPU), which can perform various appropriate actions and processes according to the program stored in the read-only memory 502 (ROM) or the program loaded from the storage part 508 to the random access memory 503 (RAM). Various programs and data required for system operation are also stored in the random access memory 503. The central processing unit 501, the read-only memory 502 and the random access memory 503 are connected to each other through a bus 504. The input / output interface 505 (Input / Output interface, i.e., I / O interface) is also connected to the bus 504.

[0121] The following components are connected to the input / output interface 505: an input section 506 including a keyboard, a mouse, etc.; an output section 507 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker; a storage section 508 including a hard disk, etc.; and a communication section 509 including a network interface card such as a local area network card, a modem, etc. The communication section 509 performs communication processing via a network such as the Internet. A drive 510 is also connected to the input / output interface 505 as needed. A removable medium 511, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 510 as needed so that a computer program read therefrom is installed into the storage section 508 as needed.

[0122] In particular, according to an embodiment of the present application, the process described in each method flow chart can be implemented as a computer software program. For example, an embodiment of the present application includes a computer program product, which includes a computer program carried on a computer readable medium, and the computer program contains a program code for executing the method shown in the flow chart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 509, and / or installed from the removable medium 511. When the computer program is executed by the central processor 501, various functions defined in the system of the present application are executed.

[0123] In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 509, and / or installed from the removable medium 511. When the computer program is executed by the central processor 501, various functions provided by the embodiment of the present application are performed.

[0124] According to another aspect of the embodiment of the present application, an electronic device for implementing the above account identification method is also provided. The electronic device may be Figure 1 The terminal device or server shown in the figure. This embodiment is described by taking the electronic device as a terminal device as an example. Figure 6 As shown, the electronic device includes a memory 602 and a processor 604. The memory 602 stores a computer program, and the processor 604 is configured to execute the steps in any of the above method embodiments through the computer program.

[0125] Optionally, in this embodiment, the electronic device may be located in at least one network device among a plurality of network devices of a computer network.

[0126] Optionally, in this embodiment, the above-mentioned processor can be configured to execute the methods in each embodiment of the present application through a computer program.

[0127] Alternatively, a person skilled in the art may understand that: Figure 6 The structure shown is for illustration only. Figure 6 The structure of the electronic device is not limited. Figure 6 More or fewer components (such as network interfaces, etc.) as shown in, or with Figure 6 Different configurations are shown.

[0128] Among them, the memory 602 can be used to store software programs and modules, such as the program instructions / modules corresponding to the account identification method and device in the embodiments of the present application. The processor 604 executes various functional applications and data processing by running the software programs and modules stored in the memory 602, that is, to implement the above-mentioned account identification method. The memory 602 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 602 may further include a memory remotely located relative to the processor 604, and these remote memories may be connected to the terminal via a network. Examples of the above-mentioned networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof. Among them, the memory 602 can be specifically, but not limited to, used to store information such as business data. As an example, such as Figure 6 As shown, the memory 602 may include, but is not limited to, the setting module 402, the acquisition module 404, and the identification module 406 in the account identification device. In addition, it may also include, but is not limited to, other module units in the account identification device, which will not be repeated in this example.

[0129] Optionally, the transmission device 606 is used to receive or send data via a network. Specific examples of the network may include a wired network and a wireless network. In one example, the transmission device 606 includes a network adapter (Network Interface Controller, NIC), which can be connected to other network devices and routers via a network cable so as to communicate with the Internet or a local area network. In one example, the transmission device 606 is a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0130] In addition, the electronic device further includes: a display 608 for displaying the business data; and a connection bus 610 for connecting various module components in the electronic device.

[0131] In other embodiments, the terminal device or server may be a node in a distributed system, wherein the distributed system may be a blockchain system, and the blockchain system may be a distributed system formed by connecting the multiple nodes through network communication. The nodes may form a peer-to-peer network, and any form of computing device, such as a server, terminal or other electronic device, may become a node in the blockchain system by joining the peer-to-peer network.

[0132] According to one aspect of the present application, a computer-readable storage medium is provided, and a processor of an electronic device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the electronic device executes the account identification method provided in various optional implementations of the above-mentioned account identification aspects.

[0133] Optionally, in this embodiment, the above-mentioned computer-readable storage medium can be configured to store data for executing the methods in various embodiments of the present application.

[0134] Optionally, in this embodiment, a person of ordinary skill in the art may understand that all or part of the steps in the various methods of the above embodiments may be completed by instructing hardware related to the terminal device through a program, and the program may be stored in a computer-readable storage medium, and the storage medium may include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a disk or an optical disk, etc.

[0135] The serial numbers of the above-mentioned embodiments of the present application are for description only and do not represent the advantages or disadvantages of the embodiments.

[0136] If the integrated units in the above embodiments are implemented in the form of software functional units and sold or used as independent products, they can be stored in the above computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for one or more electronic devices to execute all or part of the steps of the methods described in each embodiment of the present application.

[0137] In the above embodiments of the present application, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, please refer to the relevant description of other embodiments.

[0138] In the several embodiments provided in the present application, it should be understood that the disclosed application can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0139] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0140] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0141] The above is only a preferred implementation of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. An account identification method, characterized in that: include: Setting a target prompt word according to a target business scenario, wherein the target prompt word is used to instruct a target large language model to process business data based on the target business scenario; Acquire the business data corresponding to the target account, wherein the business data is matched with the target image displayed by the target application and the business process template by jumping to the target application, and when the match is successful and the target application is always running in the system foreground, perform text recognition and determination on the target image; The target prompt word and the business data are input into the target large language model to obtain a target recognition result, wherein the target recognition result includes a risk label, and the risk label is used to indicate whether the target account is a risky account.

2. The method according to claim 1, characterized in that The step of inputting the target prompt word and the business data into the target large language model to obtain a target recognition result includes: Input the target prompt word and the business data into the target large language model to determine the risk type and risk level; determining the risk label based on the risk type and the risk level; The target identification result is generated according to the risk label.

3. The method according to claim 1, characterized in that The step of inputting the target prompt word and the business data into the target large language model to obtain a target recognition result includes: Input the target prompt word and the business data into the target large language model to determine the risk type, risk level and the reason why the target account is at risk; Determine the risk label based on the risk type and the risk level, and determine the account risk information based on the risk label and the reason causing the target account to be at risk; The target identification result is generated according to the risk tag and the account risk information.

4. The method according to claim 1, characterized in that: The method further comprises: Acquire a preset target business field, wherein the target business field represents a field that can affect the risk tag; Extracting target business data from the business data according to the target business field, wherein the target business data represents the business data corresponding to the target business field; The target prompt word and the business data are concatenated to obtain target input data, and the target input data is input into the target large language model to obtain the target recognition result.

5. The method according to claim 1, characterized in that The method further comprises: Embedding the target prompt word in the initial sample data to obtain target sample data, wherein the target sample data represents the initial sample data associated with the target business scenario; The target sample data is used to train an initial large language model to obtain the target large language model.

6. The method according to claim 1, characterized in that The obtaining of the business data corresponding to the target account includes: Based on the target image, query whether there is a target template in the business process template and whether the target application is running in the system foreground, wherein the image features of the target template and the target image meet a preset matching condition; In the case where the target template exists and the target application is always running in the system foreground, a text recognition operation is performed on the target image to obtain the business data, and the business data and the target image are stored in a target database.

7. The method according to claim 5, characterized in that The method further comprises at least one of the following: If the target template does not exist, generating second prompt information, wherein the second prompt information is used to indicate that the service data needs to be re-acquired; When the target application switches from running in the system foreground to running in the system background, a third prompt information is generated, wherein the third prompt information is used to indicate that the business data needs to be re-acquired.

8. An account identification device, characterized in that: include: A setting module, used to set a target prompt word according to a target business scenario, wherein the target prompt word is used to instruct the target large language model to process business data based on the target business scenario; an acquisition module, configured to acquire the business data corresponding to the target account, wherein the business data is matched with the target image displayed by the target application and the business process template by jumping to the target application, and when the match is successful and the target application is always running in the foreground of the system, the target image is recognized and determined by text recognition; The recognition module is used to input the target prompt word and the business data into the target large language model to obtain a target recognition result, wherein the target recognition result includes a risk label, and the risk label is used to indicate whether the target account is a risk account.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored computer program, wherein the computer program can be executed by an electronic device to perform the method described in any one of claims 1 to 7.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method described in any one of claims 1 to 7 are implemented.

11. An electronic device comprising a memory and a processor, characterized in that: A computer program is stored in the memory, and the processor is configured to execute the method according to any one of claims 1 to 7 through the computer program.

Images