Risk prompting method and device, storage medium and electronic equipment

Abstract

The invention discloses a risk prompting method and device, a storage medium and electronic equipment. The risk prompting method comprises the following steps: acquiring to-be-detected information of at least one data type involved when a user executes a service; inputting the to-be-detected information into an identification model, and identifying each piece of key point information contained in the to-be-detected information; judging whether the to-be-detected information is abnormal or not according to the key point information; when it is determined that the to-be-detected information is abnormal, risk type information corresponding to risks involved in the to-be-detected information is determined according to the key point information; and inputting the risk type information into a preset big language model to generate risk prompt information for the to-be-detected information through the big language model, and performing risk prompt according to the risk prompt information. According to the scheme, diversified risk prompt information can be generated through the large language model, so that the attention of the user to the abnormal prompt is fully aroused, and the privacy and property safety of the user are ensured.

Description

Technical Field

[0001] This specification relates to the field of Internet technology, and in particular to a risk warning method, device, storage medium and electronic device. Background Art

[0002] With the rapid development of artificial intelligence and Internet technologies, users are facing increasing challenges in terms of privacy and property security when performing business operations such as financial transactions and wealth management. In particular, when users become aware of potential security risks or when abnormalities occur during business execution, further risk detection is needed to identify risks.

[0003] However, after identifying that a business has risks, users are usually prompted based on pre-generated fixed text to inform them that the business they are currently executing has security risks. However, the content and style of this prompt form are relatively simple, and readability is also limited. It is difficult to attract users' attention, causing users to ignore the prompt information, and it is difficult to ensure the privacy and property safety of users.

[0004] Therefore, how to provide users with diversified and stylized information prompts after discovering risks in the business and further protect users' privacy and property safety is an urgent problem to be solved. Summary of the invention

[0005] This specification provides a method, device, storage medium and electronic device for risk detection. Risk warning information is generated based on risk type information through a large language model to provide risk warnings to users.

[0006] This manual adopts the following technical solutions:

[0007] This instruction manual provides a risk warning method, including:

[0008] Acquire business information of at least one data type involved in the user's business execution as information to be detected;

[0009] Inputting the information to be detected into a pre-trained recognition model, so as to identify key point information contained in the information to be detected through the recognition model, wherein the key point information is used to characterize the information content corresponding to the designated detected target in the information to be detected;

[0010] According to the key point information, determining whether the information to be detected is abnormal;

[0011] When it is determined that the information to be detected has an abnormality, determining risk type information corresponding to the risk involved in the information to be detected according to the key point information;

[0012] The risk type information is input into a preset large language model to generate risk warning information for the information to be detected through the large language model, and risk warning is performed according to the risk warning information.

[0013] Optionally, before judging whether the information to be detected has a risk according to the key point information, the method further includes:

[0014] Obtaining website information involved in the user's business execution;

[0015] Judging whether the information to be detected is abnormal according to the key point information specifically includes:

[0016] It is determined whether the information to be detected is abnormal according to the website information and the key point information.

[0017] Optionally, generating risk warning information for the information to be detected through the large language model specifically includes:

[0018] Determine the target text template;

[0019] Generate risk warning text for the information to be detected through the large language model;

[0020] The risk warning text is added to the target text template to obtain the risk warning information.

[0021] Optionally, determine a target text template, specifically including:

[0022] Obtaining user information of the user;

[0023] A text template matching the user information is determined in a preset template library as the target text template.

[0024] Optionally, the method further comprises:

[0025] The case information matching the risk type information is determined in the historical case library, and the case information is pushed to the user.

[0026] Optionally, training the recognition model includes:

[0027] Get historical information to be tested;

[0028] Inputting the historical information to be detected as a training sample into the recognition model, so as to identify each key point information contained in the historical information to be detected as predicted key point information through the recognition model;

[0029] The recognition model is trained with the optimization goal of minimizing the deviation between the predicted key point information and the key point information actually marked in the historical information to be detected.

[0030] Optionally, the data type of the information to be detected includes: text, image, and audio.

[0031] This manual provides a risk warning device, including:

[0032] An acquisition module, used to acquire business information of at least one data type involved in a user's business execution as information to be detected;

[0033] A recognition module, used for inputting the information to be detected into a pre-trained recognition model, so as to recognize key point information contained in the information to be detected through the recognition model, wherein the key point information is used to characterize the information content corresponding to the designated detected target in the information to be detected;

[0034] A judgment module, used for judging whether the information to be detected has an abnormality according to the key point information;

[0035] A determination module, configured to determine, when determining that the information to be detected has an abnormality, risk type information corresponding to the risk involved in the information to be detected based on the key point information;

[0036] The prompt module is used to input the risk type information into a preset large language model to generate risk prompt information for the information to be detected through the large language model, and to provide risk prompts according to the risk prompt information.

[0037] Optionally, before judging whether the information to be detected has risks according to the key point information, the acquisition module is further used to acquire website information involved in the user's business execution;

[0038] The judgment module is specifically used to judge whether the information to be detected has an abnormality according to the website information and the key point information.

[0039] Optionally, the prompt module is specifically used to determine a target text template;

[0040] A risk warning text for the information to be detected is generated through the large language model; the risk warning text is added to the target text template to obtain the risk warning information.

[0041] Optionally, the prompt module is specifically used to obtain user information of the user; and determine a text template matching the user information in a preset template library as the target text template.

[0042] Optionally, the prompt module is further used to determine case information matching the risk type information in a historical case library, and push the case information to the user.

[0043] Optionally, the device further comprises:

[0044] A training module is used to obtain historical information to be detected; input the historical information to be detected as a training sample into the recognition model, so as to identify the key point information contained in the historical information to be detected through the recognition model as predicted key point information; and train the recognition model with the optimization goal of minimizing the deviation between the predicted key point information and the key point information actually marked in the historical information to be detected.

[0045] This specification provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the risk warning method described above is implemented.

[0046] This specification provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the risk warning method described above is implemented when the processor executes the program.

[0047] At least one of the above technical solutions adopted in this specification can achieve the following beneficial effects:

[0048] In the risk warning method provided in the present specification, information to be detected of at least one data type involved in a user's business execution is obtained; the information to be detected is input into a recognition model to identify key point information contained in the information to be detected; based on the key point information, it is determined whether the information to be detected has an abnormality; when it is determined that the information to be detected has an abnormality, the risk type information corresponding to the risk involved in the information to be detected is determined based on the key point information; the risk type information is input into a preset large language model to generate risk warning information for the information to be detected through the large language model, and risk warnings are given based on the risk warning information.

[0049] It can be seen from the above method that, on the one hand, this solution can detect anomalies for various types of business data submitted by users, and on the other hand, generate diversified risk warning information through a large language model. Compared with the existing methods, this solution not only improves the accuracy of risk identification, but also the risk warning information generated by the large language model is more diversified and stylized, thereby fully attracting users' attention to risks and ensuring users' privacy and property safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] The drawings described herein are used to provide a further understanding of this specification and constitute a part of this specification. The exemplary embodiments and descriptions of this specification are used to explain this specification and do not constitute an improper limitation on this specification.

[0051] In the figure:

[0052] Figure 1 A schematic diagram of a risk detection method provided in this specification;

[0053] Figure 2 This is a schematic diagram of a credential submission page provided in this manual;

[0054] Figure 3 A schematic diagram of a page for generating a risk warning text provided in this instruction manual;

[0055] Figure 4 This is a page diagram of risk warning information provided in this manual;

[0056] Figure 5 A schematic diagram of the overall process of risk warning provided in this instruction manual;

[0057] Figure 6 A schematic diagram of a risk warning device provided in this manual;

[0058] Figure 7 A method provided in this specification corresponding to Figure 1 Schematic diagram of electronic equipment. DETAILED DESCRIPTION

[0059] In order to make the purpose, technical solutions and advantages of this specification more clear, the technical solutions of this specification will be clearly and completely described below in combination with the specific embodiments of this specification and the corresponding drawings. Obviously, the described embodiments are only part of the embodiments of this specification, not all of them. Based on the embodiments in this specification, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this specification.

[0060] Currently, when a transaction fails, the user will submit a picture voucher explaining the purpose of the transaction through the voucher entry provided by the client, and then it will be transferred to the manual review link in the background for risk assessment and a result will be given, or semi-automatic approval will be performed through certain regular rules, but the manual review link is time-consuming and related to the review volume. At present, rule-based review is mainly implemented through a series of expert experience model regularization, which has certain bottlenecks in general capabilities, and all judgment results are encapsulated in the model, and the configuration flexibility is relatively weak.

[0061] At present, the method of risk warning is mainly through manual writing of copywriting to form a copywriting library, and the use of expert rules and policy labels to output specified content to users; in some fraud control scenarios, generative artificial intelligence (AIGC) can be used to generate some text reminders, but the readability is still limited in text form. For users who are currently accustomed to short videos and multimedia output, stronger capabilities are needed to capture user attention.

[0062] The technical solutions provided by the embodiments of this specification are described in detail below in conjunction with the accompanying drawings.

[0063] Figure 1 The following is a flow chart of a risk warning method provided in this manual, which includes the following steps:

[0064] S100: Acquire business information of at least one data type involved in a business performed by a user as information to be detected.

[0065] In the risk control scenario, due to the information gap between the business provider's understanding of the purpose of the transaction and the actual situation, the accuracy of risky transactions that the risk control system can identify is limited, and transactions that are misaudited cause a very bad experience for users. In order to improve the ability to distinguish risks, there are already a variety of products such as voice and text that interact with users and collect information to confirm the user's transaction purpose and further identify possible risks.

[0066] In actual applications, abnormal business may be packaged with rhetoric to guide users to bypass risk control identification rules, resulting in users reporting erroneous information that does not match the actual situation; in addition, some users with immature security awareness tend to ignore security prompts before they perceive the risk. Both of the above situations increase the difficulty of identifying risk situations. Therefore, more reasonable methods are needed for the two interactive purposes of risk judgment and security awakening.

[0067] Based on this, this instruction manual provides a risk warning method that uses a multimodal solution for information collection and interaction. While improving the ability to differentiate risks, it further improves the diversity and readability of risk warnings, thereby improving the user experience while better protecting user privacy and property safety.

[0068] In this specification, the execution entity used to implement a risk warning method can be a designated device such as a server. For the sake of ease of description, the following will only take the server as the execution entity as an example to illustrate a risk warning method provided in this specification.

[0069] The server may obtain at least one type of service data involved in the user's service execution as the information to be detected.

[0070] When users are performing business operations such as transactions, financial management, and borrowing, they often have doubts about the current business and worry that the business may have security risks. In this case, users can actively submit information to be tested through the credential entrance provided by the business client, thereby further identifying risks in the current business.

[0071] In addition, the server's risk control system can also perform preliminary identification of the user's current business and guide the user to submit information to be tested when possible risks are detected. For example, when a user performs a business operation different from the past, a large amount of money flows suddenly, or a transaction is abnormally interrupted, the server can suspend the current business and guide the user to submit information to be tested through the credential entry provided by the business client, thereby further identifying the risks of the current business. For ease of understanding, this manual provides a schematic diagram of the credential submission page, such as Figure 2 shown.

[0072] Figure 2 This is a schematic diagram of a credential submission page provided in this manual.

[0073] Among them, users can upload video files, audio files or documents. In addition, users can also manually enter content that requires credentials.

[0074] In this specification, the information types of the above-mentioned information to be detected may include images (such as screenshots, photos, pictures, videos, etc.), texts (such as chat records, transaction records) and audio.

[0075] In addition, the server can also obtain website information (such as website link) involved in the user's business execution. The website information can be uploaded to the server together with the information to be detected by the user, or it can be automatically obtained by the server.

[0076] S102: Inputting the information to be detected into a pre-trained recognition model to identify key point information contained in the information to be detected through the recognition model, wherein the key point information is used to characterize the information content corresponding to the specified detected target in the information to be detected.

[0077] After receiving the information to be detected, the server may input the information to be detected into a pre-trained recognition model, thereby determining key point information contained in the information to be detected through the recognition model.

[0078] Among them, the key point information is used to characterize the information content corresponding to the specified detected target in the information to be detected. In this specification, the above-mentioned detection targets may include the identities of the payee and payee, the purpose of the transaction, the subject of the transaction, and the transaction amount, etc. Of course, for some picture or video data, the detection target may also include the objects contained therein and the businesses involved. Correspondingly, the information content corresponding to the detected target may include object information (such as people's clothing, behavior, etc.) and business type.

[0079] Before using the above recognition model, the server may pre-train the recognition model.

[0080] Specifically, the server may obtain historical information to be detected, and then input the historical information to be detected into the recognition model as training samples, so as to identify key point information contained in the historical information to be detected through the recognition model as predicted key point information.

[0081] The server can train the recognition model with the optimization goal of minimizing the deviation between the predicted key point information and the key point information actually marked in the historical information to be detected. Until the training goal is met (such as the model converges to a preset range or reaches a preset number of training times), the server can deploy the recognition model.

[0082] S104: Determine whether the information to be detected has any abnormality according to the key point information.

[0083] S106: When it is determined that the information to be detected has an abnormality, determine risk type information corresponding to the risk involved in the information to be detected according to the key point information.

[0084] The server can identify risks for key point information based on preset risk control strategies. After identifying risks based on key point information, the server can further determine risk type information.

[0085] For example, when the key point information contains an abnormal object, the server can determine the risk type information (such as object abnormality, business abnormality, etc.) based on the object information corresponding to the abnormal object. For another example, when at least one of the identities of the payee and payee, the transaction purpose, the transaction subject, and the transaction amount is abnormal, it can be determined that the information to be detected is at risk and output the risk type information (such as payee abnormality, payer abnormality, transaction purpose abnormality, transaction subject abnormality, and transaction amount abnormality, etc.).

[0086] It should be noted that the above risk control strategy can be implemented through a preset risk control model. After the key point information is input into the risk control model, the risk control model can output the risk judgment result and risk type information according to the key point information. Of course, it can also be implemented through pre-set expert rules or scripts, and this manual does not make specific restrictions on this.

[0087] At the same time, the server can perform security analysis on the URL information by means of inspection to determine whether the URL information has risks. For example, when the URL information is in the blacklist preset by the server, or the URL information is marked as an abnormal URL, the server can determine that the URL information has risks.

[0088] The server may determine whether the information to be detected has risks based on the identification result of the website information and the identification result corresponding to the detection information.

[0089] When at least one of the identification result of the website information and the identification result corresponding to the information to be detected detects a risk, the server can suspend the service and output a prompt message. When the website information has a risk, the risk type information determined by the server can include: website anomaly.

[0090] Of course, the server may not obtain the website information, but only determine whether the current service is abnormal through the information to be detected uploaded by the user.

[0091] When the server detects that there is no abnormality in the data to be detected and the website information, it means that there is no abnormality in the business currently executed by the user. At this time, the service can send a prompt to the user to allow the execution of subsequent business.

[0092] S108: Inputting the risk type information into a preset large language model to generate risk warning information for the information to be detected through the large language model, and performing risk warning according to the risk warning information.

[0093] After determining that the business executed by the user is risky and obtaining risk type information, the server can input the risk type information into a preset large language model and generate risk warning text for the information to be detected based on the specified information generation conditions. For ease of understanding, this manual provides a schematic diagram of a risk warning text generation page, such as Figure 3 shown.

[0094] Figure 3 This is a schematic diagram of a page for generating a risk warning text provided in this instruction manual.

[0095] The above information generation conditions can be used to restrict the data format and data content of the data description information to be generated. For example, when the risk warning information is "abnormal transaction purpose", the corresponding information generation condition can be "use two to three sentences to remind users of abnormal transaction purpose, with no more than 50 words". The final risk warning text can be: "The transaction purpose of the current transaction has a high risk. If you continue to operate, it may cause unnecessary property losses. For the safety of your property, please be cautious!"

[0096] Of course, the server can also directly input the risk type information into a pre-trained large language model, so that the large language model directly generates a risk warning text based on the risk type information.

[0097] Furthermore, the server can obtain user information of the current user, which may include the user's age, gender, etc., and then the server can determine a text template that matches the above user information in the preset template library as the target text template. Furthermore, the server can fill the risk warning information into the target text template according to the text format corresponding to the target text template, thereby generating risk warning information, and displaying it to the user, thereby providing risk warnings to the user. In this way, the server can generate differentiated prompt graphics and texts for different groups of people (such as using different graphic styles and proportions for the elderly and young people respectively). Among them, the template library pre-stores text templates corresponding to different user information, and the data type of the text template can be an image. For ease of understanding, this instruction manual provides a page schematic diagram of risk warning information, such as Figure 4 shown.

[0098] Figure 4 This is a page diagram of risk warning information provided in this instruction manual.

[0099] Among them, risk warning information can be sent to the terminal devices used by the user (such as mobile phones, tablets), etc. through methods such as text messages, instant messaging (IM), etc., and displayed on the terminal device in the form of pop-up windows. Of course, it can also be directly displayed on other terminal devices where the user is currently performing operations (such as deposit and withdrawal machines, etc.).

[0100] It should be pointed out that the server may not determine the text template, but directly use the risk warning text output by the large language model as the risk warning information.

[0101] In addition, the server can also determine case information matching the risk type information in the historical case library, and push the case information to the user, thereby further prompting the user of the risks existing in the current business.

[0102] For ease of understanding, this manual also provides a schematic diagram of the overall process of risk warning, such as Figure 5 shown.

[0103] Figure 5 A schematic diagram of the overall process of risk warning provided in this instruction manual.

[0104] Among them, the server can extract key point information from the information to be detected through the information extraction model, and obtain the website information involved when the user performs the business. Then, based on the preset risk control strategy, it makes risk judgments based on the inspection results of the website information and the key point information, and outputs risk type information when it is determined that there is an abnormality in the business. Then, the risk type information is input into the large language model to generate risk warning information.

[0105] From the above methods, it can be seen that this solution can provide links, pictures, videos and other multi-modal media at the user input level to judge the risk situation with more factual materials; at the user output level, it can also improve the user's risk awareness in a more readable way through pictures, texts and other multi-modal methods. The above solution, with the help of the current artificial intelligence large model, can quickly judge these situations in the transaction link without human participation.

[0106] The above are one or more risk warning methods for implementing this specification. Based on the same idea, this specification also provides corresponding risk warning devices, such as Figure 6 shown.

[0107] Figure 6 A schematic diagram of a risk warning device provided for this manual includes:

[0108] An acquisition module 600 is used to acquire business information of at least one data type involved in a user's business execution as information to be detected;

[0109] The recognition module 602 is used to input the information to be detected into a pre-trained recognition model, so as to recognize each key point information contained in the information to be detected through the recognition model, wherein the key point information is used to characterize the information content corresponding to the specified detected target in the information to be detected;

[0110] A judgment module 604 is used to judge whether the information to be detected has an abnormality according to the key point information;

[0111] A determination module 606 is used to determine, when determining that the information to be detected has an abnormality, risk type information corresponding to the risk involved in the information to be detected based on the key point information;

[0112] The prompt module 608 is used to input the risk type information into a preset large language model to generate risk prompt information for the information to be detected through the large language model, and to provide risk prompts according to the risk prompt information.

[0113] Optionally, before judging whether the information to be detected has risks according to the key point information, the acquisition module 600 is further used to acquire website information involved in the user's business execution;

[0114] The judgment module 604 is specifically used to judge whether the information to be detected has an abnormality according to the website information and the key point information.

[0115] Optionally, the prompt module 608 is specifically used to determine a target text template; generate a risk warning text for the information to be detected through the large language model; and add the risk warning text to the target text template to obtain the risk warning information.

[0116] Optionally, the prompt module 608 is specifically configured to obtain user information of the user; and determine a text template matching the user information in a preset template library as the target text template.

[0117] Optionally, the prompt module 608 is further used to determine case information matching the risk type information in a historical case library, and push the case information to the user.

[0118] Optionally, the device further comprises:

[0119] The training module 610 is used to obtain historical information to be detected; input the historical information to be detected as a training sample into the recognition model, so as to identify the key point information contained in the historical information to be detected through the recognition model as predicted key point information; and train the recognition model with the optimization goal of minimizing the deviation between the predicted key point information and the key point information actually marked in the historical information to be detected.

[0120] Optionally, the data type of the information to be detected includes: text, image, and audio.

[0121] This specification also provides a computer-readable storage medium, which stores a computer program, which can be used to execute the above Figure 1 A risk warning method provided.

[0122] This manual also provides Figure 7 The one shown corresponds to Figure 1 A schematic diagram of the electronic device. Figure 7As mentioned above, at the hardware level, the electronic device includes a processor, an internal bus, a network interface, a memory, and a non-volatile memory, and may also include other hardware required for the business. The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs it to achieve the above Figure 1 The risk warning method described above. Of course, in addition to the software implementation method, this specification does not exclude other implementation methods, such as logic devices or a combination of software and hardware, etc., that is to say, the execution subject of the following processing flow is not limited to each logic unit, but can also be hardware or logic devices.

[0123] In the 1990s, improvements to a technology could be clearly distinguished as hardware improvements (for example, improvements to the circuit structure of diodes, transistors, switches, etc.) or software improvements (improvements to the method flow). However, with the development of technology, many improvements to the method flow today can be regarded as direct improvements to the hardware circuit structure. Designers almost always obtain the corresponding hardware circuit structure by programming the improved method flow into the hardware circuit. Therefore, it cannot be said that an improvement in a method flow cannot be implemented using a hardware entity module. For example, a programmable logic device (PLD) (such as a field programmable gate array (FPGA)) is such an integrated circuit whose logical function is determined by the user's programming of the device. Designers can "integrate" a digital system on a PLD by programming it themselves, without having to ask a chip manufacturer to design and produce a dedicated integrated circuit chip. Moreover, nowadays, instead of manually making integrated circuit chips, this kind of programming is mostly implemented by "logic compiler" software, which is similar to the software compiler used when developing and writing programs, and the original code before compilation must also be written in a specific programming language, which is called hardware description language (HDL). There is not only one HDL, but many kinds, such as ABEL (Advanced Boolean Expression Language), AHDL (Altera Hardware Description Language), Confluence, CUPL (Cornell University Programming Language), HDCal, JHDL (Java Hardware Description Language), Lava, Lola, MyHDL, PALASM, RHDL (Ruby Hardware Description Language), etc. The most commonly used ones are VHDL (Very-High-Speed ​​Integrated Circuit Hardware Description Language) and Verilog. Those skilled in the art should also know that it is only necessary to program the method flow slightly in the above-mentioned hardware description languages ​​and program it into the integrated circuit, and then it is easy to obtain the hardware circuit that implements the logic method flow.

[0124] The controller can be implemented in any appropriate manner, for example, the controller can take the form of a microprocessor or processor and a computer-readable medium storing a computer-readable program code (such as software or firmware) that can be executed by the (micro)processor, a logic gate, a switch, an application-specific integrated circuit (ASIC), a programmable logic controller, and an embedded microcontroller. Examples of controllers include, but are not limited to, the following microcontrollers: ARC 625D, Atmel AT91SAM, Microchip PIC18F26K20, and Silicone Labs C8051F320. The memory controller can also be implemented as part of the control logic of the memory. Those skilled in the art also know that in addition to implementing the controller in a purely computer-readable program code manner, the controller can be implemented in the form of a logic gate, a switch, an application-specific integrated circuit, a programmable logic controller, and an embedded microcontroller by logically programming the method steps. Therefore, this controller can be considered as a hardware component, and the devices included therein for implementing various functions can also be regarded as structures within the hardware component. Or even, the devices for implementing various functions can be regarded as both software modules for implementing the method and structures within the hardware component.

[0125] The systems, devices, modules or units described in the above embodiments may be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, the computer may be, for example, a personal computer, a laptop computer, a cellular phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

[0126] For the convenience of description, the above device is described in various units according to their functions. Of course, when implementing this specification, the functions of each unit can be implemented in the same or multiple software and / or hardware.

[0127] Those skilled in the art will appreciate that the embodiments of this specification may be provided as methods, systems, or computer program products. Therefore, this specification may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, this specification may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0128] This specification is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of this specification. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0129] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0130] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0131] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0132] The memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0133] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0134] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0135] It will be appreciated by those skilled in the art that the embodiments of this specification may be provided as methods, systems or computer program products. Therefore, this specification may take the form of a complete hardware embodiment, a complete software embodiment or an embodiment combining software and hardware. Moreover, this specification may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0136] This specification may be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. This specification may also be practiced in distributed computing environments where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules may be located in local and remote computer storage media, including storage devices.

[0137] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0138] The above description is only an embodiment of the present specification and is not intended to limit the present specification. For those skilled in the art, the present specification may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present specification shall be included in the scope of the claims of the present specification.

Claims

1. A risk warning method, comprising: Acquire business information of at least one data type involved in the user's business execution as information to be detected; Inputting the information to be detected into a pre-trained recognition model, so as to identify key point information contained in the information to be detected through the recognition model, wherein the key point information is used to characterize the information content corresponding to the designated detected target in the information to be detected; According to the key point information, determining whether the information to be detected is abnormal; When it is determined that the information to be detected has an abnormality, determining risk type information corresponding to the risk involved in the information to be detected according to the key point information; The risk type information is input into a preset large language model to generate risk warning information for the information to be detected through the large language model, and risk warning is performed according to the risk warning information.

2. The method according to claim 1, before judging whether the information to be detected has risks according to the key point information, the method further comprises: Obtaining website information involved in the user's business execution; Judging whether the information to be detected is abnormal according to the key point information specifically includes: It is determined whether the information to be detected is abnormal according to the website information and the key point information.

3. The method according to claim 1, generating risk warning information for the information to be detected through the large language model, specifically comprising: Determine the target text template; Generate risk warning text for the information to be detected through the large language model; The risk warning text is added to the target text template to obtain the risk warning information.

4. The method according to claim 3, wherein determining the target text template comprises: Obtaining user information of the user; A text template matching the user information is determined in a preset template library as the target text template.

5. The method of claim 1, further comprising: The case information matching the risk type information is determined in the historical case library, and the case information is pushed to the user.

6. The method of claim 1, wherein training the recognition model comprises: Get historical information to be tested; Inputting the historical information to be detected as a training sample into the recognition model, so as to identify each key point information contained in the historical information to be detected as predicted key point information through the recognition model; The recognition model is trained with the optimization goal of minimizing the deviation between the predicted key point information and the key point information actually marked in the historical information to be detected.

7. The method according to claim 1, wherein the data type of the information to be detected comprises: Text, images, audio.

8. A risk warning device, comprising: An acquisition module, used to acquire business information of at least one data type involved in a user's business execution as information to be detected; A recognition module, used for inputting the information to be detected into a pre-trained recognition model, so as to recognize key point information contained in the information to be detected through the recognition model, wherein the key point information is used to characterize the information content corresponding to the designated detected target in the information to be detected; A judgment module, used for judging whether the information to be detected has an abnormality according to the key point information; A determination module, configured to determine, when determining that the information to be detected has an abnormality, risk type information corresponding to the risk involved in the information to be detected based on the key point information; The prompt module is used to input the risk type information into a preset large language model to generate risk prompt information for the information to be detected through the large language model, and to provide risk prompts according to the risk prompt information.

9. The device according to claim 8, before judging whether the information to be detected has risks according to the key point information, the acquisition module is further used to acquire website information involved in the user's business execution; The judgment module is specifically used to judge whether the information to be detected has an abnormality according to the website information and the key point information.

10. The device according to claim 8, wherein the prompt module is specifically used to determine a target text template; A risk warning text for the information to be detected is generated through the large language model; the risk warning text is added to the target text template to obtain the risk warning information.

11. The device as described in claim 10, wherein the prompt module is specifically used to obtain user information of the user; and determine a text template matching the user information in a preset template library as the target text template.

12. The device as claimed in claim 8, wherein the prompt module is further used to determine case information matching the risk type information in a historical case library and push the case information to the user.

13. The apparatus of claim 8, further comprising: A training module is used to obtain historical information to be detected; input the historical information to be detected as a training sample into the recognition model, so as to identify the key point information contained in the historical information to be detected through the recognition model as predicted key point information; and train the recognition model with the optimization goal of minimizing the deviation between the predicted key point information and the key point information actually marked in the historical information to be detected.

14. A computer-readable storage medium storing a computer program, wherein the computer program, when executed by a processor, implements the method according to any one of claims 1 to 7.

15. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method according to any one of claims 1 to 7 when executing the program.

Images