Business processing method, electronic equipment and product

Abstract

The invention provides a service processing method, electronic equipment and a product, and the method comprises the steps: obtaining the problem description information of a target service inputted by a user through a front-end interaction interface of a service platform, and obtaining the service detail data of the target service; sending the problem description information and the business detail data to a business big language model matched with the target business for processing; obtaining a processing result of the business large language model, and forwarding the processing result to the business platform; wherein the processing result is used for indicating an answer result of the business big language model to the question description information.

Description

Technical Field

[0001] The present application relates to the technical field of computers, and in particular to a business processing method, electronic equipment, and product. Background Art

[0002] A business platform is a comprehensive system or framework that provides the foundational support and operating environment for business activities. Currently, intelligent technologies are gradually maturing and becoming increasingly prominent. In particular, the rapid development and application of large language models are enabling AI-powered upgrades to business platforms. For example, integrating large language models into business platforms can improve their processing efficiency.

[0003] When upgrading and improving a business platform, it's necessary to modify the platform's related business logic. For example, when improving a web-based business platform with artificial intelligence, the relevant business logic needs to be added and the page layout needs to be modified accordingly. This results in a large workload and consumes a lot of time and resources. Summary of the Invention

[0004] Embodiments of the present application provide a service processing method, device, electronic device, storage medium, and product.

[0005] In a first aspect, an embodiment of the present application provides a service processing method, including:

[0006] Obtaining problem description information of the target business entered by the user through the front-end interactive interface of the business platform, and obtaining business detail data of the target business;

[0007] Sending the problem description information and the business detail data to a business language model that matches the target business for processing;

[0008] Obtaining a processing result of the business large language model and forwarding the processing result to the business platform; wherein the processing result is used to indicate an answer result of the business large language model to the question description information.

[0009] This technical solution decouples the business platform from the processing of problem descriptions and business details. Because the large business language model can operate independently based on the received problem descriptions and business details, it is independent of the business platform. Therefore, there is no need to modify the relevant business logic of the business platform. This makes business processing more flexible, concise, and time-efficient.

[0010] In an optional implementation manner, after forwarding the processing result to the service platform, the method further includes:

[0011] In response to the user triggering an interactive window in the front-end interactive interface, determining first filling information for an area to be filled in the interactive window; wherein the interactive window is a window for displaying the processing result;

[0012] The first filling information is sent to the front-end interactive interface for display, so that the front-end interactive interface fills the first filling information into the to-be-filled area and displays it.

[0013] In this technical solution, after forwarding the processing results to the business platform, the first filling information of the area to be filled is determined by capturing the user's trigger operation on the processing results in the AI interactive window and pushing it to the front-end interactive interface for display, thereby achieving real-time response to user operations and facilitating users to perform subsequent operations based on the displayed information.

[0014] In an optional implementation, in response to the user triggering the interactive window in the front-end interactive interface, determining first filling information for the area to be filled in the interactive window includes:

[0015] After detecting a click operation of the user on the target processing result in the interactive window, the first filling information is generated based on result description information of the target processing result.

[0016] In this technical solution, after detecting that the user clicks on the target processing result in the interactive window of the front-end interactive interface, the first filling information of the area to be filled in the interactive window can be generated based on the result description information, so as to dynamically generate and fill information according to user operations, thereby improving the interactive experience and the targeted information display.

[0017] In an optional implementation, in response to the user triggering the interactive window in the front-end interactive interface, determining first filling information for the area to be filled in the interactive window includes:

[0018] receiving filling suggestion information input by the user in the interactive window;

[0019] First filling information of the area to be filled is generated based on the filling suggestion information.

[0020] In this technical solution, by receiving the filling suggestion information input by the user in the interactive window of the front-end interactive interface and generating the first filling information of the area to be filled in the interactive window based on this, effective utilization of user input is achieved, user participation is enhanced, and the personalization of information filling and the degree of compliance with user needs are improved.

[0021] In an optional implementation manner, forwarding the processing result to the service platform includes:

[0022] In the case where there are multiple processing results, determining a confidence level for each processing result; wherein the confidence level is used to indicate a degree of credibility of each processing result;

[0023] Forwarding each of the processing results and the confidence level of each of the processing results to the service platform.

[0024] In this technical solution, by determining the confidence level of the processing results and displaying the confidence level to the user, a trigger basis can be provided for the user to trigger the operation of the processing results, thereby guiding the user to select a more matching processing result and improving the accuracy and reliability of the business processing process.

[0025] In an optional implementation manner, after forwarding the processing result to the service platform, the method further includes:

[0026] In the case where no triggering operation of the user on the processing result in the interactive window is detected, determining the processing result with the highest confidence;

[0027] Based on the result description information of the processing result with the highest confidence, second filling information is determined, and the second filling information is sent to the front-end interactive interface for display.

[0028] In this technical solution, automatic filling and visual display of information can be achieved, further improving the user interaction experience and reducing user operation costs.

[0029] In an optional implementation, obtaining the problem description information of the target business input by the user through the front-end interactive interface of the business platform includes:

[0030] Detecting an input operation of the user in the question input area of the front-end display interface;

[0031] Based on the input content, problem description information of the target business is determined.

[0032] In this technical solution, the problem description information of the target business can be determined by detecting user operations in the problem input area of the front-end interactive interface and analyzing the input content, thereby providing a basis for business processing and ensuring the development of subsequent business.

[0033] In an optional implementation manner, after forwarding the processing result to the service platform, the method further includes:

[0034] Determining target historical operation data similar to the problem description information of the target business from the historical operation data of each business on the business platform;

[0035] Determining a historical processing opinion for the target historical operation data, and forwarding the historical processing opinion to the business platform;

[0036] Based on the user's triggering operation on the historical processing opinions and / or the processing results, third filling information is determined, and the third filling information is sent to the front-end interactive interface for display.

[0037] In this technical solution, historical operation data can be fully utilized to provide more targeted processing opinions and suggestions for problem description information, thereby improving the efficiency and quality of target business processing.

[0038] In an optional implementation manner, sending the problem description information and the business detail data to a business language model matching the target business for processing includes:

[0039] Determining target description data related to the problem description information in the business detail data;

[0040] Fusing the problem description information with the target description data to obtain target business data;

[0041] The target business data is input into the business large language model for processing to obtain a processing result of the large language model on the target business data.

[0042] In this technical solution, by determining the target description data in the business details data, fusing it with the problem description information to form the target business data, and inputting it into the business large language model for processing, the data processing scope can be effectively narrowed and the information processing efficiency can be improved.

[0043] In an optional implementation, the business language model is obtained by training in the following manner:

[0044] Acquire historical business data of a specified business module in the business platform and general data of the business large language model;

[0045] Splitting the historical business data to obtain a plurality of first sub-data, and splitting the general data to obtain a plurality of second sub-data;

[0046] The large language model to be trained is trained alternately by using the multiple first sub-data and the multiple second sub-data to obtain the business large language model.

[0047] In this technical solution, historical business data and general data of the business platform are obtained, split into first sub-data and second sub-data respectively, and then the large language model to be trained is alternately trained to obtain a business large language model. The business large language model has both specific business domain knowledge and general language capabilities, thereby improving its applicability and accuracy in the business scenarios corresponding to the business platform.

[0048] In a second aspect, an embodiment of the present application further provides a service processing device, including:

[0049] An acquisition module is used to acquire problem description information of a target business input by a user through the front-end interactive interface of the business platform, and to acquire business detail data of the target business;

[0050] A processing module, configured to send the problem description information and the business detail data to a business language model matching the target business for processing;

[0051] A forwarding module is used to obtain the processing result of the business large language model and forward the processing result to the business platform; wherein, the processing result is used to indicate the answer result of the business large language model to the question description information.

[0052] In a third aspect, an embodiment of the present application further provides an electronic device comprising: a processor and a memory, wherein the memory stores machine-readable instructions executable by the processor. When the electronic device is running, the processor communicates with the memory, and when the machine-readable instructions are executed by the processor, the steps of the above-mentioned first aspect or any possible implementation of the first aspect are performed.

[0053] In a fourth aspect, an embodiment of the present application further provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the above-mentioned first aspect or any possible implementation of the first aspect are executed.

[0054] In a fifth aspect, an embodiment of the present application further provides a computer program product, comprising a computer program / instruction, which implements the steps of the above-mentioned process processing method when executed by a processor.

[0055] In the above implementation, after obtaining the problem description information of the target business entered by the user through the front-end interactive interface of the business platform and obtaining the business details data, the problem description information and business details data are forwarded to the business big language model for processing. This can achieve decoupling between the business platform and the business big language model. At this time, the business big language model can run independently from the business platform and is not dependent on the business platform. At this time, when the business platform accesses the business big language model for text processing, there is no need to change the relevant business logic of the business platform, thereby making business processing more flexible and concise, while improving time efficiency.

[0056] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following is a brief introduction to the drawings required for use in the embodiments. The drawings herein are incorporated into and constitute a part of the specification. These drawings illustrate embodiments consistent with the present application and, together with the specification, are used to illustrate the technical solutions of the present application. It should be understood that the following drawings only illustrate certain embodiments of the present application and should not be regarded as limiting the scope. For those of ordinary skill in the art, other relevant drawings can be obtained based on these drawings without inventive effort.

[0058] Figure 1 A flowchart of a business processing method provided by an embodiment of the present application is shown;

[0059] Figure 2 A schematic diagram showing a front-end interactive interface of the business processing method provided in an embodiment of the present application;

[0060] Figure 3 A schematic diagram showing another front-end interactive interface of the business processing method provided in an embodiment of the present application is shown;

[0061] Figure 4 A flowchart of a business large language model training of a business processing method provided in an embodiment of the present application is shown;

[0062] Figure 5 A flowchart showing the target plug-in of the business processing method provided in an embodiment of the present application is shown;

[0063] Figure 6 A schematic diagram of a business processing system provided in an embodiment of the present application is shown;

[0064] Figure 7A schematic diagram of a service processing device provided in an embodiment of the present application is shown;

[0065] Figure 8 A schematic diagram of an electronic device provided in an embodiment of the present application is shown. DETAILED DESCRIPTION

[0066] In order to make the purpose, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions 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 a part of the embodiments of the present application, rather than all the embodiments. The components of the embodiments of the present application generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application for protection, but merely represents the selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without making creative work are within the scope of protection of this application.

[0067] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, it does not require further definition or explanation in subsequent drawings.

[0068] The term "and / or" herein simply describes an association relationship, indicating that three relationships can exist. For example, A and / or B can represent the existence of A alone, the simultaneous existence of A and B, and the existence of B alone. In addition, the term "at least one" herein refers to any combination of at least two of any one or more of a plurality of items. For example, "at least one of A, B, and C" can represent any one or more elements selected from the set consisting of A, B, and C.

[0069] Before going into detail, the terms used in the embodiments of the present application are explained as follows:

[0070] Artificial Intelligence: Artificial Intelligence refers to technologies and methods that enable computer systems or machines to simulate human intelligence. AI systems are able to perform tasks that typically require human intelligence, such as visual perception, speech recognition, decision-making, and language translation.

[0071] DOM (Document Object Model): This connects web pages to scripts or programming languages by storing the structure of a document (for example, the HTML representing a web page) as objects in memory. The DOM represents documents using a logical tree. Each branch of the tree ends in a node, and each node contains an object. DOM methods allow programmatic access to the tree. These methods can be used to modify the structure, style, or content of a document.

[0072] Grid management: Administratively divide urban areas into "grids," making these grids the smallest management units for target management. For example, grid management can divide the jurisdictional area into several grid-like units based on principles such as territorial management, geographical layout, and current status management, and implement dynamic and comprehensive management of each grid. It is a digital management model.

[0073] Large models: These can be understood as deep learning models with very large parameters and complex structures. Compared to traditional machine learning models, large models have stronger expressive capabilities and are better able to handle complex tasks.

[0074] To facilitate understanding of this embodiment, we first provide a detailed introduction to a business processing method disclosed in this embodiment. The execution entity of the business processing method provided in this embodiment is generally an electronic device with certain computing capabilities. In some possible implementations, this business processing method can be implemented by a processor invoking computer-readable instructions stored in a memory.

[0075] See also Figure 1 FIG. 1 is a flowchart of a business processing method provided by an embodiment of the present application. The method is applied to a target plug-in, which is set in a front-end interactive device of a business platform. The method includes steps S101 to S103, wherein:

[0076] S101: Obtain problem description information of a target business input by a user through a front-end interactive interface of a business platform, and obtain business detail data of the target business.

[0077] In an embodiment of the present application, a connection can be established with a business platform through a target plug-in, and data can be sent to the business platform and received from the business platform through the target plug-in; business detail data of the target business can also be obtained from the business platform through the target plug-in.

[0078] Here, the target plug-in can obtain element information of each UI element in the front-end interactive interface, thereby determining the element information of the UI element as the business detail data of the target business. In addition, the target plug-in can also obtain business detail data of the target business from the business platform database. Specifically, it can determine data related to the problem description information of the target business from the full data of the business platform as the business detail data.

[0079] The business platform can be built using cloud computing services, open source software (e.g., web servers), and low-code or no-code platforms. The business platform can be determined based on the actual needs of the user. For example, the business platform can be a grid-based government business platform. In addition, it can also be other business platforms. All business platforms that can establish a connection with the target plug-in are within the scope of protection of the technical solution of this application and are not listed here one by one.

[0080] The business platform can issue permissions to the target plug-in based on determining that the target plug-in is a trusted plug-in. Through this permission, the target plug-in can be allowed to access data in the business platform and send data to the business platform.

[0081] For example, if the business platform determines that the target plug-in is a trusted plug-in, the business platform can send permissions to the target plug-in and randomly generate an account and password corresponding to the target plug-in. If the target plug-in logs in to the business platform using this account and password, the target plug-in can access data in the business platform and send data to the business platform.

[0082] In an embodiment of the present application, a "question input area" may be provided in the front-end display interface. Upon detecting a user triggering the question input area in the front-end display interface, the target plug-in may obtain the question description information regarding the target business entered in the question input area by the user. The triggering operation may include at least one of the following: text entry, button click, and message feature.

[0083] S102: Send the problem description information and business detail data to a business language model that matches the target business for processing.

[0084] In an embodiment of the present application, the target plug-in can forward the problem description information and the business detail data to the business language model. Afterwards, the business language model can process the problem description information and the business detail data to obtain a processing result.

[0085] Here, the business large language model and the business platform are deployed separately. This approach not only avoids the performance bottleneck of the business platform caused by running the large language model, but also enables independent management of the business large language model, thereby better ensuring the security and stability of the business large language model and preventing the business large language model from being maliciously attacked or abused.

[0086] S103: Obtain the processing result of the business large language model and forward the processing result to the business platform; wherein the processing result is used to indicate the answer result of the business large language model to the question description information.

[0087] In an embodiment of the present application, after the business language model processes the problem description information and business detail data and obtains the processing result, the processing result can be sent to the target plug-in. After obtaining the processing result, the target plug-in can forward the processing result to the business platform.

[0088] Here, after obtaining the processing results, the target plug-in can convert the format of the processing results. For example, the processing results can be converted into a visual format. Through the visual format, the processing results can be presented to the user in a more intuitive and understandable manner. The visual format can be a report, a chart, or a text summary.

[0089] After obtaining the processing results, the front-end interactive interface can pop up the AI interactive window corresponding to the target plug-in on the front-end interactive interface, and the processing results after visualization can be displayed through the AI interactive window.

[0090] In the embodiment of the present application, in addition to obtaining the problem description information input by the user in the front-end interactive interface, the business processing purpose for the target business input by the user can also be obtained. Figure 2 The trigger operation of the "user input area" shown in the figure is used to determine the business processing objectives of the target business based on the trigger operation. The business processing objectives include but are not limited to: event analysis, event classification, automatic allocation, intelligent reminder, statistical reporting, and intelligent troubleshooting.

[0091] After obtaining the problem description information, business details data and business processing purpose of the target business, the target plug-in can transmit the problem description information, business details data and business processing purpose to the business big language model. Afterwards, the business big language model can process the problem description information and business details data according to the business processing purpose to obtain the processing result.

[0092] In an embodiment of the present application, first, the problem description information of the target business and the business detail data of the target business input by the user through the front-end interactive interface of the business platform are obtained; secondly, the problem description information and the business detail data are sent to the business big language model matching the target business for processing; wherein, the business big language model and the business platform are set on different devices; finally, the processing result of the business big language model is obtained, and the processing result is forwarded to the business platform; wherein, the processing result is used to indicate the answer result of the business big language model to the problem description information.

[0093] In the above implementation, after obtaining the problem description information of the target business entered by the user through the front-end interactive interface of the business platform and obtaining the business details data, the problem description information and business details data are forwarded to the business big language model for processing. This can achieve decoupling between the business platform and the business big language model. At this time, the business big language model can run independently from the business platform and is not dependent on the business platform. At this time, when the business platform accesses the business big language model for text processing, there is no need to change the relevant business logic of the business platform, thereby making business processing more flexible and concise, while improving time efficiency.

[0094] In an optional embodiment, after forwarding the processing result to the service platform, the following steps are further included:

[0095] First, in response to a user triggering an interactive window in a front-end interactive interface, first filling information for a to-be-filled area in the interactive window is determined; wherein the interactive window is a window for displaying processing results;

[0096] Secondly, the first filling information is sent to the front-end interactive interface, so that the front-end interactive interface fills the first filling information into the to-be-filled area and displays it.

[0097] In the embodiments of the present application, Figure 2 As shown, the interactive window includes a business interactive window and an AI interactive window. The target plug-in monitors DOM events to determine the first filling information of the area to be filled in the business interactive window.

[0098] Here, the processing result can be one or more. In the case of multiple processing results, the user can trigger the multiple processing results through the AI interaction window in the front-end interactive interface, thereby determining the first filling information of the area to be filled according to the processing result of the triggering operation.

[0099] For example, in response to a user clicking operation on the processing result 1, first filling information of the to-be-filled area in the business interaction window may be determined according to the processing result 1 clicked by the user.

[0100] Here, when the processing result is one, the first filling information of the area to be filled can be determined according to the processing result.

[0101] Here, after the first filling information is determined, the first filling information may be filled into the area to be filled and displayed.

[0102] In the above implementation, after forwarding the processing results to the business platform, the first filling information of the area to be filled is determined by capturing the user's trigger operation on the processing results in the AI interaction window and pushing it to the front-end interaction interface for display, thereby achieving real-time response to user operations and facilitating users to perform subsequent operations based on the displayed information.

[0103] In an optional embodiment, the above step determines the first filling information of the area to be filled in the interactive window in response to the user triggering the interactive window in the front-end interactive interface. Exemplarily, any one of the following embodiments may be adopted:

[0104] Implementation method one:

[0105] After detecting a click operation of the user on the target processing result in the interactive window, the first filling information is generated based on result description information of the target processing result.

[0106] In an embodiment of the present application, if the triggering operation for a processing result is a click operation, the processing result corresponding to the click operation can be determined, and the processing result clicked by the user can be determined as the target processing result. Subsequently, the first fill information for the to-be-filled area can be determined based on the result description information of the target processing result. The result description information includes at least one of the following: key information corresponding to the problem description information and the business processing purpose.

[0107] For example, the problem description information of the target business is "I went to beauty salon XXX after being introduced by a friend, but no invoice was provided, there was no effect, and the charges were unreasonable." In this case, the key information can be "consumer rights protection, medical beauty consumer rights protection, and food and drug supervision."

[0108] If the result type selected by the user is event classification, the result description information of the business large language model includes determining the solution department to which the problem description information belongs.

[0109] Reference Figure 2 , which is a schematic diagram of a front-end interactive interface of a business processing method provided in an embodiment of the present application, wherein:

[0110] Business details data include: incident number, reporter, contact information, incident source, region, location, time of incident, urgency, reporting time, item catalog, item name, appeal type, reporting time, whether it is difficult and severity.

[0111] The front-end interactive interface includes a business area and an interaction window. The business area includes a question input area and a business operation area. The interaction window includes a business interaction window and an AI interaction window. The AI interaction window is equipped with a processing result display area and a user input area. The processing result display area is used to display processing results, and the user input area can receive user input, such as the fill suggestion information in the following embodiment.

[0112] During specific implementation, the target plug-in can obtain the problem description information entered by the user in the problem input area. At the same time, the target plug-in can also obtain the element information of each UI element in the business interaction window. For example, the target plug-in can obtain the element information of the UI elements that have been filled in the business interaction window. Afterwards, the target plug-in can also obtain the business processing purpose entered by the user in the user input area. Next, the target plug-in can transmit the element information, problem description information and business processing purpose to the business big language model; then, the business big language model can process the problem description information and business details data according to the business processing purpose to obtain the processing result, and display the processing result in the processing result display. Next, the target plug-in can detect the user's triggering operation on the processing result displayed in the processing result display area, and then generate the first filling information according to the result description information of the triggered processing result, and fill the first filling information into the area to be filled, for example, Figure 2 As shown, the area to be filled can be the element display area corresponding to the "item name". In this case, the element information corresponding to the "item name" can be displayed in the corresponding element display area. Finally, the user's click operation on the business operation area can be detected, and the target business can be processed according to the content clicked by the user, where the click operation includes: acceptance, return, pending verification, etc.

[0113] Implementation method 2:

[0114] First, receiving the filling suggestion information input by the user in the interactive window;

[0115] Secondly, first filling information of the area to be filled is generated based on the filling suggestion information.

[0116] In this embodiment of the present application, after the processing result is displayed in the processing result display area, the user can click on the target processing result in the manner described in the first embodiment above, thereby generating the first fill-in information based on the result description information of the target processing result. In addition, the user can also enter fill-in suggestion information in the user input area of the AI interaction window.

[0117] After detecting the filling suggestion information input by the user in the user input area, the target plug-in can generate first filling information for the area to be filled based on the filling suggestion information. The filling suggestion information can be a secondary question about the processing result or a summary of the processing result.

[0118] After obtaining the filling suggestion information, the target plug-in may determine a processing result that matches the filling suggestion information from among multiple processing results according to the filling suggestion information.

[0119] In a specific implementation, the suggested keywords in the filling suggestion information can be extracted, and then the suggested keywords are matched with the result keywords of the processing result, and the first filling information is generated based on the result description information of the matching processing result. For example, the similarity between the suggested keywords and the result keywords of the processing result can be calculated, and the processing result with a similarity greater than a threshold is determined as a matching processing result.

[0120] In addition, the filling suggestion information can also be matched with the preset suggestion template. If the corresponding preset suggestion template is matched, the result is processed in the manner corresponding to the preset suggestion template to generate the first suggestion information. For example, if the filling suggestion information is the preset suggestion template "Please generate processing opinions", at this time, you can select the processing result with higher confidence, and generate the first filling information based on the result description information of the processing result with higher confidence.

[0121] In addition, when the result type selected by the user is event classification and the filling suggestion information is "Please generate processing opinions", the target plug-in can also forward the processing results and filling suggestion information to the business large language model for processing, and determine the current processing result of the business large language model as the first filling information.

[0122] Reference Figure 3 , which is a schematic diagram of another front-end interactive interface of the business processing method provided in an embodiment of the present application, wherein:

[0123] In the above Figure 2Based on the AI interaction window, the AI interaction window also includes a dialogue area, where the text generated by the user input area and the response to the text can be displayed. For example, if the text generated in the user input area is "Please generate a processing suggestion and fill it into the processing suggestion field", after the target plug-in fills the processing suggestion into the to-be-filled field, the response will be "The processing suggestion has been filled in for you, please check and confirm."

[0124] In a specific implementation, the target plug-in can obtain the question description information entered by the user in the question input area. Simultaneously, the target plug-in can also obtain element information for each UI element in the business interaction window. For example, the target plug-in can obtain element information for the UI elements already filled in the business interaction window. The target plug-in can then obtain the business processing purpose entered by the user in the user input area. The target plug-in can then transmit this element information, the question description information, and the business processing purpose to the business large language model. The business large language model can then process the question description information and business details data according to the business processing purpose, thereby obtaining a processing result and displaying this processing result in the processing result display. Next, the target plug-in can detect the request information entered by the user in the user input area and perform corresponding operations based on the request information, namely, generating first fill-in information and filling the first fill-in information into the to-be-filled area. Then, the target plug-in can respond to the request information in the conversation area based on a preset reply template. Finally, the target plug-in can detect the user's click operation in the business operation area and process the target business according to the user's click. These click operations include accept, return, and pending verification.

[0125] In an optional implementation, the above steps of forwarding the processing result to the business platform specifically include the following steps:

[0126] First, when there are multiple processing results, the confidence level of each processing result is determined; wherein the confidence level is used to indicate the credibility of each processing result;

[0127] Secondly, forward each processing result and the confidence level of each processing result to the business platform.

[0128] In the embodiments of the present application, each processing result can be analyzed from multiple dimensions. For example, the quality and richness of the data relied upon by the business language model when generating the processing result can be analyzed. The more authoritative the data source and the more abundant the data volume, the higher the confidence level of the corresponding processing result.

[0129] At the same time, based on historical processing results, it can also be determined whether the business language model's processing results for similar problem description information have been verified to be accurate and effective. In the case where the processing results for similar problem description information have been verified to be accurate and effective, the corresponding processing result has a higher confidence level.

[0130] After obtaining the confidence level of each processing result, the target plug-in may send each processing result and its confidence level to the business platform, and then each processing result and its confidence level may be displayed in the front-end interactive interface of the business platform.

[0131] When the number of processing results is large, a target number of processing results may be selected in descending order of confidence, and the target number of processing results and their confidence levels may be sent to the business platform.

[0132] In the above implementation, by determining the confidence level of the processing result and displaying the confidence level to the user, a trigger basis can be provided for the user to trigger the operation of the processing result, thereby guiding the user to select a more matching processing result and improving the accuracy and reliability of the business processing process.

[0133] In an optional embodiment, after forwarding the processing result to the service platform, the following steps are further included:

[0134] In the case where no user triggering operation on the processing result in the interactive window is detected, determining the processing result with the highest confidence;

[0135] Based on the result description information of the processing result with the highest confidence, the second filling information is determined, and the second filling information is sent to the front-end interactive interface for display.

[0136] In an embodiment of the present application, if the target plug-in does not detect a user triggering operation on the processing result in the interactive window within a preset time, for example, if the target plug-in does not detect any user triggering operation, such as a click or long press, on the processing result displayed in the AI interactive window within a preset time period, the second fill-in information can be determined based on the result description information of the processing result with the highest confidence. The second fill-in information can then be filled into the to-be-filled area of the front-end interactive interface.

[0137] In the above implementation, automatic filling and visual display of information can be achieved, further improving the user interaction experience and reducing user operation costs.

[0138] In an optional embodiment, obtaining problem description information of a target business input by a user through a front-end interactive interface of a business platform specifically includes the following steps:

[0139] First, the user's input operation in the question input area of the front-end display interface is detected;

[0140] Secondly, based on the input content, determine the problem description information of the target business.

[0141] In an embodiment of the present application, there may be one or more question input areas. In addition, the user may also enter question description information in the "user input area" in the AI interaction window.

[0142] Here, the question input area can be set as a text box or a voice input field, supporting users to initiate input in various forms such as text and voice.

[0143] The target plug-in can capture the user's input operation and parse the input content to obtain the problem description information of the target business; wherein the input operation can be a keyboard tapping action, voice input, etc.

[0144] For example, when the user's input operation is voice input, the voice information may be converted into text, and semantic extraction may be performed on the converted text information to obtain question description information.

[0145] In the above implementation, the problem description information of the target business can be determined by detecting the user operation in the question input area of the front-end interactive interface and analyzing the input content, thereby providing a basis for business processing and ensuring the development of subsequent business.

[0146] In an optional embodiment, after forwarding the processing result to the service platform, the following steps are further included:

[0147] First, target historical operation data similar to the problem description information of the target business is determined from the historical operation data of each business on the business platform;

[0148] Secondly, determine the historical processing opinions of the target historical operation data and forward them to the business platform;

[0149] Finally, based on the user's triggering operation on the historical processing opinions and / or processing results, the third filling information is determined and sent to the front-end interactive interface for display.

[0150] In an embodiment of the present application, the target plug-in may determine, in the summary of the historical operation data, target historical operation data whose similarity to the problem description information of the target business is greater than a preset similarity threshold.

[0151] Afterwards, the historical processing opinions can be determined in the target historical operation data and displayed in the AI interaction window.

[0152] After displaying the historical processing opinions, if a triggering operation of the user on the historical processing opinions and / or processing results in the AI interaction window is detected, the historical processing opinions and / or processing results corresponding to the triggering operation are determined to be the third filling information.

[0153] In the above implementation, historical operation data can be fully utilized to provide more targeted processing opinions and suggestions for the problem description information, thereby improving the efficiency and quality of target business processing.

[0154] In an optional implementation, the problem description information and business detail data are sent to a business language model that matches the target business for processing, specifically including the following steps:

[0155] First, determine the target description data related to the problem description information in the business detail data;

[0156] Secondly, the problem description information and target description data are integrated to obtain the target business data;

[0157] Finally, the target business data is input into the business large language model for processing to obtain the processing result of the large language model on the target business data.

[0158] In an embodiment of the present disclosure, first, the problem type of the problem description information can be determined. Second, based on the problem type, target description data similar to the problem type can be determined in the business detail data.

[0159] Here, data integration technology can be used to merge the problem description information and the target description data according to preset rules. For example, the preset rule may be to use the problem description information as the core content and add relevant supplementary information or explanatory information in the target description data in a structured or semi-structured form to the corresponding position of the problem description information.

[0160] After the target business data is obtained, the target business data can be used as input to the business language model.

[0161] In the above implementation, by determining the target description data in the business details data, fusing it with the problem description information to form the target business data, and inputting it into the business large language model for processing, the data processing scope can be effectively narrowed and the information processing efficiency can be improved.

[0162] In an optional embodiment, the business language model is trained in the following manner:

[0163] First, obtain the historical business data of the specified business module in the business platform and the general data of the business language model;

[0164] Secondly, the historical business data is split to obtain a plurality of first sub-data, and the general data is split to obtain a plurality of second sub-data;

[0165] Finally, the large language model to be trained is trained alternately by using the multiple first sub-data and the multiple second sub-data to obtain a large business language model.

[0166] In the embodiments of this application, general data refers to datasets for large model training. Public datasets can be used without special processing, such as the Belle dataset, the Chinese machine reading comprehension dataset, and the task-oriented dialogue dataset.

[0167] If the business platform is a grid-based government affairs business platform, initial business data of the grid-based government affairs business platform can be determined. The initial business data includes at least one of the following: departmental organization, personnel position, grid division, matter classification, command manual, work process, laws and regulations, and historical events.

[0168] After the initial business data is determined, the initial business data can be classified. Initial business data that does not change frequently and does not involve private information can be determined as historical business data.

[0169] For example, the departmental organizations, personnel positions and grid divisions in the initial business data of the grid-based government affairs business platform meet the above conditions. Therefore, matter classification, command manuals, work process laws and regulations, and historical events can be determined as historical business data of the grid-based government affairs business platform.

[0170] After the historical business data and the general data are determined, data processing can be performed on the historical business data and the general data, wherein the data processing includes at least one of the following: quality filtering, sensitive content filtering, and data deduplication.

[0171] When the amount of historical business data and general data is less than the preset data volume threshold, the amount of historical business data and general data can be increased through data enhancement methods so that the ratio of historical business data to general data remains between 1:10 and 1:5.

[0172] Here, the processed historical business data and the processed general data can be split separately to obtain the same number of first sub-data and second sub-data. For example, the historical business data can be divided into N equal parts to obtain N first sub-data with the same data volume; the general data can be divided into N equal parts to obtain N second sub-data with the same data volume.

[0173] Afterwards, the large language model to be trained can be trained alternately using the N first sub-data and the N second sub-data to obtain the service large language model. That is, the large language model to be trained is first trained using one first sub-data, and then the large language model to be trained is trained using one second sub-data, and this process is repeated until the large language model to be trained is trained using both the N first sub-data and the N second sub-data, thereby obtaining the service large language model.

[0174] Reference Figure 4As shown, it is a flowchart of the business large language model training of the business processing method provided in an embodiment of the present application, wherein:

[0175] S1. Determine general data and initial business data.

[0176] S2. Classify the initial business data to obtain historical business data.

[0177] S3. Process the general data and historical business data to obtain processed general data and processed historical business data.

[0178] S4. Split the processed general data and the processed historical business data respectively to obtain a plurality of first sub-data and a plurality of second sub-data.

[0179] S5. Train the large language model to be trained by alternately using the multiple first sub-data and the multiple second sub-data to obtain a large language model for the business to be evaluated.

[0180] S6. The business large language model to be evaluated is tested. If the test result meets the test conditions, the business large language model to be evaluated is determined as the business large language model.

[0181] Here, the dimensions of the inspection include at least one of the following: functional testing, performance evaluation, robustness testing, security evaluation, etc. If the inspection result does not meet the inspection conditions, the large language model of the business to be evaluated is retrained.

[0182] In the above implementation, by obtaining historical business data and general data of the business platform, splitting them into first sub-data and second sub-data respectively, and then alternately training the large language model to be trained, a business large language model is obtained, so that the business large language model has both specific business domain knowledge and general language capabilities, thereby improving the applicability and accuracy in the business scenarios corresponding to the business platform.

[0183] Reference Figure 5 , which is a workflow diagram of the target plug-in of the business processing method provided in an embodiment of the present application, wherein:

[0184] S10 . In response to changes in the question input area in the front-end interactive interface, obtain question description information of the target business and business detail data of the target business.

[0185] Here, the target plug-in can monitor whether the front-end interactive interface changes, and in the case where the front-end interactive interface changes, determine the business detail data displayed in the front-end interactive interface.

[0186] Here, it can be determined that the front-end interactive interface has changed when the user inputs problem description information of the target business through the front-end interactive interface of the business platform.

[0187] S20: Call the interface of the business language model, and send the business details data and problem description information to the business language model through the interface.

[0188] S30: Receive the processing result of the business large language model through the interface of the business large language model.

[0189] S40. In response to the target processing result selected by the user, the target processing result is displayed on the front-end interactive interface.

[0190] Here, in the case where there are multiple processing results, a DOM operation is performed in response to the target processing result clicked by the user, and the target processing result is displayed on the front-end interactive interface.

[0191] In the above implementation, after obtaining the problem description information of the target business entered by the user through the front-end interactive interface of the business platform and obtaining the business details data, the problem description information and business details data are forwarded to the business big language model for processing. This can achieve decoupling between the business platform and the business big language model. At this time, the business big language model can run independently from the business platform and is not dependent on the business platform. At this time, when the business platform accesses the business big language model for text processing, there is no need to change the relevant business logic of the business platform, thereby making business processing more flexible and concise, while improving time efficiency.

[0192] Those skilled in the art will understand that in the above-mentioned method of the specific implementation method, the writing order of each step does not mean a strict execution order and does not constitute any limitation on the implementation process. The specific execution order of each step should be determined by its function and possible internal logic.

[0193] Based on the same inventive concept, a business processing system corresponding to the business processing method is also provided in the embodiment of the present application. Since the principle of solving the problem by the system in the embodiment of the present application is similar to the above-mentioned business processing method in the embodiment of the present application, the implementation of the system can refer to the implementation of the method, and the repeated parts will not be repeated.

[0194] Reference Figure 6 , which is a schematic diagram of a business processing system provided by an embodiment of the present application, including: a business platform 61, a target plug-in 62, and a business language model 63;

[0195] The business platform 61 generates problem description information of the target business on the front-end display interface when the user inputs an operation through the problem input area on the front-end display interface.

[0196] The target plug-in 62 obtains the problem description information and the business detail data of the target business from the business platform, and forwards the problem description information and the business detail data of the target business to the business large language model.

[0197] The business language model 63 processes the problem description information and business detail data to obtain a processing result, and displays the processing result in an interactive window in the front-end display interface through the target plug-in.

[0198] Based on the same inventive concept, a business processing device corresponding to the business processing method is also provided in the embodiment of the present application. Since the principle of solving the problem by the device in the embodiment of the present application is similar to the above-mentioned business processing method in the embodiment of the present application, the implementation of the device can refer to the implementation of the method, and the repeated parts will not be repeated.

[0199] Reference Figure 7 FIG. 1 is a schematic diagram of a service processing device provided in an embodiment of the present application, wherein the device includes: an acquisition module 11, a processing module 12, and a forwarding module 13; wherein,

[0200] The acquisition module 11 is used to acquire the problem description information of the target business input by the user through the front-end interactive interface of the business platform, and acquire the business details data of the target business;

[0201] A processing module 12 is configured to send the problem description information and the business detail data to a business language model that matches the target business for processing;

[0202] The forwarding module 13 is used to obtain the processing result of the business large language model and forward the processing result to the business platform; wherein the processing result is used to indicate the answer result of the business large language model to the question description information.

[0203] The embodiment of the present application obtains the problem description information of the target business input by the user through the front-end interactive interface of the business platform, and obtains the business detail data, and then forwards the problem description information and business detail data to the business large language model for processing. This can achieve decoupling between the business platform and the business large language model. At this time, the business large language model can be separated from the business platform and run independently without relying on the business platform. At this time, when the business platform accesses the business large language model for text processing, there is no need to change the relevant business logic of the business platform, thereby making business processing more flexible and concise, while improving time efficiency.

[0204] The forwarding module is further configured to determine first filling information for the area to be filled in the interactive window in the front-end interactive interface in response to the user triggering the interactive window in the front-end interactive interface; wherein the interactive window is a window for displaying the processing result;

[0205] The first filling information is sent to the front-end interactive interface for display.

[0206] The forwarding module is further configured to generate the first filling information based on the result description information of the target processing result after detecting a click operation of the user on the target processing result in the interactive window.

[0207] The forwarding module is further configured to receive the filling suggestion information input by the user in the interactive window;

[0208] First filling information of the area to be filled is generated based on the filling suggestion information.

[0209] a forwarding module, specifically configured to determine the confidence level of each processing result when there are multiple processing results; wherein the confidence level is used to indicate the degree of credibility of each processing result;

[0210] Forwarding each of the processing results and the confidence level of each of the processing results to the service platform.

[0211] a forwarding module, specifically configured to determine the processing result with the highest confidence level when no triggering operation of the user on the processing result in the interactive window is detected;

[0212] Based on the result description information of the processing result with the highest confidence, second filling information is determined, and the second filling information is sent to the front-end interactive interface for display.

[0213] The acquisition module is further configured to detect an input operation of the user in the question input area of the front-end display interface;

[0214] Based on the input content, problem description information of the target business is determined.

[0215] For descriptions of the processing flow of each module in the device and the interaction flow between each module, reference can be made to the relevant descriptions in the above method embodiment, which will not be described in detail here.

[0216] Corresponding to Figure 1 The embodiment of the present application further provides an electronic device 800, such as Figure 8 , which is a schematic structural diagram of an electronic device 800 provided in an embodiment of the present application, wherein the electronic device may be a computer device (i.e., a personal computer (PC), a laptop computer, a server), a communication device (i.e., a mobile phone), and a wearable electronic device (i.e., a smart watch, a smart bracelet, a head-mounted display); the electronic device 800 includes:

[0217] Processor 81, memory 82; memory 82 is used to store execution instructions, including internal memory 821 and external memory 822; the memory 821 here is also called internal memory, which is used to temporarily store the operation data in the processor 81 and the data exchanged with the external memory 822 such as a hard disk. The processor 81 exchanges data with the external memory 822 through the memory 821. When the electronic device 800 is running, the processor 81 communicates with the memory 82, so that the processor 81 executes the following instructions:

[0218] Obtaining problem description information of the target business entered by the user through the front-end interactive interface of the business platform, and obtaining business detail data of the target business;

[0219] Sending the problem description information and the business detail data to a business language model that matches the target business for processing;

[0220] Obtaining a processing result of the business large language model and forwarding the processing result to the business platform; wherein the processing result is used to indicate an answer result of the business large language model to the question description information.

[0221] The present application also provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the computer program executes the steps of the business processing method described in the above method embodiment. The storage medium can be a volatile or non-volatile computer-readable storage medium.

[0222] An embodiment of the present application also provides a computer program product, which carries program code. The instructions included in the program code can be used to execute the steps of the business processing method described in the above method embodiment. For details, please refer to the above method embodiment and will not be repeated here.

[0223] The computer program product may be implemented in hardware, software, or a combination thereof. In one embodiment, the computer program product is implemented as a computer storage medium. In another embodiment, the computer program product is implemented as a software product, such as a software development kit (SDK).

[0224] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the system and device described above can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here. In the several embodiments provided in this application, it should be understood that the disclosed system, device and method can be implemented in other ways. The device embodiments described above are merely 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 communication interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0225] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0226] 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.

[0227] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a non-volatile computer-readable storage medium that is executable by a processor. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0228] Finally, it should be noted that the above-described embodiments are only specific implementation methods of the present application, which are used to illustrate the technical solutions of the present application, rather than to limit them. The scope of protection of the present application is not limited thereto. Although the present application has been described in detail with reference to the above-described embodiments, those skilled in the art should understand that any person skilled in the art can modify or easily conceive of changes to the technical solutions described in the above-described embodiments within the technical scope disclosed in the present application, or perform equivalent replacements for some of the technical features thereof. These modifications, changes, or replacements do not deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the scope of protection of the present application. Therefore, the scope of protection of the present application shall be subject to the scope of protection of the claims.

Claims

1. A business processing method, characterized in that: include: Obtaining problem description information of the target business entered by the user through the front-end interactive interface of the business platform, and obtaining business detail data of the target business; Sending the problem description information and the business detail data to a business language model that matches the target business for processing; Obtaining a processing result of the business large language model and forwarding the processing result to the business platform; wherein the processing result is used to indicate an answer result of the business large language model to the question description information.

2. The method according to claim 1, characterized in that After forwarding the processing result to the service platform, the method further includes: In response to the user triggering an interactive window in the front-end interactive interface, determining first filling information for an area to be filled in the interactive window; wherein the interactive window is a window for displaying the processing result; The first filling information is sent to the front-end interactive interface, so that the front-end interactive interface fills the first filling information into the to-be-filled area and displays it.

3. The method according to claim 2, characterized in that The determining, in response to the user triggering the interactive window in the front-end interactive interface, first filling information of the area to be filled in the interactive window includes: After detecting a click operation of the user on the target processing result in the interactive window, the first filling information is generated based on result description information of the target processing result.

4. The method according to claim 2, characterized in that The determining, in response to the user triggering the interactive window in the front-end interactive interface, first filling information of the area to be filled in the interactive window includes: receiving filling suggestion information input by the user in the interactive window; First filling information of the area to be filled is generated based on the filling suggestion information.

5. The method according to claim 1, characterized in that The forwarding the processing result to the service platform includes: In the case where there are multiple processing results, determining a confidence level for each processing result; wherein the confidence level is used to indicate a degree of credibility of each processing result; Forwarding each of the processing results and the confidence level of each of the processing results to the service platform.

6. The method according to claim 5, characterized in that After forwarding the processing result to the service platform, the method further includes: Determining the processing result with the highest confidence level when no triggering operation of the user on the processing result in the interactive window is detected; wherein the interactive window is a window for displaying the processing result; Based on the result description information of the processing result with the highest confidence, second filling information is determined, and the second filling information is sent to the front-end interactive interface for display.

7. The method according to claim 1, characterized in that After forwarding the processing result to the service platform, the method further includes: Determining target historical operation data similar to the problem description information of the target business from the historical operation data of each business on the business platform; Determining a historical processing opinion for the target historical operation data, and forwarding the historical processing opinion to the business platform; Based on the user's triggering operation on the historical processing opinions and / or the processing results, third filling information is determined, and the third filling information is sent to the front-end interactive interface for display.

8. The method according to claim 1, characterized in that The sending of the problem description information and the business detail data to a business language model matching the target business for processing includes: Determining target description data related to the problem description information in the business detail data; Fusing the problem description information with the target description data to obtain target business data; The target business data is input into the business large language model for processing to obtain a processing result of the large language model on the target business data.

9. An electronic device, characterized in that: include: A processor and a memory, wherein the memory stores machine-readable instructions executable by the processor. When the electronic device is running, the processor communicates with the memory, and when the machine-readable instructions are executed by the processor, the steps of the business processing method according to any one of claims 1 to 8 are performed.

10. A computer program product, characterized in that The computer program product is stored in a storage medium, and the program product is executed by at least one processor to implement the steps of the service processing method according to any one of claims 1 to 8.

Images