Information processing method, system and device, storage medium and program product

Abstract

The embodiment of the invention provides an information processing method, system and device, a storage medium and a program product. In the embodiment of the invention, aiming at a first to-be-processed problem, a processing flow is executed to obtain first processing information corresponding to the first to-be-processed problem, and the processing flow comprises the following steps: obtaining detail information of the to-be-processed problem according to a preset requirement; obtaining target knowledge serving as a processing basis of the to-be-processed problem from a knowledge base; based on the detailed information and the target knowledge, calling a preset model to generate processing information corresponding to the to-be-processed problem; judging whether the first processing information needs to be audited or not; when the first processing information does not need to be audited, processing the first to-be-processed problem based on the first processing information; and when the first processing information needs to be audited, sending the first processing information to a manual auditing client. According to the technical scheme, the labor cost can be reduced, and the processing efficiency can be improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computer, and particularly relates to an information processing method, system, device, storage medium and program product. BACKGROUND

[0002] In the field of electronic commerce (e-commerce for short), with the continuous expansion of online transaction scale, the number of after-sales problems (such as transaction dispute problems) also rapidly increases. At present, the processing of after-sales problems by e-commerce platforms mainly relies on manual methods, that is, customer service personnel obtain relevant materials from e-commerce platforms for manual auditing, judgment and processing.

[0003] However, due to the large number of after-sales problems at present, e-commerce platforms need to invest a large number of customer service personnel to process information, which significantly increases operating costs, and manual auditing needs to be analyzed case by case, and the processing period is relatively long, which is difficult to meet the needs of users for quickly solving after-sales problems, and affects user experience. SUMMARY

[0004] Aspects of the present application provide an information processing method, system, device, storage medium and program product to reduce labor costs and improve processing efficiency.

[0005] A first aspect of the present application provides an information processing method, comprising:

[0006] For a first to-be-processed problem, a processing flow is executed to obtain first processing information corresponding to the first to-be-processed problem, wherein the processing flow comprises: obtaining detail information of a to-be-processed problem according to a preset requirement; obtaining target knowledge from a knowledge base as a basis for processing the to-be-processed problem; and based on the detail information and the target knowledge, calling a preset model to generate processing information corresponding to the to-be-processed problem;

[0007] It is judged whether the first processing information needs to be audited;

[0008] When the first processing information does not need to be audited, the first to-be-processed problem is processed based on the first processing information;

[0009] When the first processing information needs to be audited, the first processing information is sent to a manual auditing client.

[0010] A second aspect of the present application provides an information processing method, comprising:

[0011] For the third to-be-processed problem, a processing flow is executed to obtain third processing information corresponding to the third to-be-processed problem, wherein the processing flow comprises: obtaining detail information of the to-be-processed problem according to a preset requirement; obtaining target knowledge serving as a basis for processing the to-be-processed problem from a knowledge base; and based on the detail information and the target knowledge, calling a preset model to generate processing information corresponding to the to-be-processed problem;

[0012] If the processing flow is in the first mode, the third to-be-processed problem is processed based on artificial processing information made by an artificial person on the third to-be-processed problem, wherein the third processing information is used in combination with the artificial processing information as a basis for adjusting the processing flow;

[0013] If the processing flow is in the second mode, the third to-be-processed problem is processed based on the third processing information.

[0014] The third aspect of the present application provides an information processing method, comprising:

[0015] According to a preset requirement, obtaining detail information of a to-be-processed problem;

[0016] Based on the detail information, obtaining target knowledge from a knowledge base, wherein the knowledge base is constructed based on a preset problem processing specification;

[0017] Based on the detail information and the target knowledge, using a preset model to generate processing information corresponding to the to-be-processed problem.

[0018] The fourth aspect of the present application provides an information processing method, comprising:

[0019] For the first e-commerce after-sales problem, a processing flow is executed to obtain first processing information corresponding to the first e-commerce after-sales problem, wherein the processing flow comprises: obtaining detail information of the e-commerce after-sales problem according to a preset requirement; obtaining target knowledge serving as a basis for processing the e-commerce after-sales problem from a knowledge base; and based on the detail information and the target knowledge, calling a preset model to generate processing information corresponding to the e-commerce after-sales problem;

[0020] Judging whether the first processing information needs to be audited;

[0021] When no audit is needed, processing the first to-be-processed problem based on the first processing information;

[0022] When audit is needed, sending the first processing information to an artificial audit client.

[0023] The fifth aspect of the present application provides an information processing method, comprising:

[0024] receive first processing information corresponding to the first to-be-processed question sent by the server, the first processing information being obtained by executing a processing flow, the processing flow comprising: obtaining detail information of the to-be-processed question according to a preset requirement; obtaining target knowledge serving as a basis for processing the to-be-processed question from a knowledge base; and based on the detail information and the target knowledge, calling a preset model to generate processing information corresponding to the to-be-processed question;

[0025] display the first processing information for manual auditing;

[0026] in response to a confirmation operation of a user on the first processing information, send confirmation information to the server, the confirmation information being used to instruct the server to process the first to-be-processed question based on the first processing information.

[0027] A sixth aspect of the present application provides an information processing system, comprising: a server and a manual auditing client; wherein

[0028] the server is configured to implement the method in the first aspect or the fourth aspect.

[0029] the manual auditing client is configured to: receive the first processing information sent by the server; and display the first processing information for manual auditing.

[0030] A seventh aspect of the present application provides an electronic device, comprising: a memory and a processor, wherein

[0031] the memory is configured to store a program;

[0032] the processor is coupled to the memory and is configured to execute the program stored in the memory, so as to implement the method in any one of the preceding aspects.

[0033] An eighth aspect of the present application provides a computer readable storage medium storing a computer program, the computer program being executable by a computer to implement the method in any one of the preceding aspects.

[0034] A ninth aspect of the present application provides a computer program product, comprising a computer program executable by a processor to implement the method in any one of the preceding aspects.

[0035] In the technical scheme provided by the embodiment of the application, the processing flow is used to process the to-be-processed problem based on the knowledge base and the preset model to obtain processing information, that is, the relevant knowledge is introduced into the preset model through the knowledge base, so that the training cost of the preset model can be reduced. In addition, after the processing information output by the processing flow is obtained, it is further judged whether the processing information needs to be audited by a human being, when the processing information needs to be audited by a human being, the processing information is sent to a human auditing client for human auditing, and when the processing information does not need to be audited, the to-be-processed problem is processed based on the processing information directly. It can be seen that, on the premise of automatically processing the to-be-processed problem by the processing flow to reduce the human processing cost and improve the information processing efficiency, the processing error rate can also be effectively reduced.

[0036] In the technical scheme provided by the embodiment of the application, when the processing flow is in the first mode, the to-be-processed problem is processed based on the human processing information, and the processing information output by the processing flow is used in combination with the human processing information as a basis for adjusting the processing flow. When the processing flow is in the second mode, the to-be-processed problem is processed based on the processing information output by the processing flow, that is, the processing information output by the processing flow is used to process the to-be-processed problem. That is, after the processing flow is put into operation / deployed, the optimal processing path is selected by judging the mode of the processing flow, and then a dynamic balance between efficiency and reliability is achieved. Moreover, when the processing flow is in the first mode, the basis for adjusting the processing flow can be generated based on the processing information output by the processing flow and the human processing information, so that the processing flow can be continuously optimized.

[0037] In the technical scheme provided by the embodiment of the application, the problem is automatically analyzed by the preset model and the processing information is generated, which can not only reduce the dependence on human beings and reduce the operation cost, but also shorten the processing period and improve the user satisfaction. In addition, the knowledge base can be introduced to provide the model with the problem processing specification required for problem processing, so that the training cost of the preset model can be reduced.

[0038] In the technical scheme provided by the embodiment of the application, the processing flow is used to process the to-be-processed problem based on the knowledge base and the preset model to obtain processing information, that is, the relevant knowledge is introduced into the preset model through the knowledge base, so that the training cost of the preset model can be reduced. In addition, after the processing information output by the processing flow is obtained, it is further judged whether the processing information needs to be audited by a human being, when the processing information needs to be audited by a human being, the processing information is sent to a human auditing client for human auditing, and when the processing information does not need to be audited, the to-be-processed problem is processed based on the processing information directly. It can be seen that, on the premise of automatically processing the to-be-processed problem by the processing flow to reduce the human processing cost and improve the information processing efficiency, the processing error rate can also be effectively reduced.

[0039] In the technical scheme provided by the embodiment of the present application, the processing flow is used to process the to-be-processed question based on the knowledge base and the preset model to obtain processing information, that is, the relevant knowledge is introduced into the preset model through the knowledge base, so that the training cost of the preset model can be reduced. In addition, after the processing information output by the processing flow is obtained, the processing information is sent to the artificial auditing client for artificial auditing. It can be seen that, on the premise of automatically processing the to-be-processed question by the processing flow to reduce the artificial processing cost and improve the information processing efficiency, the processing error rate can also be effectively reduced. BRIEF DESCRIPTION OF DRAWINGS

[0040] The accompanying drawings, which are included to provide a further understanding of the present application, constitute a part of the present application and illustrate the illustrative embodiments of the present application and their descriptions serve to explain the present application, but do not constitute improper limitations on the present application. In the drawings:

[0041] Figure 1 The structural schematic diagram of the information processing system provided by an exemplary embodiment of the present application is shown in the figure;

[0042] Figure 2 The flowchart of the information processing method provided by an exemplary embodiment of the present application is shown in the figure;

[0043] Figure 3 The to-be-audited interface provided by an exemplary embodiment of the present application is shown in the figure;

[0044] Figures 4-8 The flowchart of the information processing method provided by different exemplary embodiments of the present application is shown in the figure;

[0045] Figure 9 The structural schematic diagram of the electronic device provided by an exemplary embodiment of the present application is shown in the figure. DETAILED DESCRIPTION

[0046] In order to make the purpose, technical scheme and advantages of the present application clearer, the technical scheme of the present application will be described clearly and completely in combination with the specific embodiments of the present application and the corresponding drawings. Obviously, the described embodiments are only part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0047] It should be noted that in the case of the user information involved in the embodiments of the present application, the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the embodiments of the present application are all information and data authorized by the user or authorized by all parties, and the collection, use and processing of related data need to comply with relevant laws, regulations and standards of relevant countries and regions, and provide corresponding operation portal for user to choose authorization or refusal. In addition, the various models (including but not limited to language models or large models) involved in the present application are in line with relevant legal and standard regulations.

[0048] Before introducing the specific content of the embodiments of the present application, the technical terms mentioned in the present text are briefly explained.

[0049] An electronic commerce service provider, or an electronic commerce platform, can also be referred to as an e-commerce platform, is a platform that provides online transaction negotiation for enterprises or individuals. The electronic commerce service provider is a place that coordinates and integrates the orderly association and flow of information flow, goods flow and fund flow, that is, by establishing an electronic commerce portal, it provides online shopping services to ordinary consumers or provides business applications to enterprises to seek better integration between the Internet and traditional business and industrial systems. Enterprises, merchants and individuals can fully utilize the network infrastructure, payment platforms and other shared resources provided by the electronic commerce platform to carry out their own business activities. In this paper, the electronic commerce service provider can be a shop market e-commerce platform, a mall e-commerce platform or a social e-commerce platform. The shop market e-commerce platform and the mall e-commerce platform can be referred to as traditional electronic commerce. Social e-commerce is a new derivative mode of electronic commerce. It uses social websites, social APPs and other dissemination channels to assist the purchase and sale of goods through social interaction and user-generated content, and applies social elements such as attention, communication, discussion and interaction to the electronic commerce transaction process. Social e-commerce combines electronic commerce and social media.

[0050] Merchants are individuals and various organizations engaged in business activities (production, operation of related goods) through electronic commerce platforms.

[0051] E-commerce dispute: in the transaction process, the consumer causes the refund due to internal or external reasons, but the merchant refuses, and then intervenes in the platform processing process.

[0052] With the rapid development of artificial intelligence (AI) technology, its application prospects in various fields are increasingly broad. In the e-commerce field, the use of artificial intelligence is mostly focused on intelligent customer service, such as intelligent customer service replacing human customer service to have a conversation with consumers. However, the handling of after-sales problems (such as transaction disputes) in the e-commerce field mainly relies on human labor. However, there are problems such as high operating costs and low processing efficiency in processing e-commerce after-sales problems by human labor.

[0053] To solve the problems existing in the existing information processing scheme, the embodiments of the present application propose that the artificial intelligence technology is applied to information processing, which can not only reduce the labor cost, but also shorten the processing period. In addition, in order to avoid the problems such as high training cost, uncertainty of training effect, slow knowledge update caused by model fine-tuning, the knowledge base required for problem processing is introduced into the preset model to introduce the knowledge required for processing problems. That is, the preset model does not need to learn the corresponding knowledge through training. That is, the present scheme enters the Retrieval-Augmented Generation (RAG) technology in natural language processing to solve the existing technical problems.

[0054] In addition, the various models (such as the preset model, the contradiction identification model, and the text summarization model) involved in this paper can be the same language model (LM) or different language models, and the embodiments of the present application do not make specific limitations. The language model is a model based on artificial intelligence. The embodiments of the present application do not limit the number of model parameters supported by the language model, and the goal is to meet the application requirements. If the model parameters are relatively more, the size of the language model will be relatively larger, and the model performance will be relatively better, of course, more time and resources will be consumed in the inference and training process; if the model parameters are relatively less, the size of the language model will be relatively smaller, and in the case that the performance meets the requirements, the model is more lightweight, and less time and resources are consumed in the inference and training process. The language model is a deep learning model used to process and generate natural language text. The model is based on a neural network architecture and can be trained on a large amount of text data to capture the complex patterns and structures of language. In an optional embodiment, the language model includes an encoder, a decoder, a self-attention layer, and a feed-forward neural network. The encoder is used to understand the semantics of the input text.

[0055] Before introducing the information processing method provided by the embodiments of the present application, the system architecture involved in the method is introduced. Figure 1A structural schematic diagram of an information processing system provided by an embodiment of the present application is shown in FIG. 1. As shown in FIG. 1, the system includes a server 1 and an artificial auditing client 2. The server 1 can be a server, a service cluster, a virtual server, a cloud, etc. The client device running the artificial auditing client 2 can be, but is not limited to, a smartphone, a smart wearable device, a tablet computer, a notebook computer, a desktop computer, etc. Figure 1

[0056] The client can implement the steps in the method provided by the following partial method embodiments, and the server can implement the steps in the method provided by the following partial method embodiments, which will be described in detail below.

[0057] Optionally, the system can further include a plurality of participant clients (not shown).

[0058] The plurality of participant clients can be used to submit a to-be-processed problem and / or receive and display processing information sent by the server 1 after processing the to-be-processed problem.

[0059] It should be noted that each to-be-processed problem (e.g., a first to-be-processed problem, a second to-be-processed problem, and a third to-be-processed problem) in this document can exist in the form of a task or a work order. The to-be-processed problem can be submitted by any participant involved in the problem, or can be automatically generated by the server based on collected information. The source of the to-be-processed problem is not limited in the embodiments of the present application. Any to-be-processed problem in the present application can be an after-sales problem. For example, the after-sales problem can be a transaction dispute problem.

[0060] Figure 2 A flowchart of an information processing method provided by an embodiment of the present application is shown in FIG. 2. As shown in FIG. 2, the execution subject of the method provided by the embodiment can be the server 1. The method includes the following steps: Figure 2

[0061] 101. For a first to-be-processed problem, a processing flow is executed to obtain first processing information corresponding to the first to-be-processed problem.

[0062] 102. It is determined whether the first processing information needs to be audited.

[0063] 103. When the first processing information does not need to be audited, the first to-be-processed problem is processed based on the first processing information.

[0064] 104. When the first processing information needs to be audited, the first processing information is sent to an artificial auditing client.

[0065] ​​In step 101 above, the processing flow includes: obtaining detailed information about the problem to be processed according to preset requirements; obtaining target knowledge from the knowledge base as the basis for processing the problem to be processed; and, based on the detailed information and the target knowledge, invoking a preset model to generate processing information corresponding to the problem to be processed. In other words, when this processing flow is executed, the following steps can be achieved:

[0066] S1. Obtain detailed information of pending issues according to preset requirements.

[0067] S2. Obtain target knowledge from the knowledge base as the basis for processing the problem to be processed.

[0068] S3. Based on the detailed information and the target knowledge, call a preset model to generate processing information corresponding to the problem to be processed.

[0069] In the above step S1, the preset requirements may be used to specify the information items to be obtained and / or the acquisition method.

[0070] In some embodiments, the preset requirements may include preset information items to be obtained.

[0071] Taking a transaction dispute on an e-commerce platform as an example, information about both parties to the transaction, order information, refund information corresponding to the transaction order, dispute information corresponding to the transaction order, logistics information corresponding to the transaction order, attribute information of the goods involved in the transaction order, transaction message information, and chat logs between the two parties can be obtained. Dispute information refers to the information recorded in the dispute work ticket corresponding to the e-commerce after-sales issue. This information can include the membership attributes of both parties. Optionally, evidence submitted by both parties, such as photos, can also be obtained.

[0072] In the above step S2, in some embodiments, the knowledge base may store preset problem handling specifications and / or case facts and handling information of historical information handling cases.

[0073] Among them, problem handling standards may include but are not limited to: legal and regulatory standards, industry standards, and platform standards. Taking e-commerce after-sales issues as an example, problem handling standards may include: problem handling standards provided by the e-commerce platform (this standard is a platform standard).

[0074] Here, historical information processing cases can be cases that were handled correctly in the past. Alternatively, the knowledge base can store both historical cases of incorrect handling and historical cases of correct handling, with the incorrect cases used as negative examples and the correct cases used as positive examples. It is understood that these historical information processing cases can be manually processed or automatically processed using a processing flow.

[0075] In some optional embodiments, the target knowledge can be obtained from the knowledge base based on the detail information.

[0076] For example, the target knowledge that matches the problem to be processed can be screened from the knowledge base based on the detail information by using a retrieval technology. The target knowledge can include a target problem processing specification.

[0077] In the step S3, the information related to the problem to be processed and the target knowledge can be used as input parameters to call the preset model to output the processing information corresponding to the problem to be processed by the preset model.

[0078] In some optional embodiments, a prompt can be generated based on the detail information and the target knowledge, and the preset model can be called based on the prompt to output the processing information corresponding to the problem to be processed by the preset model under the prompting of the prompt. For example, a prompt template containing placeholders can be obtained, and the detail information and the target knowledge can be filled into the prompt template at positions corresponding to the placeholders to obtain the prompt. The use of the prompt template can not only improve the efficiency of constructing the prompt, but also improve the effect of constructing the prompt.

[0079] The preset model can analyze the detail information and the target knowledge by using its language understanding ability and reasoning ability to obtain the processing information corresponding to the problem to be processed.

[0080] In the step 101, the processing flow is executed for the first problem to be processed to obtain the first processing information corresponding to the first problem to be processed. That is, the processing flow processes the first problem to be processed according to the steps S1, S2 and S3 to obtain the first processing information corresponding to the first problem to be processed.

[0081] In the step 102, it is determined whether the first processing information needs to be audited. If the first processing information does not need to be audited, the step 103 is executed, and if the first processing information needs to be audited, the step 104 is executed.

[0082] In some embodiments, it can be determined whether the first processing information needs to be audited based on the first processing information and / or the detail information.

[0083] For example, when the amount involved in the processing information corresponding to the first problem to be processed exceeds a preset amount, it is determined that the processing information corresponding to the first problem to be processed needs to be audited. When the amount does not exceed the preset amount, it is determined that the processing information corresponding to the first problem to be processed does not need to be audited.

[0084] In actual applications, the size of the preset amount can be set according to actual needs, which is not limited in the embodiments of the present application.

[0085] Another example, the detail information includes a participant involved in the first pending issue. When the number of historical complaints of the participant involved in the first pending issue against issue processing exceeds a preset number, it is determined that the processing information corresponding to the first pending issue needs to be audited. When the number of complaints does not exceed the preset number, it is determined that the processing information corresponding to the first pending issue does not need to be audited.

[0086] In practical applications, the size of the preset number can be set according to actual needs, and the embodiments of the present application do not make specific limitations.

[0087] Another example, the detail information includes a participant involved in the first pending issue. When the number of historical complaints of the participant involved in the first pending issue against issue processing exceeds a preset number, it is determined that the processing information corresponding to the first pending issue needs to be audited. When the number of complaints does not exceed the preset number, it is determined that the processing information corresponding to the first pending issue does not need to be audited.

[0088] In practical applications, the difficult or easy processing error processing type can be recorded in advance to form a list. As long as the processing type involved in the pending issue hits the list, it is determined that it needs to be audited, otherwise it does not need to be audited.

[0089] In the above step 103, when the first processing information does not need to be audited, the pending issue can be processed directly based on the first processing information.

[0090] In the above step 104, when the first processing information needs to be audited, the first processing information is sent to the manual auditing client and waits for auditing.

[0091] After the manual auditing client receives the first processing information, the first processing information is displayed for manual auditing. For example, as shown in the interface 300, the interface 300 displays the judgment information (i.e. processing information) a31 for the judgment order A, and displays the "pass" control 32 and the "not pass" control 33 for the judgment information a31. In response to the triggering operation of the "pass" control 32, it is determined that the judgment information a31 is passed. In response to the "not pass" control 33, it is determined that the judgment information a31 is not passed, and a judgment information modification interface (not shown) is displayed, which displays the judgment information a for manual modification. Figure 3

[0092] ​It can be understood that the server can directly send the first processing information to the manual auditing client when it is judged that the first processing information needs to be audited. When the manual auditing client receives the first processing information, the first processing information is displayed. Alternatively, the server can store the first processing information when it is judged that the first processing information needs to be audited. When the server subsequently receives an acquisition request sent by the manual auditing client, the server sends the first processing information to the manual auditing client. When the manual auditing client receives the first processing information, the first processing information is displayed for manual auditing.

[0093] In the technical scheme provided by the embodiments of the present application, the processing flow is used to process the to-be-processed question based on the knowledge base and the preset model to obtain processing information, that is, the relevant knowledge is introduced into the preset model through the knowledge base, so that the training cost of the preset model can be reduced. In addition, after the processing information output by the processing flow is obtained, it is further judged whether the processing information needs to be manually audited. When the processing information needs to be manually audited, the processing information is sent to the manual auditing client for manual auditing. When the processing information does not need to be audited, the to-be-processed question is directly processed based on the processing information. It can be seen that, on the premise of automatically processing the to-be-processed question by the processing flow to reduce the manual processing cost and improve the information processing efficiency, the processing error rate can also be effectively reduced.

[0094] The detail information obtained in each of the above embodiments is complex and miscellaneous. Before being input into the preset model, the contradictions between the multiple participants involved in the to-be-processed question can be identified by using the contradiction identification model, and then the contradictions are input into the preset model, so that the output accuracy of the preset model can be improved.

[0095] Therefore, in some embodiments, the above processing flow can further include:

[0096] S4, calling a contradiction identification model according to the detail information to obtain contradictions between multiple participants involved in the to-be-processed question.

[0097] In the above S3, "calling a preset model to generate processing information corresponding to the to-be-processed question based on the detail information and the target knowledge", can include:

[0098] calling a preset model to generate processing information corresponding to the to-be-processed question based on the detail information, the contradictions and the target knowledge.

[0099] In the above S4, the detail information can be used as a parameter to call the contradiction identification model to output the contradictions between the multiple participants involved in the first to-be-processed question by the contradiction identification model. The contradictions between the multiple participants can include the dispute points and / or demands of the multiple participants.

[0100] In some optional embodiments, a prompt can be generated according to the detail information, and a contradiction recognition model can be invoked based on the prompt to output contradictions between the multiple parties under the prompting of the prompt. For example, a prompt template can be obtained, the prompt template including placeholders, the detail information being filled into the prompt template at positions of the corresponding placeholders to obtain the prompt. The use of the prompt template can not only improve the efficiency of constructing the prompt, but also improve the effect of constructing the prompt.

[0101] The contradiction recognition model analyzes the detail information by using its language understanding capability and reasoning capability to obtain the contradictions between the multiple parties.

[0102] In this embodiment, the contradiction recognition model is used to identify the controversial points and demands of the multiple parties involved in the question before the question is processed by the preset model, and the identified controversial points and demands are provided to the preset model, which not only helps to reduce the processing difficulty of the preset model, but also improves the processing accuracy of the preset model.

[0103] In other embodiments, the step S1 of obtaining the detail information of the question to be processed according to the preset requirement can be implemented by the following steps:

[0104] S13, obtaining initial detail information of the question to be processed according to the preset requirement.

[0105] S14, performing text summarization on the initial detail information by using a text summarization model to obtain the detail information.

[0106] The detail information includes the summary text output by the text summarization model.

[0107] Generally, the initial detail information includes multiple detail information items, and the text summarization model can be used to summarize the information items with more text content to obtain the summary text corresponding to the information items. For example, in an e-commerce platform, the detail information items of a transaction dispute include order information, message information and chat record information. The text content of the message information and the chat record information is more, and therefore, the text summarization model can be used to summarize the message information and the chat record information to obtain the summary text of the message information and the summary text of the chat record information.

[0108] Optionally, the detail information can include the initial detail information. That is, the initial detail information before text summarization is also input to the preset model to avoid loss of detail information.

[0109] The second information item can be used as a parameter to invoke the text summarization model to obtain the summary text of the second information item.

[0110] The text summarization model uses its language understanding capability to perform text summarization on the initial detail information.

[0111] In this embodiment, before calling the contradiction identification model or the preset model, the text summarization model is used to summarize and extract information with more text content, which can improve the accuracy of the model.

[0112] This automated problem solving solution can greatly improve efficiency, reduce human error, and provide more transparent and fair services to users.

[0113] In some embodiments, the processing flow executed in step 101 can be an adjusted processing flow, i.e., the processing flow executed in step 101 is a processing flow obtained by adjusting the processing flow to be adjusted.

[0114] To avoid the high training cost brought by model fine-tuning, the above adjustment can include adjustment of the input of the model called by the processing flow, and does not include adjustment of the structure and / or parameters inside the model (e.g., preset model, text summarization model, contradiction identification model) called by the processing flow. The structure refers to the network structure, and the parameters refer to the network parameters. For example, the adjustment of the input of the model called by the processing flow can include adjustment of the type of input parameters and / or adjustment of the prompt word template. The input parameters are information items related to the first problem to be processed that need to be input to the model.

[0115] Of course, in actual application, if cost is not considered, the above adjustment can also include adjustment of the structure and / or parameters inside the model called by the processing flow.

[0116] The adjustment basis of the processing flow can be obtained by the following steps:

[0117] 105. For the second problem to be processed, execute the processing flow to be adjusted to obtain second processing information corresponding to the second problem to be processed.

[0118] 106. Obtain artificial processing information corresponding to the second problem to be processed.

[0119] 107. Based on the second processing information and the artificial processing information, generate an adjustment basis for adjusting the processing flow.

[0120] In step 105, the processing flow to be adjusted can be used to execute steps S1-S3 to obtain second processing information corresponding to the second problem to be processed.

[0121] In step 106, the artificial processing information refers to the processing information made by the information processing personnel for the second problem to be processed.

[0122] In an alternative implementation of 107, when the second processing information and the manual processing information are different, information provided by the information processing personnel in the process of processing the second to-be-processed problem can be obtained, for example, information items related to the second to-be-processed problem, manually summarized text summaries, and manually extracted contradictions between multiple parties involved in the second to-be-processed problem. The second processing information, the manual processing information, and the information provided by the information processing personnel in the process of processing the second to-be-processed problem can be used to construct a basis for adjusting the processing flow. The program optimization personnel can adjust and optimize the processing flow based on the basis, for example, optimize the input of the called model, thereby improving the output accuracy of the model.

[0123] Of course, in addition to manually optimizing the processing flow, the processing flow can also be automatically adjusted or optimized based on the reference information.

[0124] The technical solution provided by the embodiments of the present application innovatively realizes the landing of the service scenario based on the existing model directly by the engineering side personnel without the algorithm personnel. That is, the existing model can be completely used, and the time-consuming and laborious work can be simplified without model fine-tuning, thereby greatly reducing the cost and improving the service efficiency. In addition, with the expansion of the application scenario, only the knowledge (for example, problem processing specification) in the knowledge base needs to be adjusted as needed.

[0125] Figure 4 A flowchart of an information processing method provided by another embodiment of the present application is shown. The execution subject of the method can be the service end 1. As shown in Figure 4 The method includes the following steps:

[0126] 201. For a third to-be-processed problem, a processing flow is executed to obtain third processing information corresponding to the third to-be-processed problem.

[0127] The processing flow includes: obtaining information related to the to-be-processed problem (i.e., detail information of the to-be-processed problem) according to a preset requirement; obtaining target knowledge as a basis for processing the to-be-processed problem from a knowledge base; and calling a preset model to generate processing information corresponding to the to-be-processed problem based on the information and the target knowledge.

[0128] 202. If the processing flow is in a first mode, the third to-be-processed problem is processed based on manual processing information made by a human being for the third to-be-processed problem.

[0129] The third processing information is used in combination with the manual processing information as a basis for adjusting the processing flow.

[0130] 203. If the processing flow is in the second mode, a third to-be-processed problem is processed based on the third processing information.

[0131] After the third processing information corresponding to the third to-be-processed problem is obtained by executing the processing flow, a mode in which the processing flow is located can be acquired.

[0132] In some embodiments, the mode in which the processing flow is located can be manually configured by a user.

[0133] In other embodiments, the mode in which the processing flow is located can also be determined based on running data of the processing flow.

[0134] After the mode in which the processing flow is located is acquired, if the mode in which the processing flow is located is the first mode, step 202 is executed, and if the mode in which the processing flow is located is the second mode, step 203 is executed.

[0135] In step 202, the third to-be-processed problem is processed based on the artificial processing information made by the human being on the third to-be-processed problem, that is, the artificial processing information takes effect.

[0136] In step 203, the third processing information is used in combination with the artificial processing information as a basis for adjusting the processing flow. The generation process of the basis for adjusting the processing flow can refer to the corresponding content in the above embodiments, which will not be described here.

[0137] In the technical scheme provided by the embodiments of the present application, when the processing flow is in the first mode, the to-be-processed problem is processed based on the artificial processing information, and the processing information output by the processing flow is used in combination with the artificial processing information as a basis for adjusting the processing flow. When the processing flow is in the second mode, the to-be-processed problem is processed based on the processing information output by the processing flow, that is, the processing information output by the processing flow is used to process the to-be-processed problem. That is, after the processing flow is put into operation / deployed, the optimal processing path is selected by judging the mode in which the processing flow is located, and then a dynamic balance between efficiency and reliability is achieved. Moreover, when the processing flow is in the first mode, the basis for adjusting the processing flow can be generated based on the processing information output by the processing flow and the artificial processing information, so that the processing flow can be continuously optimized.

[0138] In some embodiments, the mode in which the processing flow is located supports dynamic switching between the first mode and the second mode.

[0139] The dynamic switching can be manual dynamic switching or automatic dynamic switching.

[0140] For example, when the processing accuracy of the processing flow decreases and is lower than a first preset threshold, the user can switch the model in which the processing flow is located to the first mode; when the processing accuracy of the processing flow increases and is higher than a second preset threshold after being adjusted, the user can switch the model in which the processing flow is located to the second mode. The second preset threshold can be greater than the first preset threshold.

[0141] For another example, the mode in which the processing flow is located can also be dynamically determined based on running data of the processing flow. For example, the running data can include the processing accuracy of the processing flow in a preset time period (for example, the last day, the last week or the last month), and the mode in which the processing flow is currently located is determined based on the processing accuracy. For example, when the processing accuracy is less than or equal to a first threshold, it is determined that the processing flow is in the first mode. When the processing accuracy is greater than the first threshold, it is determined that the processing flow is in the second mode.

[0142] The technical scheme provided in the embodiments of the present application realizes dynamic switching of the mode of the processing flow, that is, dynamic switching of the optimal processing path, which helps to improve the flexibility of the scheme and facilitates iterative updating of the processing flow.

[0143] It should be noted that the contents of the steps in the method provided in the embodiments of the present application which are not described in detail can be referred to the corresponding contents in the above embodiments, which will not be described here. In addition, the method provided in the embodiments of the present application can include other parts or all steps in the above embodiments in addition to the above steps, which can be specifically referred to the corresponding contents of the above embodiments, which will not be described here.

[0144] Figure 5 The flowchart of the information processing method provided in another embodiment of the present application is shown. The execution subject of the method can be the above-mentioned server 1. As shown in the figure, the method includes the following steps: Figure 5

[0145] 301. According to a preset requirement, the details of the problem to be processed are obtained.

[0146] 302. The target knowledge is obtained from the knowledge base.

[0147] The knowledge base is constructed based on a preset problem processing specification.

[0148] 303. Based on the information and the target knowledge, a preset model is used to generate processing information corresponding to the dispute.

[0149] The specific implementation of the above step 301 can be referred to the specific implementation of the above step S1; the specific implementation of the above step 302 can be referred to the specific implementation of the above step S2; the specific implementation of the above step 303 can be referred to the specific implementation of the above step S3. ​

[0150] In the technical solutions provided by the embodiments of the present application, the preset model is used to automatically analyze problems and generate processing information, which can not only reduce the dependence on manual work and significantly reduce operating costs, but also shorten the processing period and improve user satisfaction. In addition, the introduction of the knowledge base can provide the model with the required problem processing specifications, thereby reducing the training cost of the preset model.

[0151] It should be noted that the details of the steps in the method provided by the embodiments of the present application are described in the corresponding content of the above embodiments, which will not be repeated here. In addition, the method provided by the embodiments of the present application can include other parts or all of the steps in the above embodiments in addition to the above steps, which can be specifically understood from the corresponding content of the above embodiments, and will not be repeated here.

[0152] Figure 6 The flowchart of the information processing method provided by another embodiment of the present application is shown. The execution subject of the method can be the above-mentioned server 1. As shown in the figure, the method comprises the following steps: Figure 6

[0153] 401. For the first e-commerce after-sales problem, a processing flow is executed to obtain first processing information corresponding to the first e-commerce after-sales problem.

[0154] The processing flow comprises: obtaining detail information of the e-commerce after-sales problem according to a preset requirement; obtaining target knowledge from a knowledge base as a basis for processing the e-commerce after-sales problem; and calling a preset model to generate processing information corresponding to the e-commerce after-sales problem based on the detail information and the target knowledge.

[0155] 402. Determine whether the first processing information needs to be audited.

[0156] 403. When no audit is required, the first e-commerce after-sales problem is processed based on the first processing information.

[0157] 404. When the audit is required, the first processing information is sent to an artificial audit client.

[0158] The first e-commerce after-sales problem can include but is not limited to: return and exchange dispute, refund dispute, fake goods dispute, etc.

[0159] The specific implementation of the above steps 401 to 404 can be understood from the corresponding content of the above embodiments, which will not be repeated here. Figure 2

[0160] ​​The technical scheme provided in the embodiments of the present application is that the processing flow is used to process the e-commerce after-sales problem based on the knowledge base and the preset model to obtain processing information, that is, the relevant knowledge is introduced into the preset model through the knowledge base, so that the training cost of the preset model can be reduced. In addition, after obtaining the processing information output by the processing flow, it is further judged whether the processing information needs to be audited by a human being. When the processing information needs to be audited by a human being, the processing information is sent to a human auditing client for human auditing. When the processing information does not need to be audited, the processing information is directly used to process the problem to be processed. It can be seen that, on the premise of automatically processing the e-commerce after-sales problem by using the processing flow to reduce the human processing cost and improve the information processing efficiency, the processing error rate can also be effectively reduced.

[0161] It should be noted that the contents not fully described in each step of the method provided in the embodiments of the present application can be referred to the corresponding contents in the above embodiments, which will not be described here. In addition, the method provided in the embodiments of the present application can further include other parts or all steps in the above embodiments in addition to the above steps, which can be specifically referred to the corresponding contents of the above embodiments, which will not be described here.

[0162] It should be noted that the processing flow in the present application can be a workflow.

[0163] The dispute processing scheme provided in the embodiments of the present application will be exemplarily introduced in combination with the e-commerce scene. As shown in the following figure, Figure 7 The method involves two flows: the first flow and the second flow.

[0164] The second flow (i.e. the above processing flow) is executed after being started by the first flow, and involves the following steps:

[0165] 74. Obtain the basic information based on the dispute work order id.

[0166] The basic information can include but is not limited to order information, refund order information, dispute order information, logistics information, transaction messages, chat records between the seller and the buyer, member attributes, product attributes, etc.

[0167] In addition, for the transaction message information and the chat record, the information amount can be large, and the core useful information is relatively scattered. Here, a text summary model is used to complete the summary and abstraction of the information.

[0168] 75. Based on the obtained information and the preset prompt word template, a contradiction recognition model is called to output the dispute points and the demands of the buyer and the seller.

[0169] 76. Based on the dispute point and appeal information, the basic information of the dispute work order, and the target knowledge retrieved from the knowledge base of dispute handling norms, a decision model (i.e., the preset model above) is invoked to output decision information.

[0170] After obtaining the decision information, the second process ends.

[0171] The first process includes the following steps:

[0172] 71. Through a timing task, the corresponding dispute work order table is scanned every certain period of time to obtain the dispute work order that needs to be executed.

[0173] 72. The dispute work order is processed to obtain a dispute work order id.

[0174] 72. Based on the dispute work order id, the second process is started.

[0175] 77. After receiving the message of the second process, the message is parsed to obtain the model result (i.e., the decision information).

[0176] If currently in the first stage (corresponding to the first mode above), step 78 is executed; if currently in the second stage, step 80 is executed.

[0177] 78. The decision result is stored in the decision information record base.

[0178] 79. After the customer service completes the decision on this dispute work order, the decision information of the customer service decision is compared with the decision information of the model. Through the comparison of data, a basis for adjusting the first process is generated.

[0179] Based on the basis, the first process can be continuously improved.

[0180] 80. It is determined whether the decision information needs to be reviewed.

[0181] If it needs to be reviewed, step 82 is executed; otherwise, step 81 is executed.

[0182] 81. The decision is executed.

[0183] Based on the decision information, the decision is executed.

[0184] 82. Review.

[0185] That is, the decision information is sent to the manual review client for manual review. If the manual review is passed, the decision is executed based on the decision information.

[0186] In the second stage, after obtaining the decision information of the model, the decision can be directly executed. However, in order to avoid excessive dependence on the model, disputes exceeding a certain refund / payout threshold need to be manually reviewed again.

[0187] It should be noted that in order not to let the consumer feel that the service process has no temperature, the judgment solution against the consumer "unfavorable" is delayed.

[0188] In practical application, after the effective judgment, the dispute ticket is completed, and thereafter, the dispute ticket can be tracked to determine whether the dispute ticket involves subsequent complaint events, and then the problems existing in the second process are continuously improved according to the tracking information.

[0189] It should be noted that the time consumption of the second process can be relatively long, and the second process can inform the first process that the dispute processing is completed through asynchronous messages (that is, offline messages), and then subsequent processing is performed.

[0190] In practical application, when the judgment accuracy rate output by the second process reaches a certain effect, the implementation of the first process can be simplified or omitted, for example: the judgment of the dispute ticket can be directly completed on the second process side, which is a complete dispute judgment intelligent agent.

[0191] Figure 8 The flowchart of the information processing method provided by another embodiment of the application is shown. The execution subject of the method can be the artificial audit client 2. As shown in the figure, the method includes the following steps: Figure 8

[0192] 501, receiving the first processing information corresponding to the first to-be-processed problem sent by the server.

[0193] The first processing information is obtained by executing a processing flow, and the processing flow includes: obtaining detail information of the to-be-processed problem according to a preset requirement; obtaining target knowledge as a basis for processing the to-be-processed problem from a knowledge base; and based on the detail information and the target knowledge, calling a preset model to generate processing information corresponding to the to-be-processed problem.

[0194] 502, display the first processing information for artificial audit.

[0195] 503, in response to the confirmation operation of the user for the first processing information, send confirmation information to the server.

[0196] The confirmation information is used to instruct the server to process the first to-be-processed problem based on the first processing information.

[0197] In the above 501, the artificial audit client 2 can actively obtain the first processing information corresponding to the first to-be-processed problem from the server, or passively receive the first processing information corresponding to the first to-be-processed problem sent by the server.

[0198] ​For example, in response to a user's acquisition operation, the manual review client 2 may send a request to the server to obtain the first processing information corresponding to the first pending issue, and receive the first processing information returned by the server in response to the request. The acquisition operation may be an operation to open a pending review interface. The pending review interface is used to display the first pending review information.

[0199] In the above 502, after receiving the first processing information, the manual review client displays the first processing information on the review interface for manual review.

[0200] In the above 503, in response to the user's confirmation operation on the first processing information, confirmation information is sent to the server, and the confirmation information is used to instruct the server to process the first to-be-processed problem based on the first processing information.

[0201] In the technical solution provided by the embodiments of this application, the processing flow processes pending issues based on a relevant knowledge base and a preset model to obtain processing information. In other words, the knowledge base is used to introduce relevant knowledge into the preset model, which can reduce the training cost of the preset model. In addition, after obtaining the processing information output by the processing flow, the processing information is sent to a manual review client for manual review. It can be seen that this solution can effectively reduce the processing error rate while utilizing the processing flow to automatically process pending issues to reduce manual processing costs and improve information processing efficiency.

[0202] In some embodiments, the above method may further include:

[0203] 504. In response to the user's modification operation on the first processing information, send the modified processing information to the server.

[0204] The modified processing information is used by the server to process the first pending issue.

[0205] For example, Figure 3 The pending review interface 300 shown in FIG. 1 displays processing information a31 for dispute ticket A. A "Pass" control 32 and a "Fail" control 33 are displayed for processing information a31. In response to a trigger operation on the "Pass" control 32, processing information a31 is determined to have passed the review. In response to a trigger operation on the "Fail" control 33, processing information a31 is determined to have failed the review and a processing information modification interface (not shown) is displayed. The processing information modification interface displays processing information a for manual modification.

[0206] It should be noted that the details of the steps in the method provided by the embodiments of the present application are not fully described and can be referred to the corresponding contents in the above embodiments. In addition, the method provided by the embodiments of the present application can include other parts or all steps in the above embodiments in addition to the above steps, and the specific contents can be referred to the corresponding contents in the above embodiments, which will not be described here.

[0207] The embodiments of the present application also provide an information processing system, as shown in Figure 1 The system comprises a server 1 and an artificial audit client 2; wherein,

[0208] The server 1 is configured to implement the information processing method in the above Figure 2 or Figure 6 corresponding embodiments.

[0209] The artificial audit client 2 is configured to receive the first processing information sent by the server; and display the first processing information for artificial audit.

[0210] It should be noted that the specific implementation of the server and the client in the system provided by the embodiments of the present application can be referred to the corresponding contents in the above corresponding embodiments, which will not be described here.

[0211] Another embodiment of the present application provides an information processing device, which is configured to:

[0212] execute a processing flow for a first to-be-processed problem to obtain first processing information corresponding to the first to-be-processed problem, wherein the processing flow comprises: obtaining detail information of the to-be-processed problem according to a preset requirement; obtaining target knowledge from a knowledge base as a basis for processing the to-be-processed problem; and calling a preset model to generate processing information corresponding to the to-be-processed problem based on the detail information and the target knowledge;

[0213] determine whether the first processing information needs to be audited;

[0214] when the first processing information does not need to be audited, process the first to-be-processed problem based on the first processing information;

[0215] when the first processing information needs to be audited, send the first processing information to an artificial audit client.

[0216] Optionally, the executed processing flow is an adjusted processing flow; and the device is further configured to:

[0217] execute a to-be-adjusted processing flow for a second to-be-processed problem to obtain second processing information corresponding to the second to-be-processed problem;

[0218] obtain artificial processing information corresponding to the second to-be-processed problem;

[0219] generate reference information as an adjustment basis of the to-be-adjusted processing flow based on the second processing information and the artificial processing information.

[0220] Optionally, the adjustment comprises adjustment of an input parameter of the preset model.

[0221] Optionally, the adjustment does not comprise adjustment of a structure and / or parameter inside the preset model.

[0222] Optionally, the preset requirement comprises a preset to-be-acquired information item.

[0223] Optionally, the processing flow further comprises:

[0224] invoke a contradiction recognition model to acquire a contradiction between a plurality of participants involved in the to-be-handled problem according to the detail information;

[0225] invoke a preset model to generate processing information corresponding to the to-be-handled problem based on the detail information and the target knowledge, comprising:

[0226] invoke a preset model to generate processing information corresponding to the to-be-handled problem based on the detail information, the contradiction and the target knowledge.

[0227] Optionally, the contradiction comprises dispute content and / or appeal information of the plurality of participants.

[0228] Optionally, the acquiring information related to the first to-be-handled problem according to a preset requirement comprises:

[0229] acquiring initial detail information of a to-be-handled problem according to the preset requirement;

[0230] performing text summarization on the initial detail information by using a preset text summarization model to obtain the detail information.

[0231] Optionally, the apparatus for judging whether the first processing information needs to be audited is specifically configured to:

[0232] judge whether the first processing information needs to be audited according to the first processing information and / or information related to the first to-be-handled problem.

[0233] It should be noted that each apparatus provided in the above embodiments can implement the technical solutions described in the corresponding interface display method embodiments, and the specific implementation principles can be referred to the corresponding content in the interface display method embodiments, which will not be described here.

[0234] Another embodiment of the present application provides an information processing apparatus, which is configured to:

[0235] For the third to-be-handled problem, a processing flow is executed to obtain third processing information corresponding to the third to-be-handled problem, where the processing flow includes: obtaining detail information of the to-be-handled problem according to a preset requirement; obtaining target knowledge serving as a basis for processing the to-be-handled problem from a knowledge base; and generating processing information corresponding to the to-be-handled problem by calling a preset model based on the detail information and the target knowledge.

[0236] If the processing flow is in the first mode, the third to-be-handled problem is processed based on artificial processing information made by an artificial for the third to-be-handled problem, where the third processing information is used in combination with the artificial processing information as a basis for adjusting the processing flow.

[0237] If the processing flow is in the second mode, the third to-be-handled problem is processed based on the third processing information.

[0238] Optionally, the mode in which the processing flow is located supports dynamic switching between the first mode and the second mode.

[0239] It should be noted that each of the apparatuses provided in the above embodiments can implement the technical solutions described in the method embodiments above, and the specific implementation principles can be referred to the corresponding content in the method embodiments, which will not be described here in detail.

[0240] Another embodiment of the present application provides an information processing apparatus, which is configured to:

[0241] obtain detail information of a to-be-handled problem according to a preset requirement;

[0242] obtain target knowledge from a knowledge base based on the detail information, where the knowledge base is constructed based on a preset problem processing specification;

[0243] generate processing information corresponding to the dispute by using a preset model based on the detail information and the target knowledge.

[0244] It should be noted that each of the apparatuses provided in the above embodiments can implement the technical solutions described in the method embodiments above, and the specific implementation principles can be referred to the corresponding content in the method embodiments, which will not be described here in detail.

[0245] Another embodiment of the present application provides an information processing apparatus, which is configured to:

[0246] For the first e-commerce after-sales problem, a processing flow is executed to obtain first processing information corresponding to the first e-commerce after-sales problem, wherein the processing flow comprises: obtaining detail information of the e-commerce after-sales problem according to a preset requirement; obtaining target knowledge serving as a basis for processing the e-commerce after-sales problem from a knowledge base; and based on the detail information and the target knowledge, calling a preset model to generate processing information corresponding to the e-commerce after-sales problem;

[0247] It is determined whether the first processing information needs to be audited.

[0248] When no audit is required, the first to-be-processed problem is processed based on the first processing information.

[0249] When the audit is required, the first processing information is sent to an artificial audit client.

[0250] It should be noted that the apparatuses provided in the above embodiments can implement the technical solutions described in the method embodiments above, and the specific implementation principles can be referred to the corresponding content in the method embodiments above, which will not be described herein.

[0251] Another embodiment of the present application provides an information processing apparatus, which is configured to:

[0252] receive first processing information corresponding to a first to-be-processed problem sent by a server, wherein the first processing information is obtained by executing a processing flow, and the processing flow comprises: obtaining detail information of the to-be-processed problem according to a preset requirement; obtaining target knowledge serving as a basis for processing the to-be-processed problem from a knowledge base; and based on the detail information and the target knowledge, calling a preset model to generate processing information corresponding to the to-be-processed problem;

[0253] display the first processing information for artificial audit;

[0254] In response to a confirmation operation of a user on the first processing information, send confirmation information to the server, wherein the confirmation information is used to instruct the server to process the first to-be-processed problem based on the first processing information.

[0255] Optionally, the apparatus is further configured to:

[0256] In response to a modification operation of a user on the first processing information, send the modified processing information to the server, wherein the modified processing information is used to process the to-be-processed problem by the server.

[0257] It should be noted that the apparatuses provided in the above embodiments can implement the technical solutions described in the method embodiments above, and the specific implementation principles can be referred to the corresponding content in the method embodiments above, which will not be described herein.

[0258] An embodiment of the present application further provides an electronic device. As shown in Figure 9 The electronic device includes a processor 42 and a memory 41. The memory 41 is configured to store one or more computer programs (or instructions). The processor 42 is coupled to the memory 41 and configured to execute the one or more computer programs to implement the steps in the method provided by the embodiments of the present application.

[0259] Further, the electronic device further includes a communication component 43, a display 44, a power supply component 45, an audio component 46, and other components. Only some components are shown here, and the electronic device does not necessarily include only these components.

[0260] The memory can be implemented by any type of volatile or nonvolatile storage devices or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read only memory (EEPROM), erasable programmable read only memory (EPROM), programmable read only memory (PROM), read only memory (ROM), magnetic storage, flash memory, magnetic disk or optical disk.

[0261] The display includes a screen, which can include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes the touch panel, the screen can be implemented as a touch screen to receive an input signal from a user. The touch panel includes one or more touch sensors to sense a touch, a slide, and a gesture on the touch panel. The touch sensor can not only sense a boundary of a touch or a slide action, but also detect a duration and a pressure related to a touch or a slide operation.

[0262] The power supply component provides power to various components of the device in which the power supply component is located. The power supply component can include a power management system, one or more power sources, and other components associated with generating, managing, and distributing power to the device in which the power supply component is located.

[0263] The audio component can be configured to output and / or input audio signals. For example, the audio component includes a microphone (MIC) that is configured to receive an external audio signal when a device in which the audio component is positioned is in an operational mode, such as a call mode, a recording mode, and a voice recognition mode. The received audio signal can be further stored in a memory or transmitted via a communication component. In some embodiments, the audio component further includes a speaker for outputting audio signals.

[0264] Accordingly, the embodiments of the present application also provide a computer readable storage medium storing a computer program, when the computer program is executed by a processor, the processor is enabled to implement each step in the above method embodiments. The computer readable storage medium includes volatile or non-volatile or a combination thereof, and can be removable or non-removable. Examples of the computer readable storage medium include, but are not limited to, phase-change random access 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), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), flash memory or other memory technology, compact disc read-only memory (CD-ROM), digital video disc (DVD) or other optical storage, magnetic cassette, magnetic tape disk storage or other magnetic storage devices, or any other non-transmission medium.

[0265] Accordingly, the embodiments of the present application also provide a computer program product, the computer program product includes a computer program or instructions, when the computer program or instructions are executed by a processor, the processor is enabled to implement each step in the above method embodiments. It should be understood that each process or a combination of multiple processes in the above method flow can be implemented by the computer program or instructions. In addition, these computer programs or instructions can be applied to the processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing devices, so that the processor of the general-purpose computer, the special-purpose computer, the embedded processor or other programmable data processing devices can be implemented as a device for implementing the corresponding functions in the above method embodiments.

[0266] It should also be noted that the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, commodity, or apparatus that includes a series of elements includes not only those elements but also other elements not explicitly listed, or includes elements inherent to such process, method, commodity, or apparatus. In the absence of further limitations, an element defined by the phrase "comprises a ..." does not exclude the presence of other identical elements in the process, method, commodity, or apparatus that includes the element.

[0267] The above are merely embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various modifications and variations. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application should all be included within the scope of the claims of the present application.

Claims

1. An information processing method, characterized in that: include: For a first problem to be processed, a processing flow is executed to obtain first processing information corresponding to the first problem to be processed, wherein the processing flow includes: obtaining detailed information of the problem to be processed according to preset requirements; obtaining target knowledge as a basis for processing the problem to be processed from a knowledge base; and based on the detailed information and the target knowledge, calling a preset model to generate processing information corresponding to the problem to be processed; determining whether the first processing information requires review; When the first processing information does not need to be reviewed, processing the first pending issue based on the first processing information; When the first processing information needs to be reviewed, the first processing information is sent to a manual review client.

2. The method according to claim 1, characterized in that The executed processing flow is the adjusted processing flow; The method further comprises: For the second problem to be processed, executing the processing flow to be adjusted to obtain second processing information corresponding to the second problem to be processed; Obtaining manual processing information corresponding to the second pending issue; Based on the second processing information and the manual processing information, reference information serving as a basis for adjusting the processing flow to be adjusted is generated.

3. The method according to claim 2, characterized in that The adjustment includes adjusting the input parameters of the preset model.

4. The method according to claim 2, characterized in that The adjustment does not include adjustment of the network structure and / or network parameters of the preset model.

5. The method according to any one of claims 1 to 4, characterized in that The preset requirements include: preset information items to be obtained.

6. The method according to any one of claims 1 to 4, characterized in that The processing flow also includes: Based on the detailed information, a contradiction identification model is called to obtain contradictions between multiple parties involved in the problem to be processed; The calling of a preset model to generate processing information corresponding to the problem to be processed based on the detailed information and the target knowledge includes: Based on the detailed information, the contradiction and the target knowledge, a preset model is called to generate processing information corresponding to the problem to be processed.

7. The method according to claim 6, characterized in that The contradiction includes the dispute content and / or appeal information of the multiple parties.

8. The method according to any one of claims 1 to 4, characterized in that Obtaining detailed information about pending issues according to preset requirements includes: Obtaining initial detailed information of the problem to be processed according to the preset requirements; The initial detailed information is summarized using a preset text summarization model to obtain the detailed information.

9. The method according to any one of claims 1 to 4, characterized in that Determining whether the first processing information requires review includes: Determine whether the first processing information needs to be reviewed based on the first processing information and / or the detailed information.

10. An information processing method, characterized in that: include: For the third problem to be processed, a processing flow is executed to obtain third processing information corresponding to the third problem to be processed, wherein the processing flow includes: obtaining detailed information of the problem to be processed according to preset requirements; obtaining target knowledge as a basis for processing the problem to be processed from a knowledge base; and based on the detailed information and the target knowledge, calling a preset model to generate processing information corresponding to the problem to be processed; If the processing flow is in the first mode, processing the third problem to be processed based on manual processing information manually generated for the third problem to be processed, wherein the third processing information is used in combination with the manual processing information as a basis for adjusting the processing flow; If the processing flow is in the second mode, the third problem to be processed is processed based on the third processing information.

11. The method according to claim 10, characterized in that The mode in which the processing flow is located supports dynamic switching between the first mode and the second mode.

12. An information processing method, characterized in that: include: Obtain detailed information on pending issues according to preset requirements; Based on the detailed information, target knowledge is acquired from a knowledge base, wherein the knowledge base is constructed based on a preset problem processing specification; Based on the detailed information and the target knowledge, a preset model is used to generate processing information corresponding to the problem to be processed.

13. An information processing method, characterized in that: include: For a first e-commerce after-sales problem, a processing flow is executed to obtain first processing information corresponding to the first e-commerce after-sales problem, wherein the processing flow includes: obtaining detailed information of the e-commerce after-sales problem according to preset requirements; obtaining target knowledge from a knowledge base as a basis for processing the e-commerce after-sales problem; and based on the detailed information and the target knowledge, calling a preset model to generate processing information corresponding to the e-commerce after-sales problem; determining whether the first processing information requires review; If no review is required, processing the after-sales issue of the first e-commerce company based on the first processing information; When review is required, the first processing information is sent to a manual review client.

14. An information processing method, characterized in that: include: receiving first processing information corresponding to a first pending problem sent by a server, the first processing information being obtained by executing a processing flow, the processing flow comprising: obtaining detailed information of the pending problem according to preset requirements; obtaining target knowledge serving as a basis for processing the pending problem from a knowledge base; and invoking a preset model to generate processing information corresponding to the pending problem based on the detailed information and the target knowledge; displaying the first processing information for manual review; In response to the user's confirmation operation on the first processing information, confirmation information is sent to the server, where the confirmation information is used to instruct the server to process the first to-be-processed issue based on the first processing information.

15. The method according to claim 14, characterized in that include: In response to the user's modification operation on the first processing information, the modified processing information is sent to the server, and the modified processing information is used by the server to process the first to-be-processed issue.

16. An information processing system, characterized in that include: Server and manual review client; in The server is configured to implement the method according to any one of claims 1 to 9 or the method according to claim 13; The manual review client is used to: receive the first processing information sent by the server; and display the first processing information for manual review.

17. An electronic device, characterized in that: include: memory and a processor, wherein The memory is used to store programs; The processor is coupled to the memory, and is configured to execute the program stored in the memory to implement the method according to any one of claims 1 to 15.

18. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a computer, the method according to any one of claims 1 to 15 can be implemented.

19. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the method according to any one of claims 1 to 15 is implemented.

Images