Information processing method and device and electronic equipment

Abstract

The invention discloses an information processing method and device and electronic equipment. Relates to the field of artificial intelligence, and the method comprises the steps: predicting whether a target message of a user to be received by a target processing system belongs to a target message type or not according to at least one preset access rule under the condition that the target processing system starts an artificial customer service, the admission rule represents a condition that the target message belongs to a target message type and needs to be met, and the target message type is a message type supporting a target processing system to replace manual answering; and if it is predicted that the target message of the user to be received by the target processing system belongs to the target message type, determining answer information according to the target message through the target processing system under the condition that the target processing system receives the target message. The technical problem of high labor cost caused by the fact that messages sent by all consumers are directly answered manually in the manual customer service stage in the prior art is solved.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence, and more specifically, to an information processing method, device and electronic device. Background Art

[0002] Online customer service, also known as online customer service or network customer service, is an indispensable part of modern e-commerce and Internet services. It uses various technical means to provide instant customer support and problem-solving services through the Internet. When consumers find a platform for help, the platform will first provide consumers with robot services, but there are still many consumers who are unwilling to talk to robots and thus look for human customer service.

[0003] At present, the relevant technology directly relies on manual answers to all messages sent by consumers at the manual customer service stage, which leads to the problem of high labor costs.

[0004] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention

[0005] The embodiments of the present application provide an information processing method, device and electronic device to at least solve the technical problem in the related art that in the manual customer service stage, all messages sent by consumers are directly answered manually, resulting in high labor costs.

[0006] According to one aspect of an embodiment of the present application, there is provided an information processing method, comprising: in a case where a target processing system initiates a manual customer service, predicting whether a target message of a user to be received by the target processing system belongs to a target message type based on at least one preset access rule, wherein the access rule represents a condition that needs to be met for the target message to belong to the target message type, and the target message type is a message type that supports the target processing system to replace manual answers; if it is predicted that the target message of the user to be received by the target processing system belongs to the target message type, then when the target processing system receives the target message, the target processing system determines answer information based on the target message.

[0007] Furthermore, determining the answer information according to the target message through the target processing system includes: determining the action to be executed from multiple actions according to the target message through the action decision model in the target processing system, wherein the multiple actions include a first action and at least one second action, the first action refers to an action for switching to manual customer service to answer the target message, and the second action refers to an action for determining the answer information according to the target message; when the action to be executed is the second action, the answer information is determined based on the second action and the target message.

[0008] Furthermore, different second actions are used to determine answer information based on different methods and / or different data, and at least one second action includes a second action belonging to the first category. The second action belonging to the first category refers to an action used to determine answer information based on retrieval enhancement generation technology and the target message. When the action to be executed is the second action, determining the answer information based on the second action and the target message includes: if the action to be executed is the second action belonging to the first category, determining a knowledge base that matches the action to be executed; and determining the answer information based on the knowledge base and the target message based on the retrieval enhancement generation technology.

[0009] Furthermore, based on the retrieval enhancement generation technology, determining the answer information according to the knowledge base and the target message includes: judging, according to the target message, whether there is at least one target knowledge fragment matching the target message in the knowledge base matching the action to be executed; in the case that there is at least one target knowledge fragment in the knowledge base matching the action to be executed, generating a target indication statement according to the at least one target knowledge fragment and the target message, wherein the target indication statement is used to guide the target big model to rewrite the at least one target knowledge fragment to determine the answer information used to answer the target message; generating the answer information according to the target indication statement through the target big model.

[0010] Furthermore, after determining whether there is at least one target knowledge fragment matching the target message in the knowledge base that matches the action to be executed, the information processing method also includes: when there is no at least one target knowledge fragment matching the action to be executed in the knowledge base, inputting the target message into an answer information generation model; and outputting answer information corresponding to the target message through the answer information generation model.

[0011] Furthermore, different second actions are used to determine answer information based on different methods and / or different data, and at least one second action includes a second action belonging to the second category. The second action belonging to the second category refers to an action used to determine the answer information based on the robot answer information and the target message of the robot customer service service for the user. When the action to be executed is the second action, determining the answer information based on the second action and the target message includes: if the action to be executed is the second action belonging to the second category, determining the problem that matches the target message, wherein the problem that matches the target message is one of the following: a problem in the target message, a problem in the historical target message received by the target processing system; judging whether the robot answer information can solve the problem that matches the target message based on the degree of matching between the robot answer information and the problem that matches the target message, and obtaining a judgment result; determining the answer information based on the judgment result.

[0012] Furthermore, determining the answer information based on the judgment result includes: when the robot answer information can solve the problem matching the target message, obtaining manual answer information matching the robot answer information; determining the manual answer information matching the robot answer information as the answer information; when the robot answer information cannot solve the problem corresponding to the target message, determining an answer information generation script matching the problem matching the target message; and generating the answer information according to the answer information generation script.

[0013] Furthermore, after the target processing system receives the target message, the information processing method also includes: in the case of receiving an instruction to transfer to manual customer service sent by manual customer service, controlling the target processing system to stop determining answer information based on the target message; in the case of not receiving an instruction to transfer to manual customer service sent by manual customer service, judging whether to switch to manual customer service to answer the target message based on the target message; if it is determined to switch to manual customer service to answer the target message, controlling the target processing system to stop determining answer information based on the target message, and sending a service request to the manual customer service, wherein the service request is used to request the manual customer service to answer the target message.

[0014] Furthermore, judging whether to switch to manual customer service to answer the target message based on the target message includes: judging whether there is a target keyword in the target message to obtain a first judgment result, wherein the target keyword is a word that represents the customer service identity of the user questioning the manual customer service; judging whether the target message belongs to a message under a target scenario to obtain a second judgment result, wherein the target scenario is a scenario that represents the existence of negative emotions by the user; judging whether the conversation round to which the target message belongs is greater than a preset round to obtain a third judgment result; and determining whether to switch to manual customer service to answer the target message based on at least one of the first judgment result, the second judgment result, and the third judgment result.

[0015] Furthermore, after the target processing system determines the answer information according to the target message, the information processing method further includes: performing quality inspection on the answer information according to at least one preset quality inspection rule; and if the answer information passes the quality inspection, feeding back the answer information to the user.

[0016] Furthermore, after determining the answer information according to the target message through the target processing system, the information processing method also includes: feeding back the answer information to the user; if the next target message sent by the user after the target message is received, determining the target action to be executed from multiple actions according to the next target message through the action decision model, and executing the target action, wherein the multiple actions include a first action and at least one second action, the first action refers to an action for switching to manual customer service to answer the target message, and the second action refers to an action for determining the answer information according to the target message.

[0017] According to another aspect of an embodiment of the present application, an information processing device is also provided, including: a prediction unit, used to predict whether a target message of a user to be received by the target processing system belongs to a target message type according to at least one preset access rule when the target processing system starts a manual customer service, wherein the access rule characterizes the conditions that need to be met for the target message to belong to the target message type, and the target message type is a message type that supports the target processing system to replace manual answers; a determination unit, used to determine answer information according to the target message through the target processing system when the target processing system receives the target message if it is predicted that the target message of the user to be received by the target processing system belongs to the target message type.

[0018] Furthermore, the determination unit includes: a first determination sub-unit, used to determine the action to be executed from multiple actions according to the target message through the action decision model in the target processing system, wherein the multiple actions include a first action and at least one second action, the first action refers to an action for switching to manual customer service to answer the target message, and the second action refers to an action for determining answer information according to the target message; a second determination sub-unit, used to determine the answer information based on the second action and the target message when the action to be executed is the second action.

[0019] Furthermore, different second actions are used to determine answer information based on different methods and / or different data, and at least one second action includes a second action belonging to the first category. The second action belonging to the first category refers to an action used to determine the answer information based on the retrieval enhancement generation technology and the target message. The second determination subunit includes: a first determination module, which is used to determine the knowledge base matching the action to be executed if the action to be executed is the second action belonging to the first category; and a second determination module, which is used to determine the answer information based on the knowledge base and the target message based on the retrieval enhancement generation technology.

[0020] Furthermore, the second determination module includes: a judgment submodule, used to judge whether there is at least one target knowledge fragment matching the target message in the knowledge base matching the action to be executed according to the target message; a first generation submodule, used to generate a target indication statement according to the at least one target knowledge fragment and the target message when there is at least one target knowledge fragment in the knowledge base matching the action to be executed, wherein the target indication statement is used to guide the target big model to rewrite the at least one target knowledge fragment to determine the answer information for answering the target message; and a second generation submodule, used to generate the answer information according to the target indication statement through the target big model.

[0021] Furthermore, the information processing device also includes: a first processing unit, used to input the target message into the answer information generation model when there is no at least one target knowledge fragment in the knowledge base matching the action to be executed; and a second processing unit, used to output the answer information corresponding to the target message through the answer information generation model.

[0022] Furthermore, different second actions are used to determine answer information based on different methods and / or different data, and at least one second action includes a second action belonging to the second category. The second action belonging to the second category refers to an action used to determine the answer information based on the robot answer information and the target message of the robot customer service service for the user. The second determination subunit includes: an acquisition module, which is used to determine the problem matched by the target message if the action to be executed is a second action belonging to the second category, wherein the problem matched by the target message is one of the following: a problem in the target message, a problem in the historical target message received by the target processing system; a judgment module, which is used to judge whether the robot answer information can solve the problem matched by the target message according to the degree of match between the robot answer information and the problem matched by the target message, and obtain a judgment result; a third determination module, which is used to determine the answer information based on the judgment result.

[0023] Furthermore, the third determination module includes: an acquisition submodule, which is used to acquire manual answer information matching the robot answer information when the robot answer information can solve the problem matching the target message; a first determination submodule, which is used to determine the manual answer information matching the robot answer information as the answer information; a second determination submodule, which is used to determine the answer information generation script matching the problem matching the target message when the robot answer information cannot solve the problem corresponding to the target message; and a third generation submodule, which is used to generate the answer information according to the answer information generation script.

[0024] Furthermore, the information processing device also includes: a third processing unit, which is used to control the target processing system to stop determining the answer information according to the target message when receiving an instruction to transfer to manual customer service sent by manual customer service; a judgment unit, which is used to judge whether to switch to manual customer service to answer the target message according to the target message when not receiving an instruction to transfer to manual customer service sent by manual customer service; a fourth processing unit, which is used to control the target processing system to stop determining the answer information according to the target message if it is determined to switch to manual customer service to answer the target message, and send a service request to the manual customer service, wherein the service request is used to request the manual customer service to answer the target message.

[0025] Furthermore, the judgment unit includes: a first judgment sub-unit, used to judge whether there is a target keyword in the target message, and obtain a first judgment result, wherein the target keyword is a word that characterizes the user's questioning of the customer service identity of the manual customer service service; a second judgment sub-unit, used to judge whether the target message belongs to a message under a target scenario, and obtain a second judgment result, wherein the target scenario is a scenario that characterizes the user's negative emotions; a third judgment sub-unit, used to judge whether the dialogue round to which the target message belongs is greater than a preset round, and obtain a third judgment result; a third determination sub-unit, used to determine whether to switch to manual customer service to answer the target message based on at least one of the first judgment result, the second judgment result, and the third judgment result.

[0026] Furthermore, the information processing device also includes: a quality inspection unit, which is used to perform quality inspection on the answer information according to at least one preset quality inspection rule; and a fifth processing unit, which is used to feed back the answer information to the user if the answer information passes the quality inspection.

[0027] Furthermore, the information processing device also includes: a sixth processing unit, used to feed back the answer information to the user; a seventh processing unit, used to determine the target action to be executed from multiple actions according to the next target message through an action decision model if the next target message sent by the user after the target message is received, and execute the target action, wherein the multiple actions include a first action and at least one second action, the first action refers to an action for switching to manual customer service to answer the target message, and the second action refers to an action for determining the answer information based on the target message.

[0028] According to another aspect of an embodiment of the present invention, there is further provided an electronic device, comprising: a memory storing an executable program; and a processor for running the program, wherein any one of the above-mentioned information processing methods is executed when the program is running.

[0029] According to another aspect of an embodiment of the present invention, a computer program product is provided, including a computer program or instructions, and the computer program or instructions implement any one of the above information processing methods when executed by a processor.

[0030] In an embodiment of the present application, when the target processing system starts the manual customer service service, it is predicted whether the target message of the user to be received by the target processing system belongs to the target message type according to at least one preset access rule, wherein the access rule represents the conditions that the target message must meet for it to belong to the target message type, and the target message type is a message type that supports the target processing system to replace manual answers; if it is predicted that the target message of the user to be received by the target processing system belongs to the target message type, then when the target processing system receives the target message, the target processing system determines the answer information according to the target message. When the target processing system starts the manual customer service service, it is predicted whether the target message of the user to be received by the target processing system belongs to the target message type according to at least one preset access rule, thereby achieving effective prediction of the messages to be received that can be answered by the system instead of manually; if it is predicted that the target message to be received belongs to the target message type, then when the target message is received, the target processing system determines the answer information according to the target message, thereby achieving the system replacing manual answers to user messages in the manual customer service stage, thereby effectively carrying a part of the manual customer service service volume and reducing the manual customer service cost. The purpose of using the system to answer user messages that meet certain conditions in the manual customer service stage is achieved, and the technical effect of reducing the cost of manual customer service is realized, thereby solving the technical problem in the related technology of directly relying on manual answers to all messages sent by consumers in the manual customer service stage, resulting in high labor costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0032] Figure 1 is a schematic diagram of a computer terminal provided according to Embodiment 1 of the present application;

[0033] Figure 2 This is the process of the information processing method provided in Example 1 of this application Figure 1 ;

[0034] Figure 3 is an optional flow chart for determining answer information provided according to Embodiment 1 of the present application;

[0035] Figure 4 This is the process of the optional information processing method provided in Example 1 of the present application. Figure 2 ;

[0036] Figure 5 This is a flow chart of the custody end detection provided according to the first embodiment of the present application;

[0037] Figure 6 is a schematic diagram of an information processing device provided according to Embodiment 2 of the present application;

[0038] Figure 7 It is a structural block diagram of an electronic device provided according to Embodiment 3 of the present application. DETAILED DESCRIPTION

[0039] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.

[0040] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0041] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant laws, regulations and standards in the relevant regions, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0042] Example 1

[0043] According to an embodiment of the present application, an information processing method is also provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0044] The method embodiment provided in the first embodiment of the present application can be executed in a mobile terminal, a computer terminal or a similar computing device. Figure 1 FIG. 1 shows a hardware structure block diagram of a computer terminal (or mobile device) for implementing an information processing method. Figure 1 As shown, the computer terminal (or mobile device) 10 may include a processor set 102 (the processor set 102 may include but is not limited to a processing device such as a microprocessor (Microcontroller Unit, referred to as MCU) or a programmable logic device (Field Programmable Gate Array, referred to as FPGA), and the processor set 102 may include a processor set, Figure 1 102a, 102b, ..., 102n are used to illustrate), a memory 104 for storing data, and a transmission device 106 for communication functions. In addition, it may also include: a display, an input / output interface (I / O interface), a Universal Serial Bus (USB) port (which may be included as one of the ports of the BUS bus), a network interface, a power supply and / or a camera. It can be understood by those skilled in the art that Figure 1 The structure shown is only for illustration and does not limit the structure of the above electronic device. Figure 1 More or fewer components as shown, or with Figure 1 Different configurations are shown.

[0045] It should be noted that the one or more processors 102 and / or other data processing circuits described above may generally be referred to herein as "data processing circuits". The data processing circuits may be embodied in whole or in part as software, hardware, firmware, or any other combination thereof. In addition, the data processing circuits may be a single independent processing module, or may be integrated in whole or in part into any of the other components in the computer terminal 10 (or mobile device).

[0046] The memory 104 can be used to store software programs and modules of application software, such as program instructions / data storage devices corresponding to the information processing method in the embodiment of the present application. The processor 102 executes various functional applications and data processing by running the software programs and modules stored in the memory 104, that is, realizing the above-mentioned information processing method. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 104 may further include a memory remotely arranged relative to the processor 102, and these remote memories may be connected to the computer terminal 10 via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0047] The transmission device 106 is used to receive or send data via a network. The specific example of the above network may include a wireless network provided by a communication provider of the computer terminal 10. In one example, the transmission device 106 includes a network adapter (Network Interface Controller, referred to as NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 106 can be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0048] The display may be a touch screen type liquid crystal display that enables a user to interact with a user interface of the computer terminal 10 (or mobile device).

[0049] Online customer service, also known as online customer service or network customer service, is an indispensable part of modern e-commerce and Internet services. It uses various technical means to provide instant customer support and problem-solving services through the Internet. When consumers find a platform for help, the platform will first provide robot services to consumers, but there are still many consumers who are unwilling to talk to robots and thus look for human customer service. However, in some scenarios, for consumer problems, the human customer service stage can only provide consumers with the same solutions as the robot customer service stage. In this case, if we still rely on manual answers to all messages sent by consumers, it will ineffectively occupy a lot of service time of human customer service, resulting in high labor costs.

[0050] In the above technical background, the present application provides Figure 2 The information processing method shown. Figure 2 This is the process of the information processing method provided in Example 1 of this application Figure 1 The method comprises:

[0051] Step S201, when the target processing system starts the manual customer service, predict whether the target message of the user to be received by the target processing system belongs to the target message type according to at least one preset access rule, wherein the access rule represents the conditions that the target message must meet to belong to the target message type, and the target message type is a message type that supports the target processing system to replace manual answers.

[0052] Optionally, the information processing method in this embodiment is executed by a target processing system. The target processing system may be a system in a target platform, the target platform includes but is not limited to a shopping platform, a consulting platform or other platforms that provide online customer service needs, and the user refers to a user in the target platform, for example, when the target platform is a shopping platform, the user may be a consumer.

[0053] Optionally, the target processing system may start the manual customer service for the current user when receiving the user's request to transfer to manual customer service. The target processing system may also start the manual customer service for the current user when the user or the target processing system triggers a specific condition.

[0054] Optionally, before the target processing system initiates manual customer service, customer service is provided to the user through a robot customer service.

[0055] Optionally, after the target processing system starts the manual customer service for the current user, it is equivalent to entering the manual customer service stage.

[0056] Optionally, since in some scenarios, manual customer service can still only provide consumers with the same solutions as in the robot customer service stage for consumer problems, when the manual customer service service is started, the target processing system can predict whether the message that the user is about to send in the manual customer service stage is a message with the same or similar content to the answer from the manual customer service and the robot customer service, and based on this, determine whether the message that the user is about to send in the manual customer service stage is a message type that supports the target processing system to replace manual answers (i.e., the aforementioned target message type).

[0057] In an optional embodiment, the target processing system may predict whether the target message of the user to be received by the target processing system belongs to the target message type according to at least one preset admission rule.

[0058] Optionally, the target processing system may determine that the target message of the user to be received by the target processing system belongs to the target message type when it is predicted that the target message of the user to be received by the target processing system satisfies all conditions in at least one access rule, and determine that the target message of the user to be received by the target processing system does not belong to the target message type when it is predicted that the target message of the user to be received by the target processing system does not satisfy at least one condition in at least one access rule.

[0059] Optionally, the target processing system may predict whether the target message of the user to be received by the target processing system belongs to the target message type based on the user information of the user to whom the target message to be received belongs and at least one admission rule. The user information at least includes the target history messages of the user in the robot customer service service, and the user information may also include at least one of the following: the user level of the user, the user's order information. Optionally, according to different actual application requirements, the user information may also include other information associated with the user.

[0060] Since the user may send the same or similar message as the robot customer service to the manual customer service when the robot customer service cannot solve his problem, it is possible to infer whether the target message of the user to be received belongs to the target message type in combination with the target history message of the user in the robot customer service. For example, in the case where the user information includes the above-mentioned target history message, the target processing system has preset corresponding relationships between multiple messages and target message types, and the access rule may be: determine whether the target history message of the user in the robot customer service belongs to the target message type, if the target history message belongs to the target message type, then determine that the target message of the user to be received by the target processing system meets the first condition of belonging to the target message type, and if the target history message does not belong to the target message type, then determine that the target message of the user to be received by the target processing system does not meet the first condition of belonging to the target message type. Among them, the target history message may be the Mth to last message sent by the user in the robot customer service service as a question statement, where M is a positive integer.

[0061] Since the message sent by the user to the manual customer service may be associated with the user's order status, it is possible to infer whether the target message belongs to the target message type in combination with the order status. For example, the user's order information includes the user's order status, and the order status includes but is not limited to successful transaction, failed transaction, refunded, refund failed, etc. In the case where the user information includes the above order information, the access rule may be: when the order status is the target order status, it is determined that the target message of the user to be received by the target processing system meets the second condition of belonging to the target message type; when the order status is not the target order status, it is determined that the target message of the user to be received by the target processing system does not meet the second condition of belonging to the target message type.

[0062] Since the target platform needs to pay more attention to the user experience of relatively important users, it is possible to infer whether the target message belongs to the target message type in combination with the user level. For example, there is a corresponding user level for the user, and the user level represents the importance of the user in providing manual customer service. For example, when the user is a consumer, the user level can be understood as the user's membership level in the platform. When the user information includes the above user level, the access rule can be: when the user's user level is lower than or equal to the preset level, it is determined that the target message of the user to be received by the target processing system meets the third condition of belonging to the target message type. When the user's user level is higher than the preset level, it is determined that the target message of the user to be received by the target processing system does not meet the third condition of belonging to the target message type.

[0063] In an optional embodiment, assuming that the access rule is the first one of the above examples, then when it is predicted that the target message of the user to be received by the target processing system meets the above first condition, it is determined that the target message of the user to be received by the target processing system belongs to the target message type.

[0064] In an optional embodiment, assuming that the access rules are the three of the above examples, when it is predicted that the target message of the user to be received by the target processing system meets the above-mentioned first condition, second condition and third condition, it is determined that the target message of the user to be received by the target processing system belongs to the target message type.

[0065] Step S202: If it is predicted that the target message of the user to be received by the target processing system belongs to the target message type, then when the target processing system receives the target message, the target processing system determines the reply information according to the target message.

[0066] Optionally, if it is predicted that the target message of the user to be received by the target processing system belongs to the target message type, the target processing system determines that it can replace the manual answer to the target message of the user to be received. In this case, if the target processing system receives the target message, the target processing system can directly determine the answer information according to the target message.

[0067] In an optional embodiment, the stage in which the target processing system replaces manual determination of the answer information can be called the trusteeship stage, or the intelligent trusteeship stage. That is, after the manual customer service is started, the customer service service is provided by the system instead of by the manual customer service.

[0068] For example, the target processing system may directly input the target message into the answer information generation model, and then the answer information generation model may output the answer information for the target message.

[0069] For another example, the target processing system may obtain the correspondence between multiple messages and multiple pieces of reply information, and thereby determine the reply information for the target message from the multiple pieces of reply information according to the correspondence.

[0070] For another example, the target processing system can determine the action to be executed from multiple actions according to the target message through an action decision model, wherein the multiple actions include a first action and at least one second action, the first action refers to an action for switching to manual customer service to answer the target message, and the second action refers to an action for determining answer information according to the target message, thereby determining the answer information based on the second action and the target message when the action to be executed is the second action.

[0071] Optionally, after the answer information is determined, the answer information may be fed back to the user.

[0072] Optionally, after the target processing system starts the manual customer service for the current user, it is equivalent to entering the manual customer service stage. Therefore, after determining the answer information, the target processing system can imitate the manual customer service to feed back the answer information to the user. For example, after the target processing system starts the manual customer service service, the customer service information of the manual customer service is displayed in the customer service interface in the terminal held by the user, and after determining the answer information, the answer information is displayed in the chat box of the manual customer service in the customer service interface, so that the user believes that the answer information is determined by the manual customer service.

[0073] In this solution, when the target processing system starts the manual customer service, it predicts whether the target message of the user to be received by the target processing system belongs to the target message type according to at least one preset access rule, and realizes the effective prediction of the message to be received that can be answered by the system instead of the human. If it is predicted that the target message to be received belongs to the target message type, then when the target message is received, the target processing system determines the answer information according to the target message, and realizes that the system replaces the human to answer the user message in the manual customer service stage, so that it can effectively carry a part of the manual customer service service volume and reduce the cost of manual customer service. The purpose of using the system to answer user messages that meet certain conditions in the manual customer service stage is achieved, and the technical effect of reducing the cost of manual customer service is achieved, thereby solving the technical problem of directly relying on human to answer all messages sent by consumers in the manual customer service stage in the related technology, resulting in high labor costs.

[0074] How to accurately determine the answer information for the target message is crucial. Therefore, in the information processing method provided in Example 1 of the present application, determining the answer information according to the target message through the target processing system includes: determining the action to be executed from multiple actions according to the target message through the action decision model in the target processing system, wherein the multiple actions include a first action and at least one second action, the first action refers to an action for switching to manual customer service to answer the target message, and the second action refers to an action for determining the answer information according to the target message; when the action to be executed is the second action, the answer information is determined based on the second action and the target message.

[0075] In an optional embodiment, after predicting that the target message of the user to be received by the target processing system belongs to the target message type, the target processing system can implement the hosting function based on the AI ​​Agent (artificial intelligence entity) mode, that is, receiving the target message based on the AI ​​Agent mode, and determining the answer information according to the target message based on the AI ​​Agent mode. Among them, AI Agent is an intelligent entity that can perceive the environment, make decisions and perform actions.

[0076] For example, after predicting that the target message of the user to be received belongs to the target message type, the target processing system can create an AI Agent instance and a Memory corresponding to the AI ​​Agent instance for the current user. Memory can be understood as a storage device or a memory function. The Memory stores long-term memory and short-term memory. Long-term memory includes but is not limited to the message records of the current user in the robot customer service, the message records of the user in the manual customer service, the user's order information, the user's user level and other information. Short-term memory includes but is not limited to information about the working process of the AI ​​Agent instance, such as the operations performed by the AI ​​Agent instance, the operation results of the performed operations, the working status of the AI ​​Agent instance, etc.

[0077] Optionally, after creating an AI Agent instance and a Memory corresponding to the AI ​​Agent instance, the target processing system may receive a target message through the AI ​​Agent instance, and determine an action to be executed from a plurality of actions according to the target message through the AI ​​Agent instance.

[0078] In an optional embodiment, the target processing system may determine the action to be executed from multiple actions according to the target message through an action decision model.

[0079] For example, the action decision model is called through the AI ​​Agent instance, and the target message is input into the action decision model, so that the action decision model makes an action decision among multiple actions according to the target message and outputs the action to be executed.

[0080] Optionally, the action decision model may be a neural network model, and the action decision model may be obtained by training based on the first training sample set, wherein the training samples in the first training sample set are sample messages of sample users, and the true labels corresponding to the training samples are actions. Thus, when the action decision model is applied, it can determine the accurate action to be executed according to the user's target message.

[0081] Since the target processing system has not received the target message of the user when predicting whether the target message of the user to be received by the target processing system belongs to the target message type according to at least one preset access rule, the target processing system may not receive the target message of the user actually received by the target processing system. That is, there may be a situation where the prediction result does not match the actual result. Therefore, the multiple actions include a first action and at least one second action. Among them, the first action refers to the action for switching to manual customer service to answer the target message, and the second action refers to the action for determining the answer information according to the target message. Therefore, when the action decision model is trained based on the first training sample set, the action decision model can not only learn the action for determining the answer information for different messages, but also learn the message type that supports the target processing system to replace the manual answer. That is, when the action decision model is applied, the action decision model can not only determine the action for determining the answer information for different messages, but also effectively filter the messages that do not support the target processing system to replace the manual answer.

[0082] Optionally, when the action to be executed is the first action, the target processing system may stop determining the answer information according to the target message and send a service request to the manual customer service to complete the execution of the first action. The service request is used to request the manual customer service to answer the target message.

[0083] For example, when the AI ​​Agent instance determines that the action to be executed output by the action decision model is the first action, it may stop calling the action decision model and send a service request to the manual customer service, and then close the AI ​​Agent instance after sending the service request to the manual customer service.

[0084] Optionally, when the action to be performed is a second action, the target processing system may determine the answer information based on the second action and the target message. Different second actions are used to determine the answer information according to different methods and / or different data.

[0085] For example, the second action is executed based on the target message through the AI ​​Agent instance. For any two second actions, the two second actions may determine the answer information in different ways and based on the same data, for example, one of the two second actions determines the answer information based on the target message, the retrieval enhancement generation technology and a certain knowledge base, and the other second action determines the answer information based on the target message, the vector similarity algorithm and the knowledge base; or, the second actions of the two second actions may determine the answer information in the same way and based on different data, for example, one of the two second actions determines the answer information based on the target message, the retrieval enhancement generation technology and a certain knowledge base, and the other of the two second actions determines the answer information based on the target message, the retrieval enhancement generation technology and another knowledge base; or, the second actions of the two second actions may determine the answer information in different ways and based on different data, for example, one of the two second actions determines the answer information based on the target message, the retrieval enhancement generation technology and a certain knowledge base, and the other of the two second actions determines the answer information based on the target message and the answer information generation model.

[0086] In an optional embodiment, the at least one second action includes a second action belonging to the first category, and the second action belonging to the first category refers to an action for determining answer information according to the retrieval enhancement generation technique and the target message.

[0087] In an optional embodiment, at least one second action includes a second action belonging to a second category, and the second action belonging to the second category refers to an action for determining answer information based on the robot answer information and target message of the robot customer service service for the user.

[0088] In an optional embodiment, at least one second action includes a second action belonging to a third category, and the second action belonging to the third category is an action used to determine answer information based on a first message and a preset answer information template, and the first message is a target message or a historical message sent by the user in the robot customer service service.

[0089] Optionally, in the process of determining the answer information based on the second action and the target message, if the action to be performed is a second action belonging to the third category, the first message is filled into a preset answer information template to obtain the answer information for the target message. An optional answer information template may be "Are you asking question A?", where "A" in the answer information template is used to fill in the first message.

[0090] Optionally, the target processing system may determine the question statement in the target message as the first message when the target message includes the question statement, and determine the first message based on the historical messages sent by the user in the robot customer service when the target message does not include the question statement. If the user has multiple historical messages sent in the robot customer service, a message including the question statement may be randomly determined from the multiple historical messages as the first message, or the Pth last message including the question statement sent by the user in the multiple historical messages may be determined as the first message, or other methods may be used to determine the first message from multiple historical messages, where P is a positive integer.

[0091] For example, assuming that a user sends the message "Why did the refund of my order fail?" in the robot customer service, and the target message in the manual customer service is "Hello", then the message "Why did the refund of my order fail?" can be determined as the first message, and the question "Are you asking the question 'Why did the refund of my order fail?'" can be determined as the answer information for the target message "Hello".

[0092] For example, the historical messages sent by the above-mentioned user in the robot customer service are obtained from the long-term memory through the AI ​​Agent instance, so as to determine the answer information based on the second action belonging to the third category.

[0093] In an optional embodiment, when there are multiple second actions, the second actions may include at least two of a second action belonging to the first category, a second action belonging to the second category, and a second action belonging to the third category.

[0094] It should be noted that by using the action decision model to determine the action to be performed from multiple actions, it is possible to make targeted decisions on the operations that the target processing system needs to perform based on the message sent by the user. By determining the answer information based on the second action and the target message when the action to be performed is the second action, it is possible to determine the answer information based on the message sent by the user, thereby improving the accuracy of the determined answer information. By setting the action to include the first action, the target processing system can effectively identify information that is not suitable for replacing manual answers, thereby facilitating the improvement of user experience.

[0095] In order to accurately determine the answer information, in the information processing method provided in Example 1 of the present application, different second actions are used to determine the answer information according to different methods and / or different data, at least one second action includes a second action belonging to the first category, and the second action belonging to the first category refers to an action used to determine the answer information based on the retrieval enhancement generation technology and the target message. When the action to be executed is the second action, determining the answer information based on the second action and the target message includes: if the action to be executed is the second action belonging to the first category, determining the knowledge base matching the action to be executed; and determining the answer information based on the knowledge base and the target message based on the retrieval enhancement generation technology.

[0096] Optionally, there is at least one second action belonging to the first category. Different second actions belonging to the first category determine the answer message in the same manner, but based on different data. The target processing system may preset a matching relationship between different second actions belonging to the first category and the knowledge base (i.e., the aforementioned based data).

[0097] Optionally, after determining that the action to be executed is a second action belonging to the first category, the target processing system may determine a knowledge base matching the action to be executed based on a matching relationship between different second actions belonging to the first category and the knowledge base.

[0098] Optionally, the knowledge base includes multiple knowledge fragments, and the knowledge fragment can be one of the following: preset answer information, historical answer information of manual customer service.

[0099] Optionally, the target processing system can determine the knowledge fragments associated with the target message from the knowledge base based on the retrieval enhancement generation technology, and then process the target message and the knowledge fragments associated with the target message through the target macro model to obtain a reply message for the target message.

[0100] For example, the AI ​​Agent instance can be used to determine the knowledge base that matches the action to be performed, and based on the retrieval enhancement generation technology, the answer information can be determined according to the knowledge base and the target message.

[0101] It should be noted that by determining the knowledge base that matches the action to be performed, the data based on the answer information can be determined in a targeted manner, thereby improving the accuracy of the determined answer information. By determining the answer information based on the retrieval enhancement generation technology, the accuracy of the determined answer information can be further improved, and the efficiency of determining the answer information can be improved.

[0102] In order to improve the personification of the answer information, in the information processing method provided in Example 1 of the present application, based on the retrieval enhancement generation technology, the answer information is determined according to the knowledge base and the target message, including: judging whether there is at least one target knowledge fragment matching the target message in the knowledge base matching the action to be executed according to the target message; when there is at least one target knowledge fragment in the knowledge base matching the action to be executed, generating a target indication statement according to the at least one target knowledge fragment and the target message, wherein the target indication statement is used to guide the target big model to rewrite the at least one target knowledge fragment to determine the answer information used to answer the target message; generating the answer information according to the target indication statement through the target big model.

[0103] Optionally, the knowledge base includes multiple knowledge fragments, and the knowledge fragment can be one of the following: preset answer information, historical answer information of manual customer service. In layman's terms, the preset answer information is equivalent to a preset fixed speech, and the historical answer information of manual customer service is equivalent to the speech provided by the manual customer service in the actual application process in the past.

[0104] Optionally, the target processing system can calculate the vector of the target message to obtain a first vector, and calculate the vector of the knowledge fragment in the knowledge base that matches the action to be executed to obtain a second vector corresponding to the knowledge fragment, thereby judging whether there is at least one target knowledge fragment matching the target message in the knowledge base that matches the action to be executed based on the similarity between the first vector and the second vector of the knowledge fragment in the knowledge base.

[0105] For example, when the similarity between the first vector and the second vector is greater than a preset threshold, the knowledge segment corresponding to the second vector is determined to be the target knowledge segment, and when the similarity between the first vector and the second vector is less than or equal to the preset threshold, the knowledge segment corresponding to the second vector is determined not to be the target knowledge segment. Optionally, the preset threshold may be determined by an administrator of the target processing system according to actual needs.

[0106] Optionally, when there is at least one target knowledge fragment in the knowledge base that matches the action to be executed, it is determined that the target message does not belong to the rejection data of the target large model. In this case, the target processing system can generate a target instruction statement based on the at least one target knowledge fragment and the target message. The rejection data of the target processing system can be understood as data that the target processing system refuses to process by the target large model, so as to prevent the target large model from processing such data and causing erroneous output.

[0107] Optionally, the target large model may be a large language model such as ChatGPT or ChatGLM. The instruction statement is a prompt for the target large model.

[0108] For example, the target processing system may fill at least one target knowledge fragment and a target message into a preset instruction statement template to obtain a target instruction statement. An optional instruction statement template may be as follows:

[0109] Please rewrite the given information to determine the response information for the user's message.

[0110] User message: What is the refund process for an order?

[0111] material:

[0112] -{Text content of document A}

[0113] -{Text content of document B}

[0114] -{Text content of document C}

[0115] Among them, "What is the refund process for an order" is equivalent to the target message, and "the text content of document A", "the text content of document B" and "the text content of document C" are equivalent to the target knowledge fragments. The instruction statement is used to guide or trigger the model to perform a specific task or generate a specific output. In this embodiment, the target instruction statement is used to guide the target large model to rewrite at least one target knowledge fragment to determine the answer information used to answer the target message.

[0116] Afterwards, the target processing system can input the target indication statement into the target macro model, thereby generating answer information according to the target indication statement through the target macro model.

[0117] In an optional embodiment, the preset trained large model can be fine-tuned to obtain the target large model. Optionally, the large model can be fine-tuned based on the second training sample set, the training samples in the second training sample set are sample knowledge fragments, and the true labels of the training samples can be the rewritten sample knowledge fragments.

[0118] For example, the above-mentioned retrieval-enhanced generation technology can be executed through an AI Agent instance to determine the process of answering information based on the knowledge base and the target message.

[0119] It should be noted that by rewriting at least one target knowledge fragment by the target big model to determine the answer information used to answer the target message, the humanization of the obtained answer message is improved, which can effectively improve the user experience and avoid directly determining at least one target knowledge fragment as an answer message, which will fix the answer message and make it easy for users to suspect that it is a robot reply, posing an answer risk.

[0120] In order to accurately determine the answer information, in the information processing method provided in Example 1 of the present application, after determining whether there is at least one target knowledge fragment matching the target message in the knowledge base that matches the action to be executed, if there is no at least one target knowledge fragment in the knowledge base that matches the action to be executed, the target message is input into the answer information generation model; and the answer information corresponding to the target message is output through the answer information generation model.

[0121] Optionally, when at least one target knowledge fragment does not exist in the knowledge base matched by the action to be executed, it is determined that the target message belongs to the rejected data of the target large model. In this case, if the target large model is still used to determine the answer information according to the target message, it will result in a relatively large probability that the determined answer information is irrelevant to the target message, or has a small correlation. Therefore, in this case, in order to ensure that the answer information can still be determined efficiently and accurately, the target processing system can input the target message into the answer information generation model.

[0122] Optionally, the answer message generation model may be a neural network model, and the answer message generation model may be trained using a third training sample set. The training samples in the third training sample set are sample messages of sample users, and the true labels corresponding to the training samples in the third training sample set are sample answer information.

[0123] Optionally, after the target message is input into the answer information generation model, the answer information generation model outputs the answer information corresponding to the target message.

[0124] For example, the above process of determining answer information when there is no at least one target knowledge fragment in the knowledge base matching the action to be executed can be performed by an AI Agent instance.

[0125] In an optional embodiment, it is possible to use Figure 3 The schematic diagram shown is used to determine the answer information. Figure 3 is a flowchart of an optional method for determining answer information provided in accordance with the first embodiment of the present application, such as Figure 3 As shown, after obtaining the target message, if the action to be performed is the second action belonging to the first category, the target processing system can use the similarity retrieval method to search for the target message in the knowledge base (i.e. Figure 3 The solution words in the solution and the historical replies of the manual customer service are used to perform knowledge recall, and then determine whether to reject the recognition based on the recall result. Optionally, if at least one target knowledge fragment is found based on the knowledge recall, it is determined that it is not rejected, and the at least one target knowledge fragment can be rewritten using the target large model to obtain answer information. If at least one target knowledge fragment is not found based on the knowledge recall, it is determined that it is rejected, and the answer information generation model is used to automatically generate the answer information.

[0126] It should be noted that by not using the target big model to determine the answer information when the target message belongs to the rejection data of the target big model, the target big model can be avoided from generating erroneous outputs. By determining the answer information based on the answer information generation model, on the one hand, the impact on the accuracy of the answer information is avoided as much as possible, and on the other hand, the efficiency of determining the answer information is guaranteed.

[0127] In order to accurately determine the answer information, in the information processing method provided in Example 1 of the present application, different second actions are used to determine the answer information according to different methods and / or different data, and at least one second action includes a second action belonging to the second category. The second action belonging to the second category refers to an action used to determine the answer information based on the robot answer information and the target message of the robot customer service service for the user. When the action to be executed is the second action, determining the answer information based on the second action and the target message includes: if the action to be executed is the second action belonging to the second category, determining the problem that matches the target message, wherein the problem that matches the target message is one of the following: a problem in the target message, a problem in the historical target message received by the target processing system; judging whether the robot answer information can solve the problem that matches the target message based on the degree of matching between the robot answer information and the problem that matches the target message, and obtaining a judgment result; determining the answer information based on the judgment result.

[0128] Optionally, when the action to be performed is a second action belonging to the second category, in the process of determining the problem that matches the target message, if the target message includes a question statement, the problem in the target message is determined as the problem that matches the target message; if the target message does not include a question statement, the problem that matches the target message is determined based on the problems in historical target messages received by the target processing system.

[0129] Optionally, the target message of the current user received by the target processing system in the manual customer service before receiving the current target message can be determined as the historical target message, and the question in the Rth to last historical target message including the question among the historical target messages received by the target processing system can be determined as the question matching the target message, where R is a positive integer.

[0130] For example, if the current target message does not include a question, but the previous target message in the current target message includes a question, then the question in the previous target message can be determined as the question that matches the current target message. For example, if the current target message is "yes", and the previous target message in the current target message is "why hasn't my order arrived yet", then "why hasn't my order arrived yet" is determined as the question that matches the target message "yes".

[0131] Optionally, if the current target message is the first message sent by the current user in the manual customer service, when the target processing system receives the current target message, it is determined that there is no target history message.

[0132] Therefore, for the first target message received by the target processing system in the manual customer service, when the target message is a question statement, the action to be executed may be the second action belonging to the second category; when the target message is not a question statement, the action to be executed is not the second action belonging to the second category.

[0133] Since users may have multiple rounds of conversations with robot customer service during the robot customer service service, that is, the robot customer service may generate multiple answer messages for the users, and users will request manual customer service only when they believe that the questions they raised have not been solved by the robot customer service, therefore, the answer information can be determined based on the robot answer information and target message of the robot customer service service for the user.

[0134] Optionally, if the target message is a question statement, and the action to be performed is a second action belonging to the second category, the target processing system may obtain robot answer information of the robot customer service service for the user.

[0135] Optionally, since the later the answer information of the robot customer service in the robot customer service service is, the more likely it is that the information is not recognized by the user, the Nth to last answer information for the user in the robot customer service service can be determined as the robot answer information, where N is a positive integer.

[0136] Since users may ask the same questions in manual customer service as in robot customer service, after obtaining the robot answer information of the robot customer service service for the user, the target processing system can judge whether the robot answer information can solve the problem matching the target message based on the degree of match between the robot answer information and the problem matching the target message, and obtain a judgment result.

[0137] For example, the target processing system has preset matching relationships between multiple questions and answer information in multiple robot customer service services. The target processing system can determine the matching degree between the above-mentioned robot answer information and the questions matching the target message based on the matching relationship between the multiple questions and the answer information in the multiple robot customer service services. For example, if the above-mentioned robot answer information and the questions matching the target message match in the above-mentioned matching relationship, it is determined that the matching degree between the robot answer information and the questions matching the target message is greater than the preset matching degree. If the above-mentioned robot answer information and the questions matching the target message do not match in the above-mentioned matching relationship, it is determined that the matching degree between the robot answer information and the questions matching the target message is less than or equal to the preset matching degree.

[0138] For another example, the target processing system may input the question of whether the robot's answer information matches the target message into a pre-trained matching degree recognition model, thereby determining the matching degree between the robot's answer information and the question of whether the target message matches through the matching degree recognition model. Optionally, the matching degree recognition model may be a neural network model, and the training samples used to train the matching degree recognition model may be question-answer pairs consisting of sample messages of sample users and sample answer information of sample robot customer service, and the real labels used to train the matching degree recognition model are used to characterize whether the question and answer in the aforementioned question-answer pairs match.

[0139] Optionally, when the matching degree between the robot's answer information and the problem of the target message matching is less than or equal to a preset matching degree, it is determined that the robot's answer information cannot solve the problem of the target message matching, and when the matching degree between the robot's answer information and the problem of the target message matching is greater than a preset matching degree, it is determined that the robot's answer information can solve the problem of the target message matching, thereby obtaining a judgment result. That is, the judgment result is used to characterize whether the robot's answer information can solve the problem of the target message matching.

[0140] Optionally, after obtaining the judgment result, the answer information is determined according to the judgment result. In an optional embodiment, if the judgment result indicates that the robot answer information can solve the problem of matching the target message, the answer information for the target message is determined according to the robot answer information. For example, the robot answer information is determined as the answer information for the target message, and for another example, the modified robot answer information is determined as the answer information for the target message; for another example, the manual answer information corresponding to the robot answer information is determined as the answer information for the target message.

[0141] In an optional embodiment, if the judgment result indicates that the robot's answer information cannot solve the problem of matching the target message, the answer information for the target message is not determined based on the robot's answer information, and the answer information for the target message can be determined based on the answer information generation script.

[0142] For example, the above process of determining the answer information based on the second action belonging to the second category can be performed by the AI ​​Agent instance. In this process, the AI ​​Agent instance can obtain the above robot answer information of the robot customer service service for the user from the long-term memory.

[0143] It should be noted that since users may ask the same questions in manual customer service as in robot customer service, by setting the second action belonging to the second category to determine the answer information based on the robot answer information and target message of the robot customer service service for the user, it is possible to determine the answer information for the user's current message in combination with the historical conversation between the robot customer service and the user. By judging whether the robot answer information can solve the problem of matching the target message, and determining the answer information based on the judgment result, it is possible to determine the answer information based on the correlation between the historical conversation between the robot customer service and the user and the user's current message, thereby improving the accuracy of determining the answer information.

[0144] In order to improve the anthropomorphism of the answer information, in the information processing method provided in the first embodiment of the present application, determining the answer information according to the judgment result includes: when the robot answer information can solve the problem matching the target message, obtaining manual answer information matching the robot answer information; determining the manual answer information matching the robot answer information as the answer information; when the robot answer information cannot solve the problem corresponding to the target message, determining an answer information generation script matching the problem matching the target message; and generating the answer information according to the answer information generation script.

[0145] Since the robot's answer information can solve the problem of matching the target message, if the robot's answer information is directly determined as the answer information, it is easy for the user to recognize that it is the robot customer service who is answering the question. Therefore, the matching relationship between the answer information in multiple robot customer service services and multiple manual answer information can be preset in the target processing system, so that the target processing system can determine the manual answer information that matches the robot's answer information based on the matching relationship.

[0146] Optionally, the manual answer information matching the robot answer information is determined as the answer information. Optionally, the manual answer information matching the robot answer information and the robot answer information solve the same problem, but the text content is different.

[0147] In the case where the robot's answer information cannot solve the problem corresponding to the target message, if the answer information for the target message is determined based on the robot's answer information, it will result in a relatively low correlation between the answer information and the target message, that is, an incorrect answer. Therefore, a matching relationship between multiple questions and multiple answer information generation scripts may be preset in the target processing system. The target processing system may determine the answer information generation script that matches the question matching the target message based on the matching relationship, thereby generating answer information based on the answer information generation script that matches the question matching the target message.

[0148] In an optional embodiment, the answer information generation script can be used to generate answer information based on the user's order information. For example, the answer information generation script is used to extract order feature information from the user's order information, and then import the order feature information into the answer information template that matches the answer information generation script, thereby obtaining answer information for the target message.

[0149] Optionally, the above-mentioned order characteristic information includes but is not limited to the order number, the logistics progress in the order, the transaction amount in the order, etc.

[0150] Optionally, the answer information template that matches the answer information generation script may be directly built into the answer information generation script, or the storage path of the answer information template may be built into the answer information generation script.

[0151] For example, an optional answer information template may be "Because the logistics progress of your order is B, the refund failed", where "B" in the aforementioned answer information template is used to fill in the order characteristic information such as the logistics progress in the order.

[0152] In an optional embodiment, the answer information generation script can also be used to obtain preset answer information from a storage area defined in the answer information generation script, and then determine the preset answer information as the answer information for the target message.

[0153] For example, the above process of determining the answer information based on the judgment result can be executed by the AI ​​Agent instance.

[0154] It should be noted that, by determining the manual answer information that matches the robot's answer information as the answer information when the robot's answer information can solve the problem of matching the target message, the personification of the answer information is improved, and the problems such as the stiff tone that exists when the robot's answer information is directly determined as the answer information for the target message are avoided. By determining the answer information based on the answer information generation script when the robot's answer information cannot solve the problem corresponding to the target message, the accuracy of the determined answer information is improved.

[0155] In order to achieve a stable hosting stage, in the information processing method provided in Example 1 of the present application, after the answer information is determined according to the target message by the target processing system, the answer information is fed back to the user; if the next target message sent by the user after the target message is received, the target action to be executed is determined from multiple actions according to the next target message through the action decision model, and the target action is executed, wherein the multiple actions include a first action and at least one second action, the first action refers to an action for switching to manual customer service to answer the target message, and the second action refers to an action for determining the answer information according to the target message.

[0156] Optionally, after the target processing system starts the manual customer service for the current user, it is equivalent to entering the manual customer service stage. Therefore, after determining the answer information, the target processing system can imitate the manual customer service to feed back the answer information to the user. For example, after the target processing system starts the manual customer service service, the customer service information of the manual customer service is displayed in the customer service interface in the terminal held by the user, and after determining the answer information, the answer information is displayed in the chat box of the manual customer service in the customer service interface, so that the user believes that the answer information is determined by the manual customer service.

[0157] Optionally, after the answer information is fed back to the user, the user's problem may not be solved at one time, so the user may send a new target message, that is, the next target message mentioned above.

[0158] Optionally, if the target processing system receives a next target message sent by the user after the target message, the target action to be executed can be determined from multiple actions according to the next target message through the action decision model.

[0159] Optionally, in the process of determining the target action to be executed from multiple actions according to the next target message through the action decision model, the next target message and the actions that have been executed by the target processing system in the manual customer service are input into the action decision model, so that the action decision model makes an action decision among multiple actions according to the next target message and the actions that have been executed by the target processing system in the manual customer service, and outputs the action to be executed.

[0160] That is, in manual customer service, if it is predicted that the target message of the user to be received by the target processing system belongs to the target message type, then when the target processing system receives the target message, the target processing system can first determine whether the target processing system has performed an action among multiple actions in the manual customer service, so that if the target processing system has not performed an action among multiple actions, the action to be performed is determined from multiple actions according to the target message through the action decision model; if the target processing system has performed the action, the action to be performed is determined from multiple actions according to the target message and the actions that have been performed by the target processing system in the manual customer service through the action decision model.

[0161] For example, when the target processing system receives the target message of the current user for the first time, since the target processing system has not performed any action, the action decision model only makes an action decision among multiple actions based on the target message and outputs the action to be executed.

[0162] For another example, suppose Q is a positive integer greater than 1. When the target processing system receives the target message of the current user for the Qth time, since the target processing system has performed an action before, the action decision model makes an action decision among multiple actions based on the target message and the actions that have been performed by the target processing system in the manual customer service, and outputs the action to be executed.

[0163] Optionally, in order to enable the action decision model to effectively combine the aforementioned executed actions to make action decisions when making action decisions, in an optional embodiment, the action decision model is a large language model, and the target processing system can preset an instruction statement in the action decision model. The instruction statement can be a prompt. The instruction statement is used to guide the target processing system to determine the action to be executed based on the target rule, target message and executed actions. The target rule refers to the rule for determining the action to be executed based on the target message and the action already executed by the target processing system in the manual customer service, so that the action decision model can accurately output the action to be executed based on the guidance of the preset instruction statement.

[0164] For example, assuming that the multiple actions include action A, action B, action C, and action D, then an optional target rule may be "If the target processing system has executed action A, then determine the action to be executed from actions B, C, and D based on the target message", or, an optional target rule may be "If the target processing system has executed action A three times, then determine action D as the action to be executed", or other content.

[0165] For example, in the above process, the action decision model is called by the AI ​​Agent instance to determine the above target action to be executed, and the AI ​​Agent instance can determine the actions that the AI ​​Agent instance has executed from the short-term memory.

[0166] Optionally, when the target action to be executed is the first action, the target processing system may stop determining the answer information according to the next target message and send a service request to the manual customer service to complete the execution of the target action to be executed. The service request is used to request the manual customer service to answer the target message.

[0167] For example, when the AI ​​Agent instance determines that the target action to be executed output by the action decision model is the first action, it can stop calling the action decision model and send a service request to the manual customer service. Then, after sending the service request to the manual customer service, the AI ​​Agent instance can be closed.

[0168] Optionally, when the target action to be executed is the second action, the target processing system can determine the response information based on the second action and the next target message to complete the execution of the target action to be executed. Optionally, the process of the target processing system executing the target action to be executed is the same as the process of the target processing system executing the above-mentioned action to be executed, so it is not repeated here.

[0169] For example, the second action is performed by the AI ​​Agent instance based on the next target message.

[0170] It should be noted that, through the above process, when the user and the target processing system have a continuous dialogue, the target processing system can continuously and stably implement the hosting stage, thereby further reducing labor costs.

[0171] In order to reduce the risk of users identifying the answers as non-manual answers during the manual customer service stage, in the information processing method provided in Example 1 of the present application, after the target processing system receives the target message, when an instruction to transfer to manual customer service is received from manual customer service, the target processing system is controlled to stop determining the answer information based on the target message; when an instruction to transfer to manual customer service is not received from manual customer service, it is determined based on the target message whether to switch to manual customer service to answer the target message; if it is determined to switch to manual customer service to answer the target message, the target processing system is controlled to stop determining the answer information based on the target message, and a service request is sent to the manual customer service, wherein the service request is used to request the manual customer service to answer the target message.

[0172] Optionally, before the target processing system predicts whether the target message of the user to be received by the target processing system belongs to the target message type based on at least one preset access rule, the target processing system can pre-determine a manual customer service representative for providing customer service to the user, that is, pre-assign a manual customer service representative to the user.

[0173] Afterwards, if it is predicted that the target message of the user to be received by the target processing system belongs to the target message type, the manual customer service will not process the target message of the user, and the target processing system will determine the answer information according to the target message. However, during the operation of the target processing system, the manual customer service can monitor the conversation between the target processing system and the user, and actively command the target processing system to stop answering when it is identified that the user may have negative emotions or the target processing system may not be able to answer accurately, that is, send a transfer instruction to the manual customer service to the target processing system so that the manual customer service can answer by itself.

[0174] Optionally, the target processing system may stop determining the answer information according to the target message upon receiving an instruction to transfer to the manual customer service sent by the manual customer service.

[0175] For example, the above-mentioned process of controlling the target processing system to stop determining the answer information according to the target message can be implemented through an AI Agent instance, and the AI ​​Agent instance can be closed after stopping determining the answer information according to the target message. Optionally, the above-mentioned process of determining whether to switch to manual customer service to answer the target message can also be implemented through an AI Agent instance, and the AI ​​Agent instance can be closed after sending a service request to the manual customer service.

[0176] Optionally, the target processing system may also monitor the conversation between it and the user, that is, determine whether to switch to manual customer service to answer the target message based on the target message.

[0177] For example, the target processing system may determine to switch to manual customer service to answer the target message if it is determined based on the target message that the user has negative emotions. For another example, the target processing system may determine to switch to manual customer service to answer the target message if it is determined based on the target message that the user questions the customer service identity of the manual customer service. For another example, the target processing system may determine to switch to manual customer service to answer the target message if the number of rounds of the user sending the target message exceeds a threshold.

[0178] Optionally, if it is determined to switch to manual customer service to answer the target message, the target processing system can control itself to stop determining the answer information based on the target message, and send a service request to the manual customer service, wherein the service request is used to request the manual customer service to answer the target message.

[0179] For example, the above process of determining whether to switch to manual customer service to answer the target message based on the target message can be implemented through an AI Agent instance.

[0180] It should be noted that through the above process, on the one hand, it is realized that in the intelligent hosting stage, the manual customer service acts as a supervisor and can intervene in the conversation at any time, thereby reducing the risk of being identified as non-human work by users and improving user experience. On the other hand, it is realized that the target processing system can actively and directly end the hosting for risky sessions and remind the manual customer service to continue the service, thereby further reducing the risk of being identified as non-human work by users and improving user experience.

[0181] In order to accurately determine whether it is necessary to switch to manual customer service to answer the target message, in the information processing method provided in Example 1 of the present application, judging whether to switch to manual customer service to answer the target message according to the target message includes: judging whether there is a target keyword in the target message to obtain a first judgment result, wherein the target keyword is a word that represents the user's questioning of the customer service identity of the manual customer service; judging whether the target message belongs to a message under a target scenario to obtain a second judgment result, wherein the target scenario is a scenario that represents the user's negative emotions; judging whether the conversation round to which the target message belongs is greater than a preset round to obtain a third judgment result; and determining whether to switch to manual customer service to answer the target message based on at least one of the first judgment result, the second judgment result, and the third judgment result.

[0182] Optionally, the target processing system may be preset with a questioning identity rule, which includes words used by users to question the customer service identity of the manual customer service service, that is, includes the target keyword. The target processing system may determine whether the target keyword exists in the target message based on the questioning identity rule to obtain a first determination result.

[0183] For example, assuming that the target keyword is "robot", when the target message is "You are a robot, right?", it is determined that the first judgment result indicates that the target keyword exists in the target message.

[0184] Optionally, a scene recognition model may be preset in the target processing system, and the target processing system may input the target message into the scene recognition model to identify whether the target message belongs to the target scene through the scene recognition model. Optionally, the scene recognition model may be a neural network model, and the scene recognition model may be obtained by training through the fourth training sample set, and the training samples in the fourth training sample set may be sample messages of sample users, and the true labels of the training samples in the fourth training sample set may be used to characterize whether the sample message belongs to the target scene.

[0185] Optionally, the target processing system may also preset a vocabulary corresponding to the target scenario, so that when there are words in the target message that fall into the vocabulary corresponding to the target scenario, it is determined that the target message belongs to the target scenario; when there are no words in the target message that fall into the vocabulary corresponding to the target scenario, it is determined that the target message does not belong to the target scenario.

[0186] For example, assuming that the vocabulary corresponding to the target scenario includes words such as "315" and "angry", then when the target message is "I am very angry" or "I am going to call 315", it is determined that the second judgment result represents that the target message belongs to the target scenario.

[0187] Optionally, when receiving a target message, the target processing system may count the conversation turns between it and the user. For example, the conversation turns may be determined based on the number of target messages received from the user, that is, the conversation turns are updated after receiving the target message from the user. For another example, the conversation turns may be determined based on the number of answer messages generated by the target processing system, that is, the conversation turns are updated after the answer information is fed back to the user.

[0188] Optionally, after determining the conversation turn to which the target message belongs, the target processing system may determine whether the conversation turn to which the target message belongs is greater than a preset turn to obtain a third determination result.

[0189] In an optional embodiment, the target processing system may determine whether to switch to manual customer service to answer the target message based on at least one of the first judgment result, the second judgment result, and the third judgment result.

[0190] For example, if the target processing system determines whether to switch to manual customer service to answer the target message based on the first judgment result, the second judgment result, and the third judgment result, the target processing system may determine not to switch to manual customer service to answer the target message when the first judgment result indicates that the target keyword does not exist in the target message, and the second judgment result indicates that the target message does not belong to the message under the target scenario, and the third judgment result indicates that the conversation turn to which the target message belongs is less than or equal to the set turn.

[0191] When the first judgment result indicates that the target keyword exists in the target message, or the second judgment result indicates that the target message belongs to a message under the target scenario, or the third judgment result indicates that the conversation turn to which the target message belongs is greater than the preset turn, it is determined to switch to manual customer service to answer the target message.

[0192] It should be noted that, through the above judgment, it is possible to judge whether the user questions the identity of the customer service, whether the user is emotionally agitated, and whether the user's problem is difficult to solve for a long time. In other words, it is possible to judge whether there is a risk that the user may identify the answer as non-human. By determining whether to switch to manual customer service to answer the target message based on the above judgment results, it is possible to accurately determine the timing of switching to manual customer service to answer, thereby reducing the risk of users identifying non-human answers and avoiding affecting the user experience.

[0193] In order to improve the standardization of answer information, in the information processing method provided in Example 1 of the present application, after the answer information is determined according to the target message by the target processing system, the answer information is quality-checked according to at least one preset quality inspection rule; if the answer information passes the quality inspection, the answer information is fed back to the user.

[0194] Optionally, the quality inspection rules are used to perform quality inspection on the standardization of the answer information, and the quality inspection rules include standardization conditions that the answer information needs to meet.

[0195] For example, the quality inspection rule may be "the answer information cannot include the following sensitive words: A, B, C". Another example is that the quality inspection rule may be "the answer information is subjected to emotion recognition using an emotion recognition model, and the recognition result cannot indicate that the answer information contains negative emotions".

[0196] Optionally, when it is determined that the answer information meets at least one normative condition in the quality inspection rule, it is determined that the answer information has passed the quality inspection; when it is determined that there are normative conditions in the quality inspection rule that the answer information does not meet, it is determined that the answer information has failed the quality inspection.

[0197] Optionally, if the answer information passes the quality check, the answer information will be fed back to the user; if the answer information fails the quality check, the answer information may not be sent to the user, and a service request may be sent to the manual customer service, wherein the service request is used to request the manual customer service to answer the target message.

[0198] It should be noted that by conducting additional quality checks on the answer information before feeding it back to the user, the standardization of the answer information can be effectively improved, that is, the effectiveness of the answer information can be improved, thereby facilitating improving user experience and reducing answer risks.

[0199] In order to further improve the standardization of the answer information, in the information processing method provided in the first embodiment of the present application, before the target processing system starts the manual customer service, the target processing system can use the quality inspection model to perform adversarial simulation testing to verify the standardization of the answer information determined by the target processing system. Optionally, the quality inspection model is used to perform a standardization quality inspection on the answer information determined by the target processing system according to at least one of the aforementioned preset quality inspection rules.

[0200] It should be noted that by performing quality inspection on the answer information determined by the target processing system before the target processing system starts the manual customer service, the answer capability of the target processing system can be verified before the target processing system is applied, thereby further improving the effectiveness of the answer information determined by the target processing system during actual application.

[0201] In an optional embodiment, it is possible to use Figure 4 The schematic diagram shown realizes intelligent hosting in the manual customer service stage. Figure 4 This is the process of the optional information processing method provided in Example 1 of the present application. Figure 2 ,like Figure 4As shown, the target processing system may include a channel module, a hosting module, an AI customer service module, an action decision model, and a customer service assistant module. Optionally, when a user needs manual service, he or she may send a request to transfer to manual service to the target processing system. After receiving the request to transfer to manual service, the channel module creates a session of manual customer service for the user and assigns manual customer service to the user. After that, the hosting module triggers the hosting detection. The hosting module may predict whether the target message to be received by the user belongs to the target message type according to at least one preset access rule, thereby determining access when it is predicted that the target message to be received by the user belongs to the target message type, that is, allowing entry into the intelligent hosting stage.

[0202] Alternatively, if Figure 4 As shown, after the admission is determined, the AI ​​customer service module can be initialized, that is, an AI Agent instance for the current user and the Memory corresponding to the AI ​​Agent instance are created. Afterwards, the Agent is triggered, that is, the target message sent by the user through the channel module is received by the AIAgent instance, and the AI ​​Agent instance performs a status check on the working status of the AI ​​Agent instance. If it is detected that the AI ​​Agent instance is in an abnormal working state, it stops determining the answer information according to the target message, and sends a service request to the manual customer service, wherein the service request is used to request the manual customer service to answer the target message; if it is detected that the AI ​​Agent instance is not in an abnormal working state, the AI ​​Agent instance makes an action decision. In the action decision process, the AI ​​Agent instance determines the action to be executed by calling the action decision model, and then the AI ​​Agent instance executes the action to be executed. Optionally, as Figure 4 As shown, during the working process of the AI ​​Agent instance, the AI ​​Agent instance can update the Memory based on the operation information during its working process.

[0203] Alternatively, if Figure 4 As shown, the AI ​​Agent instance can implement the execution of the action to be executed through the customer service assistant model. If the action to be executed is the second action, after the action to be executed is completed, the answer information for the user's target message will be obtained. In this case, the AI ​​Agent instance can perform quality inspection on the obtained answer information, and then the channel module will feed back the answer information to the user if it is determined that the answer information has passed the quality inspection.

[0204] Alternatively, if Figure 4 As shown, in an optional embodiment, the multiple actions include at least a problem confirmation action, a solution execution action, a solution explanation action, a solution questioning action, and a transfer to manual service action ( Figure 4(not shown in the figure). The transfer to manual service action is equivalent to the first action, and the problem confirmation action, solution execution action, solution explanation action and solution follow-up action are equivalent to the second action. The problem confirmation action can be understood as the second action belonging to the third category mentioned above, the problem confirmation action can be understood as the second action belonging to the second category mentioned above, and the solution explanation action and solution follow-up action can be understood as the second action belonging to the first category mentioned above.

[0205] For example, suppose that the user sends the message "Why did the refund of my order fail?" in the robot customer service service. Because the user is not satisfied with the answer of the robot customer service to the message, he requests to enter the manual customer service service. Then, the first target message sent in the manual customer service service is "Hello". Then the action to be executed determined by the action decision model may be the problem confirmation action. The problem confirmation action can be understood as being triggered when the user's problem needs to be determined. Therefore, executing this action (based on the above-mentioned method of executing the second action belonging to the third category) may output the answer information "Are you asking the question 'Why did the refund of my order fail?'" and feedback it to the user. Afterwards, the user may send the target message "Yes". In this case, the action to be executed determined by the action decision model may be the solution execution action. The solution execution action can be understood as being triggered when the solution to the user's problem needs to be determined. Therefore, executing this action (based on the above-mentioned method of executing the second action belonging to the second category) may output the answer information "Because the logistics progress of your order is signed for, the refund failed" and feedback it to the user.

[0206] Afterwards, the user may ask further questions about the aforementioned solution (i.e., the aforementioned answer information). For example, the next target message is "Why did the refund fail after signing for it?" In this case, the action to be executed determined by the action decision model may be a solution explanation action. The solution explanation action can be understood as being triggered when the solution needs to be explained. Executing this action (based on the above-mentioned method of executing the second action belonging to the first category) may output the answer information "Because the refund can only be made after the goods are returned", and feedback it to the user. Afterwards, the user may think that the answer information solves his problem, but there are still some other questions. For example, the next target message is "How many days after the goods are returned at the latest?" In this case, the action to be executed determined by the action decision model may be a solution questioning action. The solution questioning action can be understood as being triggered when it is necessary to answer questions that are less related to the solution. Executing this action may output the answer information "Are you asking the question 'How many days after the goods are returned at the latest after signing for it'?" and feedback it to the user. In this way, the intelligent hosting of the entire session in the manual customer service can be realized, thereby effectively reducing labor costs. Optionally, the method provided in the present application can achieve a full hosting completion rate of more than 50% within a given range of scenarios.

[0207] In an optional embodiment, it is possible to use Figure 5 The schematic diagram shown in the figure implements the detection of the end of custody. Figure 5 This is a flowchart of the custody end detection provided in Example 1 of the present application, such as Figure 5 As shown, the AI ​​Agent instance can determine whether there is a risk of answering based on the target message (that is, the above-mentioned determination of whether to switch to manual customer service to answer the target message based on the target message). If it is determined that there is a risk of answering, a service request is sent to the manual customer service, and then the AI ​​Agent instance is closed to end the hosting. Optionally, the manual customer service can also actively send a transfer instruction to the manual customer service to the hosting module, and then the hosting module ends the hosting when receiving the transfer instruction to the manual customer service. Optionally, in the manual customer service service, the hosting module can also perform disconnection detection. For example, if the user has not sent the target message for a long time, it is determined that the disconnection hosting is triggered. In this case, the hosting is ended and the channel module ends the session.

[0208] In an embodiment of the present application, when the target processing system starts the manual customer service service, it is predicted whether the target message of the user to be received by the target processing system belongs to the target message type according to at least one preset access rule, wherein the access rule represents the conditions that the target message must meet for it to belong to the target message type, and the target message type is a message type that supports the target processing system to replace manual answers; if it is predicted that the target message of the user to be received by the target processing system belongs to the target message type, then when the target processing system receives the target message, the target processing system determines the answer information according to the target message. When the target processing system starts the manual customer service service, it is predicted whether the target message of the user to be received by the target processing system belongs to the target message type according to at least one preset access rule, thereby achieving effective prediction of the messages to be received that can be answered by the system instead of manually; if it is predicted that the target message to be received belongs to the target message type, then when the target message is received, the target processing system determines the answer information according to the target message, thereby achieving the system replacing manual answers to user messages in the manual customer service stage, thereby effectively carrying a part of the manual customer service service volume and reducing the manual customer service cost. The purpose of using the system to answer user messages that meet certain conditions in the manual customer service stage is achieved, and the technical effect of reducing the cost of manual customer service is realized, thereby solving the technical problem in the related technology of directly relying on manual answers to all messages sent by consumers in the manual customer service stage, resulting in high labor costs.

[0209] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present application.

[0210] Through the description of the above implementation methods, those skilled in the art can clearly understand that the method according to the above embodiment can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, computer, server, or network device, etc.) to execute the methods of each embodiment of the present application.

[0211] Example 2

[0212] According to an embodiment of the present application, an information processing device for implementing the above information processing method is also provided. Figure 6 As shown, the device includes: a prediction unit 601 and a determination unit 602.

[0213] The prediction unit 601 is used to predict whether the target message of the user to be received by the target processing system belongs to the target message type according to at least one preset admission rule when the target processing system starts the manual customer service, wherein the admission rule represents the conditions that the target message must meet for it to belong to the target message type, and the target message type is a message type that supports the target processing system to replace the manual answer;

[0214] The determination unit 602 is used to determine the reply information according to the target message through the target processing system when the target processing system receives the target message if it is predicted that the target message of the user to be received by the target processing system belongs to the target message type.

[0215] In the information processing device provided in the second embodiment of the present application, when the target processing system starts the manual customer service, the prediction unit 601 predicts whether the target message of the user to be received by the target processing system belongs to the target message type according to at least one preset access rule, wherein the access rule represents the conditions that the target message must meet to belong to the target message type, and the target message type is a message type that supports the target processing system to replace manual answers; if the determination unit 602 predicts that the target message of the user to be received by the target processing system belongs to the target message type, then when the target processing system receives the target message, the target processing system determines the answer information according to the target message. In this scheme, when the target processing system starts the manual customer service, it predicts whether the target message of the user to be received by the target processing system belongs to the target message type according to at least one preset access rule, thereby realizing the effective prediction of the message to be received that can be replaced by the system instead of manual answers. If it is predicted that the target message to be received belongs to the target message type, then when the target message is received, the target processing system determines the answer information according to the target message, thereby realizing that the system replaces manual answers to user messages in the manual customer service stage, thereby effectively carrying a part of the service volume of manual customer service and reducing the service cost of manual customer service. The purpose of using the system to answer user messages that meet certain conditions in the manual customer service stage is achieved, and the technical effect of reducing the cost of manual customer service is realized, thereby solving the technical problem in the related technology of directly relying on manual answers to all messages sent by consumers in the manual customer service stage, resulting in high labor costs.

[0216] Optionally, in the information processing device provided in Example 2 of the present application, the determination unit includes: a first determination sub-unit, used to determine the action to be executed from multiple actions according to the target message through the action decision model in the target processing system, wherein the multiple actions include a first action and at least one second action, the first action refers to an action used to switch to manual customer service to answer the target message, and the second action refers to an action used to determine the answer information according to the target message; a second determination sub-unit, used to determine the answer information based on the second action and the target message when the action to be executed is the second action.

[0217] Optionally, in the information processing device provided in Example 2 of the present application, different second actions are used to determine answer information based on different methods and / or different data, and at least one second action includes a second action belonging to the first category. The second action belonging to the first category refers to an action used to determine the answer information based on the retrieval enhancement generation technology and the target message. The second determination subunit includes: a first determination module, which is used to determine the knowledge base that matches the action to be executed if the action to be executed is the second action belonging to the first category; and a second determination module, which is used to determine the answer information based on the knowledge base and the target message based on the retrieval enhancement generation technology.

[0218] Optionally, in the information processing device provided in Example 2 of the present application, the second determination module includes: a judgment submodule, used to judge whether there is at least one target knowledge fragment matching the target message in the knowledge base that matches the action to be executed based on the target message; a first generation submodule, used to generate a target indication statement based on the at least one target knowledge fragment and the target message when there is at least one target knowledge fragment in the knowledge base that matches the action to be executed, wherein the target indication statement is used to guide the target big model to rewrite the at least one target knowledge fragment to determine the answer information used to answer the target message; and a second generation submodule, used to generate answer information based on the target indication statement through the target big model.

[0219] Optionally, in the information processing device provided in Example 2 of the present application, the information processing device also includes: a first processing unit, used to input the target message into the answer information generation model when there is no at least one target knowledge fragment in the knowledge base matching the action to be executed; and a second processing unit, used to output the answer information corresponding to the target message through the answer information generation model.

[0220] Optionally, in the information processing device provided in Example 2 of the present application, different second actions are used to determine answer information based on different methods and / or different data, and at least one second action includes a second action belonging to the second category. The second action belonging to the second category refers to an action used to determine the answer information based on the robot answer information and the target message of the robot customer service service for the user. The second determination subunit includes: an acquisition module, which is used to determine the problem matched by the target message if the action to be executed is the second action belonging to the second category, wherein the problem matched by the target message is one of the following: a problem in the target message, a problem in the historical target message received by the target processing system; a judgment module, which is used to judge whether the robot answer information can solve the problem matched by the target message according to the degree of match between the robot answer information and the problem matched by the target message, and obtain a judgment result; a third determination module, which is used to determine the answer information based on the judgment result.

[0221] Optionally, in the information processing device provided in Example 2 of the present application, the third determination module includes: an acquisition sub-module, used to acquire manual answer information that matches the robot answer information when the robot answer information can solve the problem matching the target message; a first determination sub-module, used to determine the manual answer information that matches the robot answer information as the answer information; a second determination sub-module, used to determine an answer information generation script that matches the problem matching the target message when the robot answer information cannot solve the problem corresponding to the target message; and a third generation sub-module, used to generate answer information according to the answer information generation script.

[0222] Optionally, in the information processing device provided in Example 2 of the present application, the information processing device also includes: a third processing unit, used to control the target processing system to stop determining answer information based on the target message when receiving an instruction to transfer to manual customer service sent by manual customer service; a judgment unit, used to judge whether to switch to manual customer service to answer the target message based on the target message when not receiving an instruction to transfer to manual customer service sent by manual customer service; a fourth processing unit, used to control the target processing system to stop determining answer information based on the target message if it is determined to switch to manual customer service to answer the target message, and send a service request to the manual customer service, wherein the service request is used to request the manual customer service to answer the target message.

[0223] Optionally, in the information processing device provided in Example 2 of the present application, the judgment unit includes: a first judgment sub-unit, used to judge whether there is a target keyword in the target message, and obtain a first judgment result, wherein the target keyword is a word that characterizes the user's questioning of the customer service identity of the manual customer service service; a second judgment sub-unit, used to judge whether the target message belongs to a message under a target scenario, and obtain a second judgment result, wherein the target scenario is a scenario that characterizes the user's negative emotions; a third judgment sub-unit, used to judge whether the dialogue turn to which the target message belongs is greater than a preset round, and obtain a third judgment result; a third determination sub-unit, used to determine whether to switch to manual customer service to answer the target message based on at least one of the first judgment result, the second judgment result, and the third judgment result.

[0224] Optionally, in the information processing device provided in Example 2 of the present application, the information processing device also includes: a quality inspection unit, used to perform quality inspection on the answer information according to at least one preset quality inspection rule; and a fifth processing unit, used to feed back the answer information to the user if the answer information passes the quality inspection.

[0225] Optionally, in the information processing device provided in Example 2 of the present application, the information processing device also includes: a sixth processing unit, used to feed back answer information to the user; a seventh processing unit, used to determine the target action to be executed from multiple actions according to the next target message through an action decision model if the next target message sent by the user after the target message is received, and execute the target action, wherein the multiple actions include a first action and at least one second action, the first action refers to an action for switching to manual customer service to answer the target message, and the second action refers to an action for determining the answer information based on the target message.

[0226] It should be noted that the above-mentioned prediction unit 601 and determination unit 602 correspond to steps S201 to S202 in Example 1, and the above-mentioned units and corresponding steps implement the same examples and application scenarios, but are not limited to the contents disclosed in the above-mentioned Example 1. It should be noted that the above-mentioned modules as part of the device can be run in the computer terminal 10 provided in Example 1.

[0227] It should be noted that the preferred implementation scheme involved in the above embodiments of the present application is the same as the scheme provided in Example 1, as well as the application scenario and implementation process, but is not limited to the scheme provided in Example 1.

[0228] Example 3

[0229] The embodiment of the present application may provide an electronic device, which may be any electronic device in a group of electronic devices. Optionally, in this embodiment, the electronic device may also be replaced by a terminal device such as a mobile terminal.

[0230] Optionally, in this embodiment, the electronic device may be located in at least one network device among a plurality of network devices of a computer network.

[0231] In this embodiment, the above-mentioned electronic device can execute the program code of the following steps in the information processing method: when the target processing system starts the manual customer service, predict whether the target message of the user to be received by the target processing system belongs to the target message type according to at least one preset access rule, wherein the access rule represents the conditions that must be met for the target message to belong to the target message type, and the target message type is a message type that supports the target processing system to replace manual answers; if it is predicted that the target message of the user to be received by the target processing system belongs to the target message type, then when the target processing system receives the target message, the target processing system determines the answer information according to the target message.

[0232] The above-mentioned electronic device can also execute the program code of the following steps in the information processing method: determine the action to be executed from multiple actions according to the target message through the action decision model in the target processing system, wherein the multiple actions include a first action and at least one second action, the first action refers to an action used to switch to manual customer service to answer the target message, and the second action refers to an action used to determine the answer information according to the target message; when the action to be executed is the second action, the answer information is determined based on the second action and the target message.

[0233] The electronic device can also execute the program code of the following steps in the information processing method: if the action to be executed is the second action belonging to the first category, determine the knowledge base that matches the action to be executed; based on the retrieval enhancement generation technology, determine the answer information according to the knowledge base and the target message.

[0234] The above-mentioned electronic device can also execute the program code of the following steps in the information processing method: based on the target message, determine whether there is at least one target knowledge fragment matching the target message in the knowledge base matching the action to be executed; when there is at least one target knowledge fragment in the knowledge base matching the action to be executed, generate a target indication statement based on the at least one target knowledge fragment and the target message, wherein the target indication statement is used to guide the target big model to rewrite the at least one target knowledge fragment to determine the answer information used to answer the target message; generate the answer information according to the target indication statement through the target big model.

[0235] The above-mentioned electronic device can also execute the program code of the following steps in the information processing method: after determining whether there is at least one target knowledge fragment matching the target message in the knowledge base matching the action to be executed, if there is no at least one target knowledge fragment in the knowledge base matching the action to be executed, input the target message into the answer information generation model; and output the answer information corresponding to the target message through the answer information generation model.

[0236] The above-mentioned electronic device can also execute the program code of the following steps in the information processing method: if the action to be executed is a second action belonging to the second category, then determine the problem that matches the target message, wherein the problem that matches the target message is one of the following: a problem in the target message, a problem in the historical target message received by the target processing system; based on the degree of match between the robot's answer information and the problem that matches the target message, judge whether the robot's answer information can solve the problem that matches the target message, and obtain a judgment result; determine the answer information based on the judgment result.

[0237] The above-mentioned electronic device can also execute the program code of the following steps in the information processing method: when the robot answer information can solve the problem matching the target message, obtain the manual answer information matching the robot answer information; determine the manual answer information matching the robot answer information as the answer information; when the robot answer information cannot solve the problem corresponding to the target message, determine the answer information generation script matching the problem matching the target message; generate the answer information according to the answer information generation script.

[0238] The above-mentioned electronic device can also execute the program code of the following steps in the information processing method: after the target processing system receives the target message, when an instruction to transfer to manual customer service is received from manual customer service, the target processing system is controlled to stop determining the answer information based on the target message; when an instruction to transfer to manual customer service is not received from manual customer service, whether to switch to manual customer service to answer the target message is determined based on the target message; if it is determined to switch to manual customer service to answer the target message, the target processing system is controlled to stop determining the answer information based on the target message, and a service request is sent to the manual customer service, wherein the service request is used to request the manual customer service to answer the target message.

[0239] The above-mentioned electronic device can also execute the program code of the following steps in the information processing method: determine whether there is a target keyword in the target message to obtain a first judgment result, wherein the target keyword is a word that represents the user's questioning of the customer service identity of the manual customer service service; determine whether the target message belongs to a message under a target scenario to obtain a second judgment result, wherein the target scenario is a scenario that represents the user's negative emotions; determine whether the conversation round to which the target message belongs is greater than a preset round to obtain a third judgment result; determine whether to switch to manual customer service to answer the target message based on at least one of the first judgment result, the second judgment result, and the third judgment result.

[0240] The above-mentioned electronic device can also execute the program code of the following steps in the information processing method: after determining the reply information according to the target message through the target processing system, the reply information is quality checked according to at least one preset quality inspection rule; if the reply information passes the quality inspection, the reply information is fed back to the user.

[0241] The above-mentioned electronic device can also execute the program code of the following steps in the information processing method: after determining the answer information according to the target message through the target processing system, the answer information is fed back to the user; if the next target message sent by the user after the target message is received, the target action to be executed is determined from multiple actions according to the next target message through the action decision model, and the target action is executed, wherein the multiple actions include a first action and at least one second action, the first action refers to an action used to switch to manual customer service to answer the target message, and the second action refers to an action used to determine the answer information based on the target message.

[0242] Optionally, Figure 7 is a structural block diagram of an electronic device provided according to Embodiment 3 of the present application. Figure 7 As shown, the electronic device 70 may include: one or more ( Figure 7 Only one is shown in the figure) processor 702 and memory 704. The electronic device 70 may also include a storage controller, through which the memory 704 is controlled and managed; the electronic device 70 may also include a peripheral interface, through which the radio frequency module, audio module and display screen are connected.

[0243] Among them, the memory can be used to store software programs and modules, such as program instructions / modules corresponding to the information processing method and device in the embodiment of the present application. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory, that is, realizing the above-mentioned information processing method. The memory may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory may further include a memory remotely arranged relative to the processor, and these remote memories may be connected to the terminal 10 via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0244] The processor can call the information and application programs stored in the memory through the transmission device to perform the following steps: when the target processing system starts the manual customer service, predict whether the target message of the user to be received by the target processing system belongs to the target message type according to at least one preset access rule, wherein the access rule represents the conditions that must be met for the target message to belong to the target message type, and the target message type is a message type that supports the target processing system to replace manual answers; if it is predicted that the target message of the user to be received by the target processing system belongs to the target message type, then when the target processing system receives the target message, the target processing system determines the answer information according to the target message.

[0245] Optionally, the processor may also execute program code for the following steps: determining an action to be executed from a plurality of actions according to a target message through an action decision model in a target processing system, wherein the plurality of actions include a first action and at least one second action, the first action refers to an action for switching to manual customer service to answer the target message, and the second action refers to an action for determining answer information according to the target message; when the action to be executed is the second action, the answer information is determined based on the second action and the target message.

[0246] Optionally, the processor may also execute program code of the following steps: if the action to be executed is the second action belonging to the first category, determining a knowledge base that matches the action to be executed; and determining answer information according to the knowledge base and the target message based on retrieval enhancement generation technology.

[0247] Optionally, the processor may also execute program code for the following steps: based on the target message, determining whether there is at least one target knowledge fragment matching the target message in the knowledge base that matches the action to be executed; when there is at least one target knowledge fragment in the knowledge base that matches the action to be executed, generating a target indication statement based on the at least one target knowledge fragment and the target message, wherein the target indication statement is used to guide the target big model to rewrite the at least one target knowledge fragment to determine answer information for answering the target message; and generating answer information based on the target indication statement through the target big model.

[0248] Optionally, the processor may also execute the program code of the following steps: after determining whether there is at least one target knowledge fragment matching the target message in the knowledge base that matches the action to be executed, if there is no at least one target knowledge fragment in the knowledge base that matches the action to be executed, input the target message into the answer information generation model; and output the answer information corresponding to the target message through the answer information generation model.

[0249] Optionally, the processor may also execute the program code of the following steps: if the action to be executed is a second action belonging to the second category, determining the problem that matches the target message, wherein the problem that matches the target message is one of the following: a problem in the target message, a problem in a historical target message received by the target processing system; judging whether the robot's answer information can solve the problem that matches the target message based on the degree of match between the robot's answer information and the problem that matches the target message, and obtaining a judgment result; and determining the answer information based on the judgment result.

[0250] Optionally, the processor may also execute program codes for the following steps: when the robot's answer information can solve the problem matching the target message, obtaining manual answer information matching the robot's answer information; determining the manual answer information matching the robot's answer information as answer information; when the robot's answer information cannot solve the problem corresponding to the target message, determining an answer information generation script matching the problem matching the target message; and generating answer information according to the answer information generation script.

[0251] Optionally, the processor may also execute the program code of the following steps: after the target processing system receives the target message, if an instruction to transfer to manual customer service is received from manual customer service, the target processing system is controlled to stop determining answer information based on the target message; if an instruction to transfer to manual customer service is not received from manual customer service, the target processing system is determined based on the target message whether to switch to manual customer service to answer the target message; if it is determined to switch to manual customer service to answer the target message, the target processing system is controlled to stop determining answer information based on the target message, and a service request is sent to the manual customer service, wherein the service request is used to request the manual customer service to answer the target message.

[0252] Optionally, the processor may also execute the program code of the following steps: determining whether there is a target keyword in the target message to obtain a first judgment result, wherein the target keyword is a word that represents the user's questioning of the customer service identity of the manual customer service service; determining whether the target message belongs to a message under a target scenario to obtain a second judgment result, wherein the target scenario is a scenario that represents the user's negative emotions; determining whether the conversation round to which the target message belongs is greater than a preset round to obtain a third judgment result; and determining whether to switch to manual customer service to answer the target message based on at least one of the first judgment result, the second judgment result, and the third judgment result.

[0253] Optionally, the processor may also execute the program code of the following steps: after determining the answer information according to the target message through the target processing system, performing quality inspection on the answer information according to at least one preset quality inspection rule; if the answer information passes the quality inspection, feeding back the answer information to the user.

[0254] Optionally, the processor may also execute program code for the following steps: after determining answer information according to the target message through the target processing system, feeding back the answer information to the user; if the next target message sent by the user after the target message is received, determining the target action to be executed from multiple actions according to the next target message through the action decision model, and executing the target action, wherein the multiple actions include a first action and at least one second action, the first action refers to an action for switching to manual customer service to answer the target message, and the second action refers to an action for determining the answer information according to the target message.

[0255] It can be understood by those skilled in the art that Figure 7 The structure shown is for illustration only, and the electronic device may also be a smart phone (such as an Android phone, an iOS phone, etc.), a tablet computer, a PDA, a mobile Internet device (Mobile Internet Devices, MID), a PAD, or other terminal devices. Figure 7 The structure of the electronic device is not limited. Figure 7 More or fewer components (such as network interfaces, display devices, etc.) shown in, or having Figure 7 Different configurations are shown.

[0256] A person of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing the hardware related to the terminal device through a program, and the program can be stored in a computer-readable storage medium, and the storage medium may include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0257] Example 4

[0258] The embodiment of the present application further provides a computer program product. Optionally, in this embodiment, the computer program product may include a computer program, and when the computer program is executed by a processor, the information processing method provided in the first embodiment is implemented.

[0259] The serial numbers of the above-mentioned embodiments of the present application are for description only and do not represent the advantages or disadvantages of the embodiments.

[0260] In the above embodiments of the present application, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, please refer to the relevant description of other embodiments.

[0261] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic, for example, the division of units is only a logical function division, and there may be other division methods in actual implementation, for example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0262] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0263] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0264] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art or all or part of the technical solution, can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, server or network device, etc.) to perform all or part of the steps of each embodiment method of the present application. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, disk or optical disk, etc. Various media that can store program codes.

[0265] The above are only preferred implementations of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. An information processing method, characterized in that: include: In the case where the target processing system starts the manual customer service, predicting whether the target message of the user to be received by the target processing system belongs to the target message type according to at least one preset access rule, wherein the access rule represents the conditions that the target message must meet for it to belong to the target message type, and the target message type is a message type that supports the target processing system to replace manual answers; If it is predicted that the target message of the user to be received by the target processing system belongs to the target message type, then when the target processing system receives the target message, the target processing system determines the reply information according to the target message.

2. The method according to claim 1, characterized in that Determining the answer information according to the target message by the target processing system includes: Determining an action to be executed from a plurality of actions according to the target message through an action decision model in the target processing system, wherein the plurality of actions include a first action and at least one second action, the first action refers to an action for switching to a human customer service to answer the target message, and the second action refers to an action for determining the answer information according to the target message; In a case where the action to be performed is the second action, the answer information is determined based on the second action and the target message.

3. The method according to claim 2, characterized in that Different second actions are used to determine the answer information according to different methods and / or different data, the at least one second action includes a second action belonging to the first category, the second action belonging to the first category refers to an action used to determine the answer information according to the retrieval enhancement generation technology and the target message, and when the action to be performed is the second action, determining the answer information based on the second action and the target message includes: If the action to be performed is a second action belonging to the first category, determining a knowledge base matching the action to be performed; Based on the retrieval enhancement generation technology, the answer information is determined according to the knowledge base and the target message.

4. The method according to claim 3, characterized in that Based on the retrieval enhancement generation technology, determining the answer information according to the knowledge base and the target message includes: According to the target message, determining whether there is at least one target knowledge fragment matching the target message in the knowledge base matching the action to be executed; In the case where the at least one target knowledge fragment exists in the knowledge base matched by the action to be executed, a target instruction statement is generated according to the at least one target knowledge fragment and the target message, wherein the target instruction statement is used to guide the target macro model to rewrite the at least one target knowledge fragment to determine answer information for answering the target message; The answer information is generated according to the target instruction sentence by the target macro model.

5. The method according to claim 4, characterized in that After determining whether there is at least one target knowledge fragment matching the target message in the knowledge base matching the action to be executed, the method further includes: In the case that the at least one target knowledge fragment does not exist in the knowledge base matching the action to be executed, inputting the target message into the answer information generation model; The answer information corresponding to the target message is outputted through the answer information generation model.

6. The method according to claim 2, characterized in that Different second actions are used to determine the answer information according to different methods and / or different data, the at least one second action includes a second action belonging to a second category, the second action belonging to the second category refers to an action for determining the answer information according to the robot answer information of the user and the target message of the robot customer service service, and when the action to be executed is the second action, determining the answer information based on the second action and the target message includes: If the action to be performed is a second action belonging to the second category, determining a problem matched by the target message, wherein the problem matched by the target message is one of the following: a problem in the target message, a problem in a historical target message received by the target processing system; According to the matching degree between the robot answer information and the problem matched by the target message, judging whether the robot answer information can solve the problem matched by the target message, and obtaining a judgment result; The answer information is determined according to the judgment result.

7. The method according to claim 6, characterized in that Determining the answer information according to the judgment result includes: In the case where the robot answer information can solve the problem of matching the target message, obtaining human answer information matching the robot answer information; determining the manual answer information matching the robot answer information as the answer information; In the case where the robot answer information cannot solve the problem corresponding to the target message, determining an answer information generation script that matches the problem matching the target message; The answer information is generated according to the answer information generation script.

8. The method according to claim 1, characterized in that After the target processing system receives the target message, the method further includes: When receiving a transfer instruction from a manual customer service, controlling the target processing system to stop determining the answer information according to the target message; In the case where the instruction to switch to a manual customer service sent by the manual customer service is not received, determining whether to switch to a manual customer service to answer the target message according to the target message; If it is determined to switch to manual customer service to answer the target message, the target processing system is controlled to stop determining the answer information based on the target message, and a service request is sent to the manual customer service, wherein the service request is used to request the manual customer service to answer the target message.

9. The method according to claim 8, characterized in that Determining whether to switch to manual customer service to answer the target message according to the target message includes: Determine whether a target keyword exists in the target message to obtain a first determination result, wherein the target keyword is a word representing that the user questions the customer service identity of the manual customer service service; Determine whether the target message belongs to a message under a target scenario, and obtain a second determination result, wherein the target scenario is a scenario representing that the user has negative emotions; Determine whether the conversation round to which the target message belongs is greater than a preset round, and obtain a third determination result; Determine whether to switch to manual customer service to answer the target message according to at least one of the first judgment result, the second judgment result, and the third judgment result.

10. The method according to claim 1, characterized in that After determining the answer information according to the target message by the target processing system, the method further includes: Performing quality inspection on the answer information according to at least one preset quality inspection rule; When the answer information passes the quality inspection, the answer information is fed back to the user.

11. The method according to claim 1, characterized in that After determining the answer information according to the target message by the target processing system, the method further includes: Feedback the answer information to the user; If a next target message sent by the user after the target message is received, the target action to be executed is determined from multiple actions according to the next target message through the action decision model, and the target action is executed, wherein the multiple actions include a first action and at least one second action, the first action refers to an action for switching to manual customer service to answer the target message, and the second action refers to an action for determining answer information based on the target message.

12. An information processing device, characterized in that: include: a prediction unit, configured to predict, when the target processing system starts a manual customer service, whether a target message of a user to be received by the target processing system belongs to a target message type according to at least one preset admission rule, wherein the admission rule represents a condition that the target message must meet in order to belong to the target message type, and the target message type is a message type that supports the target processing system to replace manual answers; A determination unit is used to determine reply information according to the target message through the target processing system when the target processing system receives the target message if it is predicted that the target message of the user to be received by the target processing system belongs to the target message type.

13. An electronic device, characterized in that: include: A memory storing an executable program; A processor, configured to run the program, wherein the program, when run, executes the information processing method of any one of claims 1 to 11.

14. A computer program product, characterized in that The invention comprises a computer program or instructions, which, when executed by a processor, implements the information processing method according to any one of claims 1 to 11.

Images