Multi-scene dialogue system and multi-scene dialogue method for virtual human customer service

Abstract

The invention discloses a multi-scene dialogue system for a virtual human customer service, and the system comprises a scene dialogue platform which is used for carrying out the scene dialogue between the virtual human customer service and a user, transmitting related information to a feature analysis module, and feeding back the related information selected from a cloud database to the user; the feature analysis module is used for performing feature confirmation on the question information sent by the user, selecting related information with related features from the cloud database based on the confirmed related features, and transmitting the selected related information to the scene dialogue platform; the period analysis module is used for monitoring dialogue information generated by the scene dialogue platform in a certain monitoring period and determining a to-be-optimized partition based on a monitoring result; and the related optimization module is used for optimizing related information of a specified partition of the cloud database according to the partition to be optimized. The system solves the problem that the information answered by the related AI cannot fit the user well, and solves the related problems of the user.

Description

Technical Field

[0001] The present application belongs to the field of intelligent interaction, and in particular relates to a multi-scenario dialogue system and a multi-scenario dialogue method for virtual human customer service. Background Art

[0002] Virtual customer service refers to an intelligent customer service system that uses artificial intelligence technology to provide interactive services to customers. This system can understand customer needs, answer customer questions, and provide solutions through natural language processing technology, speech recognition technology, machine learning technology, etc.

[0003] In the prior art, the application with publication number CN116860934A discloses a multi-scenario multi-round dialogue method and system for virtual customer service, the system includes a dialogue scheduling component, a business processing component, a knowledge base query component and a chat component. The dialogue scheduling component is used to manage the running order and calling relationship of each scenario dialogue component, the business processing component is responsible for processing dialogues related to business processing, the knowledge base query component is responsible for processing general knowledge retrieval dialogues of non-business processing types, and the chat component is responsible for processing general dialogues of non-business processing and non-knowledge base query types.

[0004] During the scene dialogue process, the virtual customer service of this method generally determines the relevant interaction information based on the corresponding interaction questions and displays it through the dialogue platform. However, in the actual interaction process, the information answered by the relevant AI cannot fit the user well and cannot solve the user's related problems well, resulting in poor interaction effect and failure to achieve better interaction effect.

[0005] Therefore, there is a need for a multi-scenario dialogue system for virtual human customer service that can solve the above problems. Summary of the invention

[0006] In order to address the deficiencies of the prior art, the present application provides a multi-scenario dialogue system and a multi-scenario dialogue method for virtual human customer service, which solves the problem that the information answered by the relevant AI cannot fit the user well and cannot solve the user's related problems well.

[0007] The technical effects to be achieved by this application are achieved through the following solutions: According to a first aspect of the present application, a multi-scenario dialogue system for virtual human customer service is provided, including a scenario dialogue platform, a feature analysis module, a cycle analysis module, and a related optimization module, wherein: The scenario dialogue platform is used for scenario dialogue between the virtual customer service and the user, and sends relevant information to the feature analysis module, and feeds back relevant information selected from the cloud database to the user; The feature analysis module is used to confirm the features of the question information sent by the user, and based on the confirmed relevant features, select relevant information with relevant features from the cloud database, and transmit the selected relevant information to the scene dialogue platform; A cycle analysis module, used to monitor the dialogue information generated by the scene dialogue platform in a certain monitoring cycle, and determine the partition to be optimized based on the monitoring results; The relevant optimization module is used to optimize the relevant information of the specified partition of the cloud database according to the partition to be optimized.

[0008] Preferably, the cloud database is provided with a plurality of different partitions, each partition is provided with different relevant information, and different partitions have different feature sequences.

[0009] Preferably, the method by which the feature analysis module performs feature confirmation on the problem information is: Compare the question information sent by the user with the feature words in the feature comparison table to obtain a comparison result; Determine the corresponding feature assignment based on the comparison result, and sort the feature assignment according to the sequence of the feature words in the question information to generate a feature assignment sequence; Compare the feature assignment sequence with the feature sequence of each partition of the cloud database, determine a group of partitions with the highest degree of overlap, and select relevant information of this partition; The selected relevant information is displayed on the scene dialogue platform.

[0010] Preferably, during the comparison of the feature assignment sequence with the feature sequence of each partition of the cloud database, if there are different partitions with the highest degree of comparison overlap, then the relevant information of a group of partitions is randomly selected for display; During the same dialogue, the relevant information selected this time will not be selected again. If the same feature assignment sequence still appears, the relevant information of other partitions will be selected for display.

[0011] Preferably, the specific method for the periodic analysis module to determine the partition to be optimized is: Confirming different conversation information monitored within a preset monitoring period, and processing the confirmed different conversation information; Recognize "transfer to manual" in the dialogue information as a delimiting character, mark the dialogue information before the delimiting character as human-machine dialogue information, and mark the dialogue information after the delimiting character as customer service dialogue information; According to the number of the human-computer dialogue information and the number of the customer service dialogue information, the associated partition is confirmed and marked as the partition to be optimized.

[0012] Preferably, according to the number of human-computer dialogue information and the number of customer service dialogue information, the specific method of confirming the associated partition and marking it as the partition to be optimized is: The number of times of confirming human-machine dialogue information from a single group of dialogue information accounts for Z1 and the number of times of confirming customer service dialogue information accounts for Z2, where: Z1 = number of human-computer dialogue messages ÷ total number of dialogue messages; Z2 = number of customer service conversations ÷ total number of conversations; The feedback value Fk of this single group dialogue information is confirmed by: Fk=Z1×C1+Z2×C2, where C1 and C2 are both preset fixed coefficient factors; The feedback value Fk corresponding to the single group of dialogue information is compared with the preset value Y1. If Fk<Y1, the relevant information of the human-computer dialogue information in the single group of dialogue information is confirmed, and then the associated partition corresponding to the relevant information is confirmed and marked as the partition to be optimized.

[0013] Preferably, when the feedback value Fk≥Y1, no processing is performed.

[0014] Preferably, the periodic analysis module performs periodic analysis on the partitions to be optimized that have completed optimization. When the partitions to be optimized that have completed optimization are still determined to be partitions to be optimized, they are sent to the signal module, and the signal module generates and displays an abnormal signal.

[0015] Preferably, according to the abnormal signal sent by the signal module, the partition to be optimized is manually processed or a new partition is added.

[0016] According to a second aspect of the present application, a multi-scenario dialogue method using the multi-scenario dialogue system for virtual human customer service is provided, comprising the following steps: Step 1: Based on the scene dialogue platform, the characteristics of the question information sent by the user are confirmed, and based on the confirmed relevant characteristics, relevant information with relevant characteristics is selected from the cloud database, and the selected relevant information is transmitted to the scene dialogue platform; Step 2: Define a set of monitoring cycles, monitor the dialogue information generated by the scene dialogue platform, determine the feedback value of the corresponding dialogue scene based on the monitoring results, determine the partition to be optimized based on the numerical value of the feedback value, and transmit the determined partition to be optimized to the relevant optimization module; Step 3: The relevant optimization module optimizes the relevant information of the specified partition of the cloud database according to the partition to be optimized; Step 4: Monitor the partitions to be optimized that have been optimized. If they are still determined to be partitions to be optimized during subsequent periodic analysis, an abnormal signal is generated and displayed through the signal module.

[0017] According to an embodiment of the present application, the beneficial effects of using the multi-scenario dialogue system for virtual human customer service are: by analyzing the interaction scenarios in the human-computer interaction process, the relevant feedback values ​​are evaluated, and then based on the evaluated feedback values, the scene interaction effect in this interaction process is confirmed, and whether it is necessary to optimize the relevant partitions, and the information optimization is performed for the relevant partitions that need to be optimized to ensure the overall effect of the multi-scenario dialogue, and through the relevant numerical analysis, a better recognition effect can be achieved, and then the calibrated associated partitions are optimized in a timely manner to ensure the subsequent related interaction effects; In the subsequent interaction process, if the associated partition is still determined to be an associated partition that needs to be optimized, it is necessary to make a new determination to identify whether it is necessary to add relevant feature sequences and related information to ensure the relevant effect of its human-computer interaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the embodiments of the present application or the existing technical solutions, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0019] Figure 1 This is a structural block diagram of a multi-scenario dialogue system for virtual human customer service in one embodiment of the present application; Figure 2 This is a flowchart of a multi-scenario dialogue method for virtual human customer service in one embodiment of the present application. DETAILED DESCRIPTION

[0020] In order to make the purpose, technical solution and advantages of the present application clearer, the technical solution of the present application will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. 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 the field without creative work are within the scope of protection of the present application.

[0021] like Figure 1As shown, a multi-scenario dialogue system for virtual human customer service in an embodiment of the present application includes a scenario dialogue platform, a feature analysis module, a cycle analysis module, a cloud database, a signal module and a related optimization module. The scenario dialogue platform and the feature analysis module are bidirectionally connected, and the feature analysis module and the cloud database are bidirectionally connected. The scenario dialogue platform, the cycle analysis module, the related optimization module and the cloud database are electrically connected from the output node to the input node in sequence, and the cycle analysis module is electrically connected to the signal module input node; wherein: The scene dialogue platform is a dialogue platform between the virtual human customer service and the user, which is used for scene dialogue between the virtual human customer service and the user, and sends relevant information to the feature analysis module, and feeds back relevant information selected from the cloud database to the user; The feature analysis module is used to confirm the features of the question information sent by the user, and based on the confirmed relevant features, select relevant information with relevant features from the cloud database, and transmit the selected relevant information to the scene dialogue platform; the specific method of feature confirmation is: Based on the question information sent by the user, feature comparison is performed in combination with a feature comparison table, wherein the feature comparison table is a preset table prepared in advance by relevant operators, and has relevant feature words and relevant feature assignments set inside. The relevant words reflected in the question information are compared with the feature words inside the feature comparison table, and the corresponding feature assignment is determined based on the comparison result, and then the feature assignment is sorted according to the order of the feature words inside the question information to generate a feature assignment sequence; Compare the feature assignment sequence with the feature sequence of each different partition of the cloud database, where the feature sequences are all preset, and are preset in advance by the operator based on the response conditions of different partitions. Based on the comparison results, determine a group of partitions with the highest comparison overlap, select the relevant information of this partition, and directly display it in the scene dialogue platform. If there are different partitions with the highest comparison overlap, a group of partitions will be randomly selected for relevant information display. In the same dialogue process, the relevant information selected this time will no longer be selected. If the same feature assignment sequence still appears, select the relevant information of other partitions for display; Specifically, in the actual human-computer interaction process, the human-computer performs feature analysis based on the relevant information sent by the relevant users to determine the relevant feature assignments, and then uses the feature assignments for search to quickly lock the relevant information and display it to ensure the corresponding dialogue effect.

[0022] The cloud database in this embodiment is provided with several different partitions, and each partition is provided with different related information, and each different partition has a different feature sequence, and the different related information and feature sequences within the cloud database are set in advance by the operator.

[0023] The cycle analysis module is used to monitor the dialogue information generated by the scene dialogue platform in a certain monitoring cycle, and determine the feedback value of the corresponding dialogue scene based on the monitoring result, determine the partition to be optimized based on the numerical value of the feedback value, and transmit the determined partition to be optimized to the relevant optimization module; the relevant optimization module optimizes the relevant information of the specified partition of the cloud database.

[0024] The specific method for determining the partition to be optimized is: A set of monitoring cycles T is defined, where T is a preset value, and its specific value is determined by the operator based on experience. Different dialogue information monitored within this monitoring cycle T is confirmed, and the confirmed different dialogue information is processed; Select "transfer to manual" as the delimiter character in the conversation information, mark the relevant conversation information before transfer to manual as human-machine conversation information, and mark the relevant conversation information after transfer to manual as customer service conversation information; Confirm the proportion of human-machine dialogue information Z1 and customer service dialogue information Z2 in a single group of dialogue information (for example, in a dialogue information, a user asks a question, and the dialogue center feeds back a group of related information, which is recorded as a single group of dialogue information), where Z1 = the number of human-machine dialogue information ÷ the total number of dialogue information dialogues, and Z2 = the number of customer service dialogue information ÷ the total number of dialogue information dialogues. Use: Fk = Z1 × C1 + Z2 × C2 to confirm the feedback value Fk of this single group of dialogue information, where C1 and C2 are both preset fixed coefficient factors, and their specific values ​​are determined by the operator based on experience; The feedback value Fk corresponding to this single group of dialogue information is compared with the preset value Y1, where the specific value of Y1 is determined by the operator based on experience. If Fk≥Y1, no processing is performed, which means that the feedback value is large, the number of human-computer interactions is large, and the number of customer service interactions is small. If Fk<Y1, the relevant information of the human-computer dialogue information in this single group of dialogue information is confirmed, and then the associated partition corresponding to the relevant information is confirmed and marked as the partition to be optimized, and the confirmed partition to be optimized is displayed in the relevant optimization module.

[0025] Its external personnel operate the relevant optimization modules and optimize the relevant information of the specified partition of the cloud database based on the confirmed partition to be optimized to ensure the subsequent human-computer interaction results; Specifically, in the actual human-computer interaction process, based on the relevant interaction data, by confirming the specific time of transfer to manual, the relevant accuracy of the corresponding AI feedback information can be identified. If the feedback information is relatively in line with the expectations of the relevant personnel, then the feedback effect is better. Based on the specific number of times and relevant duration of the entire interaction process, the relevant feedback value can be identified, and the relevant feedback result can be confirmed based on the relevant feedback value. Through relevant numerical analysis, better recognition effect can be achieved. Subsequently, timely optimization can be carried out for the calibrated associated partitions to ensure the subsequent related interaction effects.

[0026] In this embodiment, it also includes: for the partition to be optimized that has been optimized, in the subsequent other periodic analysis processes, if the partition to be optimized that has been optimized is still determined to need to be optimized, an abnormal signal is generated and displayed through the signal module, and an external operator evaluates whether a new partition is needed based on the displayed abnormal signal to ensure the subsequent human-computer interaction results; Because some conversation processes are optimized and then still judged as related partitions that need to be optimized, in the subsequent analysis process, it is necessary to make additional judgments to identify whether it is necessary to add relevant feature sequences and related information to ensure the relevant effects of human-computer interaction.

[0027] like Figure 2 As shown, in one embodiment of the present application, the multi-scenario dialogue method of the multi-scenario dialogue system for virtual human customer service is adopted, including the following steps: Step 1: Based on the scene dialogue platform, the characteristics of the question information sent by the user are confirmed, and based on the confirmed relevant characteristics, relevant information with relevant characteristics is selected from the cloud database, and the selected relevant information is transmitted to the scene dialogue platform; Step 2: Define a set of monitoring cycles, monitor the dialogue information generated by the scene dialogue platform, determine the feedback value of the corresponding dialogue scene based on the monitoring results, determine the partition to be optimized based on the numerical value of the feedback value, and transmit the determined partition to be optimized to the relevant optimization module; Step 3: The relevant optimization module optimizes the relevant information of the specified partition of the cloud database according to the partition to be optimized; Step 4: Monitor the partitions to be optimized that have been optimized. If they are still determined to be partitions to be optimized during subsequent periodic analysis, an abnormal signal is generated and displayed through the signal module.

[0028] According to an embodiment of the present application, the beneficial effects of using the multi-scenario dialogue system for virtual human customer service are: by analyzing the interaction scenarios in the human-computer interaction process, the relevant feedback values ​​are evaluated, and then based on the evaluated feedback values, the scene interaction effect in this interaction process is confirmed, and whether it is necessary to optimize the relevant partitions, and the information optimization is performed for the relevant partitions that need to be optimized to ensure the overall effect of the multi-scenario dialogue, and through the relevant numerical analysis, a better recognition effect can be achieved, and then the calibrated associated partitions are optimized in a timely manner to ensure the subsequent related interaction effects; In the subsequent interaction process, if the associated partition is still determined to be an associated partition that needs to be optimized, it is necessary to make a new determination to identify whether it is necessary to add relevant feature sequences and related information to ensure the relevant effect of its human-computer interaction.

[0029] It should be noted that the above detailed descriptions are exemplary and are intended to provide further explanation of the present application. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the art to which the present application belongs.

[0030] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present application. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should also be understood that when the terms "comprise" and / or "include" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0031] 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 terms 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.

[0032] In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units that are not explicitly listed or inherent to these processes, methods, products, or apparatuses.

[0033] For ease of description, spatially relative terms, such as "above", "above", "on the upper surface of", "above", etc., may be used herein to describe the spatial positional relationship between a device or feature and other devices or features as shown in the figure. It should be understood that spatially relative terms are intended to include different orientations of the device in use or operation in addition to the orientation described in the figure. For example, if the device in the accompanying drawings is inverted, the device described as "above other devices or structures" or "above other devices or structures" will be positioned as "below other devices or structures" or "below other devices or structures". Thus, the exemplary term "above" may include both "above" and "below". The device may also be positioned in other different ways, such as rotated 90 degrees or in other orientations, and the spatially relative descriptions used herein are interpreted accordingly.

[0034] In the above detailed description, reference is made to the accompanying drawings, which form a part of this document. In the accompanying drawings, similar symbols typically identify similar components unless the context indicates otherwise. The illustrated embodiments described in the detailed description, drawings, and claims are not meant to be limiting. Other embodiments may be used, and other changes may be made, without departing from the spirit or scope of the subject matter presented herein.

[0035] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A multi-scenario dialogue system for virtual human customer service, characterized in that: It includes scene dialogue platform, feature analysis module, cycle analysis module and related optimization modules, among which: The scenario dialogue platform is used for scenario dialogue between the virtual customer service and the user, and sends relevant information to the feature analysis module, and feeds back relevant information selected from the cloud database to the user; The feature analysis module is used to confirm the features of the question information sent by the user, and based on the confirmed relevant features, select relevant information with relevant features from the cloud database, and transmit the selected relevant information to the scene dialogue platform; A cycle analysis module, used to monitor the dialogue information generated by the scene dialogue platform in a certain monitoring cycle, and determine the partition to be optimized based on the monitoring results; The relevant optimization module is used to optimize the relevant information of the specified partition of the cloud database according to the partition to be optimized.

2. The multi-scenario dialogue system for virtual human customer service according to claim 1, characterized in that: The cloud database is provided with a plurality of different partitions, each partition is provided with different relevant information, and different partitions have different feature sequences.

3. The multi-scenario dialogue system for virtual human customer service according to claim 2, characterized in that: The method for the feature analysis module to confirm the features of the problem information is as follows: Compare the question information sent by the user with the feature words in the feature comparison table to obtain a comparison result; Determine the corresponding feature assignment based on the comparison result, and sort the feature assignment according to the sequence of the feature words in the question information to generate a feature assignment sequence; Compare the feature assignment sequence with the feature sequence of each partition of the cloud database, determine a group of partitions with the highest degree of overlap, and select relevant information of this partition; The selected relevant information is displayed on the scene dialogue platform.

4. The multi-scenario dialogue system for virtual human customer service according to claim 3, characterized in that: When comparing the feature assignment sequence with the feature sequence of each partition of the cloud database, if there are different partitions with the highest degree of comparison overlap, a group of partitions will be randomly selected for relevant information display; During the same dialogue, the relevant information selected this time will not be selected again. If the same feature assignment sequence still appears, the relevant information of other partitions will be selected for display.

5. The multi-scenario dialogue system for virtual human customer service according to claim 1, characterized in that: The specific method for the cycle analysis module to determine the partition to be optimized is: Confirming different conversation information monitored within a preset monitoring period, and processing the confirmed different conversation information; Recognize "transfer to manual" in the dialogue information as a delimiting character, mark the dialogue information before the delimiting character as human-machine dialogue information, and mark the dialogue information after the delimiting character as customer service dialogue information; According to the number of the human-computer dialogue information and the number of the customer service dialogue information, the associated partition is confirmed and marked as the partition to be optimized.

6. The multi-scenario dialogue system for virtual human customer service according to claim 5, characterized in that: According to the number of human-computer dialogue information and the number of customer service dialogue information, the specific method of confirming the associated partition and marking it as the partition to be optimized is: The number of times of confirming human-machine dialogue information from a single group of dialogue information accounts for Z1 and the number of times of confirming customer service dialogue information accounts for Z2, where: Z1 = number of human-computer dialogue messages ÷ total number of dialogue messages; Z2 = number of customer service conversations ÷ total number of conversations; The feedback value Fk of this single group dialogue information is confirmed by: Fk=Z1×C1+Z2×C2, where C1 and C2 are both preset fixed coefficient factors; The feedback value Fk corresponding to the single group of dialogue information is compared with the preset value Y1. If Fk<Y1, the relevant information of the human-computer dialogue information in the single group of dialogue information is confirmed, and then the associated partition corresponding to the relevant information is confirmed and marked as the partition to be optimized.

7. The multi-scenario dialogue system for virtual human customer service according to claim 6, characterized in that: When the feedback value Fk≥Y1, no processing is performed.

8. The multi-scenario dialogue system for virtual human customer service according to claim 6, characterized in that: The periodic analysis module performs periodic analysis on the partitions to be optimized that have completed optimization. When the partitions to be optimized that have completed optimization are still determined to be partitions to be optimized, they are sent to the signal module, and the signal module generates and displays an abnormal signal.

9. The multi-scenario dialogue system for virtual human customer service according to claim 8, characterized in that: According to the abnormal signal sent by the signal module, the partition to be optimized is manually processed or a new partition is added.

10. A multi-scenario dialogue method for a multi-scenario dialogue system for virtual human customer service using any one of claims 1 to 9, characterized in that: The steps include: Step 1: Based on the scene dialogue platform, the characteristics of the question information sent by the user are confirmed, and based on the confirmed relevant characteristics, relevant information with relevant characteristics is selected from the cloud database, and the selected relevant information is transmitted to the scene dialogue platform; Step 2: Define a set of monitoring cycles, monitor the dialogue information generated by the scene dialogue platform, determine the feedback value of the corresponding dialogue scene based on the monitoring results, determine the partition to be optimized based on the numerical value of the feedback value, and transmit the determined partition to be optimized to the relevant optimization module; Step 3: The relevant optimization module optimizes the relevant information of the specified partition of the cloud database according to the partition to be optimized; Step 4: Monitor the partitions to be optimized that have been optimized. If they are still determined to be partitions to be optimized during subsequent periodic analysis, an abnormal signal is generated and displayed through the signal module.

Images