A dialogue interactive service closed-loop ticket point method

Abstract

The application relates to a dialogue interactive service closed loop small ticket point method, and belongs to the technical field of artificial intelligence, which solves the problems that an existing small ticket point interactive process cannot be flexibly adjusted and cannot realize service closed loop. The method comprises the following steps: constructing a small ticket point interactive process and loading the process into a memory; guiding a user to upload a small ticket picture through a dialogue node in the process, obtaining and saving small ticket element information; identifying whether the small ticket element information is complete through a judgment node, and identifying whether the small ticket element information is correct according to user feedback information; feeding back the small ticket element information to the user through a reply node, and receiving user feedback information; when it is identified that the small ticket element information is correct according to the user feedback information, sending the small ticket element information to a point system through a query node; and when it is identified that the small ticket element information is incorrect according to the user feedback information for multiple times, feeding back into an artificial auditing system to the user through the reply node. Interactive service closed loop of small ticket points is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of artificial intelligence, and particularly relates to a dialog interactive service closed-loop small ticket point method. BACKGROUND

[0002] As a common marketing means, small ticket points can effectively attract the attention of consumers and promote sales growth.

[0003] However, the traditional small ticket points require customers to go to the small ticket point place, show the small ticket to the staff, and enter the small ticket points by the staff. This way brings great inconvenience to customers and increases the cost of businesses.

[0004] With the advancement of technology, in order to solve the problems existing in the traditional small ticket points, the market provides some solutions, such as a shopping small ticket submission small ticket point system, which automatically submits small ticket information to the point system and cooperates with manual auditing. However, this method still has problems such as low information extraction accuracy, long manual auditing time, poor customer experience, high comprehensive cost, and inability to realize service closed loop. SUMMARY

[0005] In view of the above analysis, the embodiments of the present application aim to provide a dialog interactive service closed-loop small ticket point method to solve the problem that the existing small ticket point interaction process cannot be flexibly adjusted and cannot realize service closed loop.

[0006] The embodiments of the present application provide a dialog interactive service closed-loop small ticket point method, comprising the following steps:

[0007] A small ticket point interaction process is constructed, and each node and its configuration information are saved and loaded into the memory; each node includes a dialog node, a judgment node, a reply node and a query node;

[0008] When it is recognized that the user's intention is related to small ticket points, the small ticket point interaction process is started; the user is guided to upload a small ticket picture through the dialog node, and small ticket element information is obtained and saved; whether the small ticket element information is complete is identified through the judgment node, and whether the small ticket element information is correct is identified according to user feedback information; the small ticket element information is fed back to the user through the reply node, and user feedback information is received;

[0009] When it is identified according to the user feedback information that the small ticket element information is correct, the small ticket element information is sent to the point system through the query node, and the small ticket point interaction process is ended; when it is identified according to the user feedback information that the small ticket element information is incorrect for multiple times, the user is fed back to enter the manual auditing system through the reply node, and the small ticket point interaction process is ended.

[0010] Further improvement based on the above method, the same number of slot point entities as the invoice element information is set in the dialogue node, and the type of the slot point entity is set as a system entity. The system entity is obtained by sending the user-uploaded invoice picture to an optical character recognition model to recognize the invoice text information, and then sending the invoice text information to an entity recognition model.

[0011] Further improvement based on the above method, the entity recognition model sequentially includes a two-dimensional feature extraction module, a feature fusion module, a decoder module and an output module; wherein the two-dimensional feature extraction module includes two parallel branches, the first branch includes a text embedding layer and an encoder layer connected in sequence; the second branch includes a Bert embedding layer, a width embedding layer and a concatenation layer connected in sequence; the feature fusion module fuses the feature vectors output by the two branches in the two-dimensional feature extraction module through a linear layer, and then transmits them to the decoder module to generate an output sequence, and finally obtains the recognized entity through the output module.

[0012] Further improvement based on the above method, the dialogue node further includes a guide technique, a timeout time, a timeout number and a counter-question strategy; after entering the dialogue node, it is identified whether the slot point entity set in the dialogue node has a value, if not, the first guide technique set is sent to the user to remind the user to upload the invoice picture, when the set timeout time is reached and the invoice picture is still not obtained, the first guide technique set is sent to the user again according to the counter-question strategy, and the actual timeout number is recorded, when the actual timeout number exceeds the set timeout number, the dialogue node is exited and the manual review system is entered.

[0013] Further improvement based on the above method, after entering the dialogue node, it is identified whether the slot point entity set in the dialogue node has a value, if yes, the second guide technique set is sent to the user to remind the user to upload a clear invoice picture, when the invoice picture is obtained within the timeout time, it is identified whether the newly obtained information of each invoice element is consistent with the value in the slot point entity, if yes, the value in the slot point entity is supplemented according to the newly obtained information, if not, the newly obtained invoice element information is used to cover the value of the corresponding slot point entity.

[0014] Further improvement based on the above method, when judging whether the invoice element information is complete, two branches are set in the judgment node, and one of them is a default branch. A condition group composed of multiple conditions is set for the non-default branch; each condition introduces a slot point entity and sets the slot point entity to have a value.

[0015] Further improvement based on the above method, when the judgment node identifies whether the ticket element information is correct according to the user feedback information, two branches are set in the judgment node, and one of them is a default branch, and a condition group consisting of one condition is set for the non-default branch; the condition is set according to the button variable and button return value set in the reply node.

[0016] Further improvement based on the above method, the reply node includes: reply type, reply content type and reply content; by setting the reply type as a template, the reply content type as text, and setting the reply template in the reply content and introducing the slot entity as a variable, the ticket element information is fed back to the user through the reply node, which is to replace the variable in the reply template with the value of the obtained slot entity to send the ticket element information to the user.

[0017] Further improvement based on the above method, the reply node further includes: whether to enable button setting, when the button setting is enabled, set the button variable, select the button type and set the button label and button return value; the user feedback information received by the reply node is the button return value corresponding to the button selected by the user, and is set in the button variable.

[0018] Further improvement based on the above method, the query node includes: setting request type, request method, input parameter and its value, and output parameter; the ticket element information is sent to the point system through the query node, which is to call the point system set in the request method according to the set request type by taking the slot entity and its value as the input parameter and its value.

[0019] Compared with the prior art, the present application can at least realize one of the following beneficial effects:

[0020] 1. Through flexible construction of automatic interaction process, the demand for small ticket points is quickly met; through the cooperative work of the dialogue node, the judgment node and the reply node in the automatic interaction process, multi-round dialogue is realized, complex business scenarios are dealt with, and user dialogue experience is improved.

[0021] 2. By guiding the user to upload the small ticket picture multiple times, it is ensured that the collected small ticket information is accurate and correct, the overall recognition accuracy of the small ticket is greatly improved, the process is dynamically executed according to the customer's reply and operation, the point process under different conditions is completed in real time, the interactive service closed loop small ticket point process is realized, the time and workload of manual processing of small ticket points are reduced, the human investment cost is reduced, and the efficiency of point processing is improved.

[0022] The technical solutions in the present application can be combined with each other to realize more preferred combination solutions. Other features and advantages of the present application will be described in the following description, and some advantages will become apparent from the description or can be understood by implementing the present application. The objects and other advantages of the present application can be realized and obtained by the contents particularly pointed out in the description and the drawings. BRIEF DESCRIPTION OF DRAWINGS

[0023] The accompanying drawings are included to provide a further understanding of the present application, and are incorporated herein and constitute a part of the detailed description. The drawings illustrate embodiments of the present application and, together with the description, serve to explain the principles of the present application.

[0024] Figure 1 A flow chart of a dialogue interactive service closed-loop ticket point method in an embodiment of the present application is shown in FIG. 1.

[0025] Figure 2 A ticket point interactive flowchart in an embodiment of the present application is shown in FIG. 2. DETAILED DESCRIPTION

[0026] The preferred embodiments of the present application are described in detail below with reference to the accompanying drawings, which form a part of this application. The drawings show embodiments of the present application and, together with the description, serve to explain the principles of the present application, but are not used to limit the scope of the present application.

[0027] One specific embodiment of the present application discloses a dialogue interactive service closed-loop ticket point method, as shown in FIG. 1, which includes the following steps: Figure 1

[0028] S1, a ticket point interactive flowchart is constructed, and each node and its configuration information are saved and loaded into the memory; each node includes a dialogue node, a judgment node, a reply node, and a query node;

[0029] S2, when it is identified that the user's intention is related to ticket points, the ticket point interactive flowchart is started; the user is guided to upload a ticket picture through the dialogue node, and ticket element information is obtained and saved; whether the ticket element information is complete is identified through the judgment node, and whether the ticket element information is correct is identified according to user feedback information; the ticket element information is fed back to the user through the reply node, and user feedback information is received;

[0030] S3, when it is identified according to the user feedback information that the ticket element information is correct, the ticket element information is sent to a point system through the query node, and the ticket point interactive flowchart is ended; when it is identified according to the user feedback information that the ticket element information is incorrect for multiple times, the user is fed back to enter an artificial audit system through the reply node, and the ticket point interactive flowchart is ended.

[0031] ​During implementation, the automated interaction process is flexibly constructed to quickly meet the needs of receipt points; the collaborative work of dialogue nodes, judgment nodes and response nodes in the automated interaction process enables multi-turn dialogues, which can cope with complex business scenarios, improve the user dialogue experience, and improve the efficiency and accuracy of points processing.

[0032] Specifically, the ticket points interaction process in step S1 is constructed through a visual interface, including: start node, dialogue node, collection node, query node, judgment node, reply node, assignment node, and end node. Each node has a unique node ID, and is associated with the node ID of the parent node through the parent node ID, and with the node ID of the next-level node through the child node ID. Different nodes are identified by node type.

[0033] The following is combined with Figure 2 This is a schematic diagram of the points interaction process on receipts, detailing the configuration and implementation methods of each node.

[0034] (1) Start Node

[0035] The start node serves only as the initiation node for the process, and there is only one. In this embodiment, when it is recognized that the user's intent is related to the receipt points, the receipt points interaction process is initiated, and the process proceeds from the start node to its child nodes.

[0036] It should be noted that users can communicate with the interactive dialogue system via mini-programs or official accounts, using either text or voice. In the case of voice, the system first converts the user's speech into text using an automatic speech recognition model, then uses a natural language processing model to identify the text intent. If the text intent is related to receipt points, the receipt points interaction process is initiated.

[0037] (2) Dialogue Node

[0038] Dialogue nodes are used to collect necessary information by engaging in multi-turn conversations with users through guided dialogue. Within these nodes, information is obtained by constructing "scratching" entities. These entities include: enumerated entities, regular expression entities, intent entities, and system entities. Enumerated entities use predefined enumeration values; when used, they retrieve entity information matching the predefined enumeration values ​​based on the user-uploaded information. Regular expression entities use predefined regular expressions; when used, they extract entity information from the user-uploaded information that conforms to the regular expression. Intent entities use predefined similarity thresholds; when used, they identify the intent from the user-uploaded information and extract the intent with the highest similarity exceeding the similarity threshold as the entity information. System entities use predefined entity recognition models; when used, they extract the corresponding entity information from the user-uploaded information using these models.

[0039] Further, the dialogue node further comprises a guide dialogue, a timeout time, a timeout number and a counterquestion strategy; the guide dialogue is a fixed dialogue and is used to remind the user of the information to be provided; the timeout time is used to set the time for waiting for the user to upload the information; the timeout number is used to set the number of times for reminding the user to upload the information; and the counterquestion strategy is used in cooperation with the timeout time, if the user does not reply within the timeout time, the guide dialogue is actively sent to the user according to the counterquestion strategy, so as to guide the user to provide the information again, thereby facilitating the collection of the values of the slot entities. Preferably, the counterquestion strategy comprises a custom dialogue and a recent dialogue, when the custom dialogue is selected, the configuration of the counterquestion dialogue is increased, and the counterquestion dialogue is sent to the user; and when the recent dialogue is selected, the dialogue sent last time is automatically sent to the user.

[0040] The node information involved in the small ticket point interaction flow in the embodiment and the parent-child association information between the nodes are all stored in a database, and when the small ticket point service is started, the related records of the small ticket point interaction flow are read from the database and loaded into the memory. Preferably, the redis server is used to store the data loaded in the memory.

[0041] It should be noted that the small ticket element information to be collected in the embodiment comprises a brand name, a mall name, an order number, a transaction time and a transaction amount, and the same number of slot entities as the small ticket element information is set in the dialogue node, and the types of the slot entities are all set as system entities, that is, the small ticket picture uploaded by the user is sent to an optical character recognition model to recognize the small ticket text information, and the small ticket text information is sent to an entity recognition model to obtain the values of the slot entities.

[0042] Specifically, after entering the dialogue node, it is identified whether the slot entities set in the dialogue node have values, if not, the first guide dialogue set is sent to the user to remind the user to upload the small ticket picture. This case corresponds to the situation that the user does not upload the small ticket picture. Figure 2 The dialogue node 1 behind the start node has no value in the five slot entities, and the first guide dialogue is sent, such as "please upload the small ticket picture".

[0043] Further, when the small ticket picture is still not obtained when the set timeout time is reached, the first guide dialogue set is sent to the user again according to the set counterquestion strategy (i.e. the recent dialogue is configured in the counterquestion strategy), and the actual timeout number is recorded, when the actual timeout number exceeds the set timeout number, the dialogue node is exited and the manual auditing system is entered.

[0044] After entering the dialogue node, it is identified whether the slot entities set in the dialogue node have values, if yes, the second guide dialogue set is sent to the user to remind the user to upload the small ticket picture. This case corresponds to the situation that the user uploads the small ticket picture but the small ticket picture is not clear. Figure 2The intermediate dialogue node 2, when the first dialogue node 1 has been executed, collects part or complete ticket element information, but because it is incomplete or the user considers it incorrect, sends a second guiding dialogue such as: "Please upload a clear ticket picture".

[0045] Further, when the ticket picture is obtained within the timeout time, for each ticket element, identify whether the newly obtained information is consistent with the value in the slot point entity, if consistent, according to the newly obtained information, fill in the value in the slot point entity, if inconsistent, cover the newly obtained ticket element information with the value of the corresponding slot point entity. That is, the embodiment performs consistency check on the collected data, ensures that the information of the same ticket element remains consistent in different times of collection, thereby improving the accuracy and reliability of the data.

[0046] It should be noted that the entity recognition model in the embodiment sequentially includes a two-dimensional feature extraction module, a feature fusion module, a decoder module and an output module; wherein the two-dimensional feature extraction module includes two parallel branches, the first branch includes a text embedding layer and an encoder layer connected in sequence; the second branch includes a Bert embedding layer, a width embedding layer and a concatenation layer connected in sequence; the feature fusion module fuses the feature vectors output by the two branches in the two-dimensional feature extraction module through a linear layer, and then transmits them to the decoder module to generate an output sequence, and finally obtains the recognized entity through the output module.

[0047] Specifically, the first branch of the two-dimensional feature extraction module focuses on capturing the context information in the text and understanding the overall structure and semantics of the text; wherein the text embedding layer includes a trainable word embedding matrix, which maps each word in the text to a fixed-size vector space, and then adds position encoding to the encoder layer; the encoder layer includes a plurality of connected encoders, each encoder includes a multi-head self-attention mechanism layer and a feedforward neural network layer connected in sequence, and each layer is connected with a residual connection and a layer normalization, and the encoder layer is used to convert the input sequence into a feature vector containing rich semantic information to obtain a first feature vector.

[0048] The second branch of the two-dimensional feature extraction module focuses on identifying the width and boundary of the entity; wherein the Bert embedding layer is to obtain token embedding vectors and a special token CLS vector covering the whole sentence information by using the Bert model; the width embedding layer includes a trainable width embedding matrix, which contains a fixed size embedding vector for each width (from 1 to the maximum width); the width embedding vector is obtained by indexing the width embedding matrix, that is, the width of the entity is k+1, which means that the entity contains k+1 tokens, and the width embedding vector of the entity is obtained by indexing the embedding vector with a width of k+1 in the width embedding matrix; the token embedding vector, the width embedding vector and the special token CLS vector are spliced into the second feature vector through the splicing layer.

[0049] The feature fusion module fuses the first feature vector and the second feature vector through a full connection layer to obtain a fusion feature vector, which is transmitted to the decoder module.

[0050] The decoder module includes a plurality of sequentially connected decoders, each decoder including a masked multi-head self-attention sublayer, an encoder-decoder attention layer and a feedforward neural network layer connected in sequence; each layer is connected with a residual connection and a layer normalization.

[0051] The output module is mapped to the dimension of the number of labels through a linear layer, and then performs entity labeling through a softmax function or other classification function (such as a CRF layer), and outputs the recognized entity.

[0052] (3) Judgment node

[0053] The judgment node includes a plurality of branches, each branch including at least one condition group, and each condition group including at least one condition; the judgment symbol of the condition includes: greater than, less than, equal to, not equal to, contains, does not contain, has a value and has no value. There must be one default branch in the plurality of branches, which is used to enter the node pointed by the default branch when the conditions of all branches are not met.

[0054] In Figure 2 , the judgment node 1 and the judgment node 3 are used to identify whether the small ticket element information is complete, and two branches are set in the two judgment nodes, a condition group composed of multiple conditions is set in the non-default branch, each condition introduces a slot entity, which corresponds to the brand name, the mall name, the order number, the transaction time and the transaction amount respectively, and the judgment symbol is set to have a value, that is, all five slot entities have values, and correspondingly, the default branch indicates that not all five slot entities have values.

[0055] In Figure 2The judgment node 2 and the judgment node 4 are used to identify whether the ticket element information is correct according to the user feedback information, two branches are also set, and one condition composed of one condition is set for the non-default branch; the condition is set according to the button return value set in the reply node. Figure 2 The parent nodes of the judgment node 2 and the judgment node 4 are the reply nodes, the user feedback information is received in the reply node by setting the button variable and the button return value and stored in the memory, and the condition in the judgment node corresponds to an expected user feedback situation by referencing the button variable and setting the judgment symbol as equal, for example, the user feedback collected ticket element information is correct.

[0056] (4) Reply node

[0057] The reply node includes: a reply type, a reply content type and a reply content; the reply type includes: display and template; the display means directly displaying the reply content according to the reply content type; the template means obtaining dynamic reply content by replacing the variable in the reply template set in the reply content with a dynamic value, and displaying the dynamic reply content according to the reply content type. The reply content type includes: text, picture, audio, video, file and link; the reply content is configured according to the selected reply content type.

[0058] Figure 2 The settings of the reply node 1 and the reply node 2 in the are the same, both of which are set as template for the reply type, text for the reply content type, and the reply template set in the reply content and the slot point entity as a variable, and the values of the five slot point entities are obtained from the memory to replace the variable in the reply template to display the text to the user to confirm whether the collected ticket element information is correct.

[0059] Further, the reply node further includes: whether to enable button setting, when the button setting is enabled, set the button variable, select the button type and set the button label and the button return value; the button setting is used to display the button label according to the button type in the front end to interact with the user. Figure 2 The sub-nodes of the reply node 1 and the reply node 2 in the are judgment nodes, and the judgment nodes need to jump to different branches according to the user feedback on the reply content displayed by the reply node, therefore, the button setting is enabled in the reply node 1 and the reply node 2, the button variable is set, the button type is selected and the button label and the button return value are set, after the user selects the button to feedback, the user feedback information is received and stored in the memory.

[0060] Exemplarily, the button variable is set as btnparam, the button type is selected as radio button, the button label is added as "correct", and the corresponding button return value is "1"; the button label is "incorrect", and the corresponding button return value is "0". Two radio buttons are generated in the front-end interface, and after the user selects, the button return value of the selected button is set to the button variable btnparam and stored in the memory. In the branch condition of the judgment node, btnparam is set to be equal to 1, indicating that the small ticket element information collected by the user feedback is correct.

[0061] It should be noted that, Figure 2 In the embodiment, the user is fed back into the artificial audit system through the reply node 3, at this time, no feedback information of the user needs to be received, and therefore, the button setting in the reply node 3 does not need to be enabled, the reply type is set as display, and the reply content type is set as text or picture. The text content input in the reply content or the picture uploaded is displayed in the front-end interface, and the user is informed to enter the artificial audit system by text content or the subsequent operation process by picture.

[0062] (5) Query node

[0063] The query node includes: setting the request type, the request method, the input parameter and its value, and the output parameter; the request type includes: GET and POST. Figure 2 The query node 1 and the query node 2 in the embodiment are both used to send the small ticket element information to the credit system, that is, by taking the slot point entity and its value as the input parameter and its value, the credit system set in the request method is called according to the set request type.

[0064] (6) End node

[0065] The end node is used to mark the end of the flow running, and there is only one.

[0066] In step S3, the flow running ends when two conditions are met: the user feedbacks that the small ticket element information is correct, the small ticket element information is sent to the credit system, and the user normally completes the credit record; and the correct and complete small ticket element information is not collected for two times in succession, and the artificial audit is entered, and the credit record is further completed.

[0067] (7) Collection node and assignment node

[0068] Preferably, the embodiment further includes the collection node and the assignment node when the interactive flow is constructed.

[0069] The assignment node is used to process the variable, add a new variable and assign a value, or modify the variable value after referring to the existing variable, thereby increasing the flexibility of the flow.

[0070] The collection node is used for collecting information filled by users through a set table, including: designing a table, selecting a table, prompting a speech, a timeout time and a timeout number. The table for collecting information is created in the form of a form through "designing a table", the designed table is selected through "selecting a table", and the user is reminded to enter information in the table through "prompting a speech"; when entering the collection node, the front-end interface displays a table link, and if the user does not complete the filling within the set timeout time, the front-end interface outputs a prompt speech, and if the set timeout number is exceeded, the collection node is exited; otherwise, the next node is entered.

[0071] Compared with the prior art, the dialog interactive service closed-loop small ticket point method provided by the embodiment meets the demand of small ticket points quickly by flexibly constructing an automated interactive process; multi-round dialog is realized through the collaborative work of the dialog node, the judgment node and the reply node in the automated interactive process, complex business scenarios are coped with, and user dialog experience is improved; the small ticket information collected is ensured to be accurate and correct through guiding the user to upload small ticket pictures multiple times, the accuracy of overall small ticket recognition is greatly improved, the process is dynamically executed according to the reply and operation of the customer, the point process in different situations is completed in real time, the interactive service closed-loop small ticket point process is realized, the time and workload of manually processing small ticket points are reduced, the cost of human input is reduced, and the efficiency of point processing is improved.

[0072] Those skilled in the art can understand that all or part of the processes of the above-mentioned embodiment methods can be completed by a computer program instructing related hardware, and the program can be stored in a computer readable storage medium. The computer readable storage medium is a disk, an optical disk, a read-only memory or a random access memory, etc.

[0073] The above is only a preferred specific embodiment of the present application, but the protection scope of the present application is not limited thereto, any person skilled in the art can easily think of changes or replacements within the technical range disclosed by the present application, which should be covered within the protection scope of the present application.

Claims

1. A conversational interactive service closed loop ticketing points method, characterized by, The method comprises the following steps: constructing a small ticket point interaction flow, saving each node and its configuration information and loading them into the memory; The nodes include a dialogue node, a judgment node, a reply node and a query node; the dialogue node is provided with the same number of slot point entities as the small ticket element information, and the types of the slot point entities are all set as system entities, which are obtained by sending the small ticket picture uploaded by the user to an optical character recognition model to recognize small ticket text information, and then sending the small ticket text information to an entity recognition model; the entity recognition model sequentially includes a two-dimensional feature extraction module, a feature fusion module, a decoder module and an output module; the two-dimensional feature extraction module includes two parallel branches, the first branch includes a text embedding layer and an encoder layer connected in sequence, and the second branch includes a Bert embedding layer, a width embedding layer and a concatenation layer connected in sequence; the feature fusion module fuses the feature vectors output by the two branches of the two-dimensional feature extraction module through a linear layer, and then transmits the fused feature vectors to the decoder module to generate an output sequence, and finally obtains the recognized entities through the output module; When the user's intention is identified as being related to small ticket points, the small ticket point interaction flow is started; the user is guided to upload a small ticket picture through the dialogue node, and small ticket element information is obtained and saved; the small ticket element information is identified as being complete through the judgment node, and whether the small ticket element information is correct is identified according to user feedback information; the small ticket element information is fed back to the user through the reply node, and user feedback information is received; When the small ticket element information is identified as being correct according to the user feedback information, the small ticket element information is sent to the point system through the query node, and the small ticket point interaction flow is ended; when the small ticket element information is identified as being incorrect according to the user feedback information for multiple times, the user is fed back to enter the manual review system through the reply node, and the small ticket point interaction flow is ended.

2. The conversational interactive service closed loop ticketing credit method of claim 1 wherein, The dialogue node further includes a guide phrase, a timeout time, a timeout number and a counterquestion strategy; after entering the dialogue node, it is identified whether the slot point entities set in the dialogue node have values; if not, the first guide phrase set is sent to the user to remind the user to upload a small ticket picture; when the small ticket picture is still not obtained after reaching the set timeout time, the first guide phrase set is sent to the user again according to the counterquestion strategy, and the actual timeout number is recorded; when the actual timeout number exceeds the set timeout number, the dialogue node is exited and the manual review system is entered.

3. The conversational interactive service closed loop ticketing credit method of claim 2, wherein, After entering the dialogue node, it is identified whether the slot point entities set in the dialogue node have values; if yes, the second guide phrase set is sent to the user to remind the user to upload a clear small ticket picture; when the small ticket picture is obtained within the timeout time, it is identified whether the newly obtained information of each small ticket element is consistent with the values in the slot point entities; if yes, the values in the slot point entities are supplemented according to the newly obtained information; if not, the newly obtained small ticket element information replaces the values of the corresponding slot point entities.

4. The conversational interactive service closed loop ticketing credit method of claim 1 wherein, When the judging node identifies whether the ticket element information is complete, two branches are set in the judging node, one of which is a default branch, and a condition group composed of multiple conditions is set for the non-default branch; each condition is set by introducing a slot entity and setting the slot entity to have a value.

5. The conversational interactive service closed loop ticketing credit method of claim 1 wherein, When the judging node identifies whether the ticket element information is correct according to user feedback information, two branches are set in the judging node, one of which is a default branch, and a condition group composed of one condition is set for the non-default branch; the condition is set according to the button variable and button return value set in the reply node.

6. The conversational interactive service closed loop ticketing credit method of claim 1, wherein, The reply node includes a reply type, a reply content type and a reply content; by setting the reply type as a template, the reply content type as text, and setting a reply template in the reply content and introducing a slot entity as a variable, the ticket element information is fed back to the user through the reply node by replacing the variable in the reply template with the value of the obtained slot entity and sending the ticket element information to the user.

7. The conversational interactive service closed loop ticketing credit method of claim 1 wherein, The reply node further includes a button setting; when the button setting is enabled, a button variable is set, a button type is selected, and a button label and a button return value are set; the user feedback information received by the reply node is the button return value corresponding to the button selected by the user, and is set in the button variable.

8. The conversational interactive service closed loop ticketing credit method of claim 1 wherein, The query node includes setting a request type, a request method, input parameters and their values, and output parameters; the ticket element information is sent to the credit system through the query node by setting the slot entity and its value as the input parameters and their values, and calling the credit system set in the request method according to the set request type.

Images