WO2022037019A1 - System, method and device for implementing man-machine multi-round conversation - Google Patents

Abstract

Provided are a system, a method and a device for implementing a man-machine multi-round conversation. The method comprises: providing a user configuration interface, a plurality of types of intention nodes being provided in the user configuration interface; constructing an intention unit on the basis of operations, including selecting the intention nodes, configuring the intention nodes and connecting the intention nodes, of a first user on the user configuration interface; and implementing a man-machine multi-round conversation with a second user on the basis of the intention unit, so as to acquire a key intention of the second user in the man-machine multi-round conversation, and achieve the determination and prediction of the conversation.

Description

System, method and device for realizing multi-round dialogue between man and machine

This application claims the priority of the Chinese patent application with the application number of 202010844657.1 and the filing date of August 20, 2020, entitled "Method and Device for Realizing Human-Machine Multi-round Dialogue", wherein the content disclosed in the above application is incorporated by reference in this application.

technical field

The present disclosure relates to the field of artificial intelligence, and more particularly, to a system, method and apparatus for realizing multi-round dialogue between humans and machines.

Background technique

With the development of artificial intelligence technology, the application of human-machine multi-round dialogue is becoming more and more extensive, such as intelligent dialogue robots for hotel reservations, air tickets, etc. The purpose of multi-round human-machine dialogue is to obtain the necessary information through a series of questions after obtaining the user's initial intention to finally provide information and services corresponding to the user's intention.

In the multi-round human-machine dialogue, users usually come with initial intentions, hoping to obtain information or services corresponding to their intentions (for example, booking hotels, booking tickets, booking air tickets, information consultation, etc.). However, in the face of complex dialogue situations of users, intelligent dialogue bots are often difficult to deal with flexibly, and cannot effectively acquire the key intentions of users, thus failing to provide accurate information and services.

SUMMARY OF THE INVENTION

The purpose of the present disclosure is to provide a system, method and device for realizing multi-turn human-machine dialogue.

A first aspect of the present disclosure provides a system for implementing multi-turn human-machine dialogue including at least one computing device and at least one storage device, wherein the at least one storage device stores instructions, and the instructions, when executed by the at least one computing device, cause the At least one computing device executes the following steps of a method for implementing a multi-round human-machine dialogue: providing a user configuration interface, wherein multiple types of intent nodes are provided in the user configuration interface; selecting an intent node on the user configuration interface based on the first user, Configure the intent node and the operation of connecting the intent node, and construct the intent unit; realize the multi-round dialogue with the second user based on the intent unit, so as to obtain the key intention of the second user in the multi-round dialogue between the human and the computer, and realize the judgment and prediction of the dialogue .

A second aspect of the present disclosure provides a method for implementing a multi-turn human-machine dialogue, the method comprising: providing a user configuration interface, wherein multiple types of intent nodes are provided in the user configuration interface; based on the first user On the user configuration interface, select the intent node, configure the intent node, and connect the intent node to construct an intent unit; based on the intent unit, implement multiple rounds of human-machine dialogue with the second user to obtain the second user's human-computer Key intentions in multiple rounds of dialogue to realize judgment and prediction of dialogue.

A third aspect of the present disclosure provides an apparatus for implementing a multi-turn human-machine dialogue, the apparatus comprising: a user configuration interface providing unit for providing a user configuration interface, wherein the user configuration interface providing unit provides in the user configuration interface multiple types of intent nodes; an intent unit construction unit, based on the operation of the first user to select an intent node, configure an intent node, and connect an intent node on the user configuration interface, construct an intent unit; a man-machine multi-round dialogue implementation unit, Based on the intention unit, a multi-round dialogue with the second user is implemented, so as to obtain the key intention of the second user in the multi-round dialogue between the human and the computer, so as to realize the judgment and prediction of the dialogue.

A fourth aspect of the present disclosure provides a computer-readable storage medium having stored thereon a computer program that, when executed by one or more computing devices, causes the one or more computing devices The apparatus implements any of the methods described above.

The present disclosure can flexibly select intent nodes, configure intent nodes, and connect intent nodes on the user configuration interface according to the first user's own business needs and business experience, so as to construct the second user-oriented intent for the dialog robot program. unit. Therefore, the dialog robot program can flexibly respond to the user's complex dialog situation based on the intent unit oriented to the second user, and effectively acquire the user's key intent, thereby providing accurate information and services.

Description of drawings

The above and other objects and features of the present disclosure will become more apparent from the following description in conjunction with the accompanying drawings, which illustrate an example, wherein:

1 shows a flowchart of a method for implementing a multi-round dialogue between humans and machines according to an embodiment of the present disclosure;

2 shows a schematic diagram of constructing an intent unit in a user configuration interface according to an embodiment of the present disclosure;

3 shows a schematic diagram of a configuration of an input node according to an embodiment of the present disclosure;

4 shows a schematic diagram of a configuration of a judgment node according to an embodiment of the present disclosure;

5 shows a schematic diagram of the configuration of an input node according to an embodiment of the present disclosure;

6 shows a schematic diagram of the configuration of an assignment node according to an embodiment of the present disclosure;

7 shows a schematic diagram of the configuration of an interface node according to an embodiment of the present disclosure;

FIG. 8 shows an apparatus for realizing multi-round dialogue between humans and machines according to an embodiment of the present disclosure;

FIG. 9 shows an example of a multi-turn dialogue between humans and machines according to an embodiment of the present disclosure;

Fig. 10A shows the construction intent unit of the first round of dialogue for the example of the man-machine multi-round dialogue shown in Fig. 9;

10B shows the configuration of the input node and the configuration of the judgment node of the construction intent unit shown in 10A;

10C shows the configuration of the output node corresponding to parameter logic 1 and the configuration of the output node corresponding to parameter logic 2 and the configuration of adding task parameters of the build intent unit shown in 10A;

10D illustrates a configuration interface for adding task parameters according to an embodiment of the present disclosure;

FIG. 11A shows a further constructed intent unit for the second-round dialogue of the example of the man-machine multi-round dialogue shown in FIG. 9;

11B shows the configuration of the input node and the configuration of the judgment node of the construction intent unit shown in 11A;

11C shows the configuration of the output node corresponding to parameter logic 1 and the configuration of the output node corresponding to parameter logic 2 of the build intent unit shown in 11A;

FIG. 12A shows a further constructed intent unit for the third-round dialogue of the example of the man-machine multi-round dialogue shown in FIG. 9;

12B shows the configuration of the input node and the configuration of the judgment node of the construction intent unit shown in 12A;

12C shows the configuration of the output node corresponding to parameter logic 1 and the configuration of the output node corresponding to parameter logic 2 of the build intent unit shown in 12A;

FIG. 13A shows a further constructed intent unit for the fourth round of the example of the man-machine multi-round dialogue shown in FIG. 9;

13B shows the configuration of the input node and the configuration of the judgment node of the construction intent unit shown in 13A;

13C shows the configuration of the output node corresponding to parameter logic 1 and the configuration of the output node corresponding to parameter logic 2 of the build intent unit shown in 13A;

FIG. 14A shows a further constructed intent unit for the fifth-round dialogue and the sixth-round dialogue of the example of the human-machine multi-round dialogue shown in FIG. 9;

Figure 14B shows the configuration of the fifth-round dialogue-based input node, the assignment node, and the judgment node with the construction intent unit shown in 14A;

Fig. 14C shows the configuration of the output node corresponding to parameter logic 1 and the configuration of the output node corresponding to parameter logic 2 based on the fifth round of dialogue with the construction intent unit shown in 14A;

14D shows the configuration of the sixth-round dialogue-based input node and judgment node of the construction intent unit shown in 14A;

Fig. 14E shows the configuration of the output node corresponding to parameter logic 1 and the configuration of the output node corresponding to parameter logic 2 based on the sixth round of dialogue of the construction intent unit shown in 14A;

FIG. 15A shows a further constructed intent unit for the eighth-round dialogue of the example of the man-machine multi-round dialogue shown in FIG. 9;

15B shows the configuration of the input node and the configuration of the judgment node of the construction intent unit shown in 15A;

Figure 15C shows the configuration of the output node corresponding to parameter logic 1 and the configuration of the output node corresponding to parameter logic 2 and the assignment node of the build intent unit shown in 15A;

FIG. 16A shows a further constructed intent unit for the ninth round of the example of the man-machine multi-round dialogue shown in FIG. 9;

16B shows the configuration of the input node and the configuration of the judgment node of the construction intent unit shown in 16A;

FIG. 16C shows the configuration of the output node corresponding to the parameter logic 1 and the configuration of the output node corresponding to the parameter logic 2 and the assignment node of the build intent unit shown in 16A.

detailed description

In order for those skilled in the art to better understand the present disclosure, the exemplary embodiments of the present disclosure will be described in further detail below in conjunction with the accompanying drawings and the specific embodiments, so as to help a comprehensive understanding of the examples of the present disclosure defined by the claims and their equivalents. Sexual Example. The description includes various specific details to assist in that understanding, but these details are considered to be exemplary only. Accordingly, those of ordinary skill in the art will recognize that various changes and modifications of the embodiments described herein can be made without departing from the scope and spirit of the present disclosure. Also, descriptions of well-known functions and constructions may be omitted for clarity and conciseness. It should be noted here that "at least one of several items" in the present disclosure all means including "any one of the several items", "a combination of any of the several items", The three categories of "the whole of the several items" are juxtaposed. In the present disclosure, "and/or" all means at least one of the preceding two or more items joined by it. For example, "including at least one of A and B" and "including A and/or B" include the following three parallel situations: (1) including A; (2) including B; (3) including A and B . For another example, "execute at least one of step 1 and step 2", "execute step 1 and/or step 2" means the following three parallel situations: (1) execute step 1; (2) execute step 2; (3) Execute step one and step two. That is to say, "A and/or B" can also be expressed as "at least one of A and B", and "execute step 1 and/or step 2" can also be expressed as "execute step 1 and step 2" at least one of".

FIG. 1 shows a flowchart of a method for implementing a multi-turn human-machine dialogue according to an embodiment of the present disclosure.

Referring to FIG. 1, in step S110, a user configuration interface is provided, wherein the user configuration interface provides multiple types of intent nodes.

Here, different types of intent nodes can be used to perform different actions. According to an embodiment of the present disclosure, the various types of intent nodes include one or more of the following: an input node, a judgment node, an output node, an assignment node, and an interface node. These nodes will be described in more detail later.

In step S120, an intent unit is constructed based on the first user's operations of selecting an intent node, configuring an intent node, and connecting an intent node on the user configuration interface.

That is, the first user may be the user who builds the intent unit. For example, the first user may be a service provider that provides information or services (eg, hotel reservation, bus ticket reservation, air ticket reservation, information consultation, etc.).

Here, the first user can flexibly choose to select intent nodes, configure intent nodes, and connect intent nodes on the user configuration interface according to their own business needs and business experience, so as to build a flexible second-user-oriented dialog robot program. Intent unit.

For example, the first user may be based on historical data (eg, related to needs, habits, and/or preferences, etc.) based on business context (eg, its target users (eg, second users using information or services provided by the first user) data)) to select intent nodes, configure intent nodes, and connect intent nodes on the user configuration interface, thereby constructing a flexible intent unit for the second user. However, the present disclosure is not limited thereto, and the first user may also select intent nodes, configure intent nodes, and connect intent nodes on the user configuration interface according to any other method, thereby constructing a flexible intent unit for the second user. In the present disclosure, a single intent unit can execute multiple rounds of a single-intent human-machine dialogue.

In step S130, based on the intention unit, the human-machine multi-round dialogue with the second user is realized, so as to obtain the key intention of the second user in the human-machine multi-round dialogue, so as to realize the judgment and prediction of the dialogue.

Since the intent unit is flexibly configured by the first user as an intent unit for the second user used by the conversational robot program, the conversational robot program can flexibly and effectively implement multi-turn human-machine conversations with the second user based on the intent unit, so as to Obtain explicit instructions from the second user.

As an example, the method of implementing a multi-turn dialogue between humans and machines may be performed by a dialogue robot program that provides services or information.

FIG. 2 shows a schematic diagram of constructing an intent unit in a user configuration interface according to an embodiment of the present disclosure.

Referring to FIG. 2, in FIG. 2, a user configuration interface 200 may include a first area 210 in which various types of intent nodes are arranged and a second area 220 for constructing an intent unit.

Here, please note that the sizes and positions of the first area 210 and the second area 220 may be arbitrarily set, and are not limited to the example in FIG. 2 . Furthermore, although FIG. 2 shows five types of intent nodes (ie, input nodes, judgment nodes, output nodes, assignment nodes, and interface nodes), the intent nodes of the present disclosure are not limited to the above-mentioned nodes, or may include input nodes , one or more of a judgment node, an output node, an assignment node, and an interface node. In addition, all the intent nodes here can be configured by the first user, so personalized services or functions can be provided according to the needs of the first user.

Here, as an example, when a first user wants to build an intent unit, the first user may select (eg, drag, click, etc.) an intent node from the first area, so that the selected corresponding intent node appears in the second area. In addition, the first user can flexibly configure the selected intent node, so that the intent node can perform respective functions or operations according to the flexible configuration, so that the key intent of the second user can be acquired and the usage requirements or intent requirements of the second user can be met. .

In addition, the first user can also connect the selected intent nodes, so that the connected intent nodes can perform operations corresponding to the intent units according to the connected logical relationship.

As a non-limiting example, when an intent node is selected, a bezier curve can appear until the next intent node is dragged in, the front and rear nodes can be automatically connected, and the connected nodes and arrows can be changed. However, the above examples are non-limiting, and the present disclosure does not limit the specific selection manner and connection manner of the intended nodes.

Since the first user can flexibly configure the intent unit for the second user according to business scenarios or actual needs, the dialogue robot program with the flexibly configured intent unit can effectively provide the second user with information or services consistent with his intent.

FIG. 3 shows a schematic diagram of the configuration of an input node according to an embodiment of the present disclosure.

Here, please note that the configuration of the input nodes shown in FIG. 3 is schematic, and the present disclosure is not limited thereto. An input node of the present disclosure may include one or more of the configuration items shown in FIG. 3 , or may include other configuration items.

In an embodiment of the present disclosure, the configuration items of the input node include: a trigger method for triggering a task, a method similar to the trigger method, a slot, and a task parameter, wherein the slot is based on the trigger method and the Trigger-like questions are set, wherein task parameters are set based on trigger questions and/or trigger-like questions.

Referring to FIG. 3, the configuration item of the input node may include a trigger method for triggering the task "hotel reservation" (eg, the second user talks to the dialog robot program "please book hotel A for me" and "book hotel for me") and similar questions to trigger questions (eg, the second user talks to the conversational bot "I want to book a hotel" and "help me book a hotel").

In addition, the configuration item of the input node may include the slot "hotel name" set based on the triggering question "please help me book hotel A". In this case, the information of the slot "hotel name" may be "hotel A".

In addition, the configuration items of the input node may also include task parameters (eg, hotel name, check-in time, etc.) that are set based on the trigger method and/or a method similar to the trigger method.

During the process of configuring the input node by the first user, the first user can also delete the configured image or add a new configuration item according to the situation, so that the configuration item can meet the usage requirement of the second user.

Preferably, the configuration item of the input node may further include a jumping method for jumping to other input nodes. Therefore, even when the current dialog information of the second user does not match the current input node, the current dialog information of the second user can be jumped to the corresponding input node by using the jumping method for jumping to other input nodes, Therefore, the multi-round man-machine dialogue can be carried out normally, and the smooth use of the second user can be ensured.

In one example, the initial input node does not have a forward jump query setting.

Furthermore, throughout the drawings, the symbol of the pen may indicate that information can be entered, and the trash can indicate that it can be deleted.

FIG. 4 shows a schematic diagram of a configuration of a judgment node according to an embodiment of the present disclosure.

Here, please note that the configuration of the judgment node shown in FIG. 4 is schematic, and the present disclosure is not limited thereto. The judgment node of the present disclosure may include one or more of the configuration items shown in FIG. 4 , or may include other configuration items.

In one embodiment, the configuration item of the judgment node includes parameter logic, wherein the parameter logic performs logical condition judgment on the value obtained by the previous node and the known value or the user-defined value, and determines the judgment node based on the mutually exclusive judgment results the next node.

Preferably, the parameter logic may include parameter logic corresponding to an empty slot. Therefore, when the slot of the second user is empty in the multi-round man-machine dialogue, the dialogue robot program can also provide a corresponding logical judgment mechanism, so that the multi-round man-machine dialogue can proceed smoothly.

Referring to FIG. 4 , the configuration items of the judgment node include parameter logic 1 and parameter logic 2 . Here, parameter logic 1 may correspond to the case where "hotel name" is empty, and parameter logic 2 may correspond to the case where "hotel name" is A hotel. The first user can save the configured parameter logic through the button "Save Logic". At this time, the task of the judgment node can be configured as "hotel". Here, parameter logic 1 and parameter logic 2 are parallel parameter logics.

For example, when the second user talks to the dialogue robot program "Please help me book hotel A", the judgment node can judge that the parameter logic 1 (ie, hotel name=[A hotel]) is established, and then judge the following according to the parameter logic 1. a node. For another example, when the second user talks "please book a hotel for me" to the dialogue robot program, the judgment node can judge that parameter logic 2 (ie, the hotel name is empty) is established, and then judge the next node according to parameter logic 2.

In one example, the configuration items of the parameter logic may include a user-defined judgment value, a user-defined logical relationship, and a user-defined reference value, wherein the parameter logic is configured to determine whether the user-defined judgment value and the user-defined reference value satisfy the user-defined logic relationship , and determine the next node of the judgment node based on the determined result.

In other words, when the parameter logic determines that the custom judgment value and the custom reference value satisfy the custom logic relationship, the next node of the judgment node is determined to correspond to the situation where the custom judgment value and the custom reference value satisfy the custom logic relationship When the parameter logic determines that the user-defined judgment value and the user-defined reference value do not satisfy the user-defined logical relationship, the next node of the judgment node is determined to correspond to the user-defined judgment value and the user-defined reference value that do not satisfy the self-defined logical relationship. The next node of the case that defines the logical relationship.

Here, the custom judgment value includes at least one of the task parameter value in the data received from the previous node, the parameter value in the parameter dictionary, and the interface return value. Here, the parameter dictionary may be a pre-designed database including parameter values. The interface return value can be a value or data set returned from the interface node. Note that the parameter dictionary cannot use the logical relationship "equals" because it has multiple parameter values.

Additionally, the custom reference value may include at least one of a set value in the parameter dictionary, an interface return value, and a custom value. The custom logical relationship is the logical relationship between the custom judgment value set by the first user according to the intent node and the custom reference value.

As non-limiting examples, custom logical relationships may include "is empty", "not empty", "equal to", "not equal to", "greater than", "less than", "belongs to", "does not belong to", " includes", "excluding", "and" and/or "or".

FIG. 5 shows a schematic diagram of a configuration of an input node according to an embodiment of the present disclosure.

Here, please note that the configuration of the output nodes shown in FIG. 5 is schematic, and the present disclosure is not limited thereto. The output node of the present disclosure may include one or more of the configuration items shown in FIG. 5 , or may include other configuration items.

According to an embodiment of the present disclosure, the configuration item of the output node includes a configuration item of outputting the reply content to the second user in a predetermined form.

Referring to Figure 5, the output node is configured to perform the task "Robot asks for hotel name". Here, the first user can configure multiple reply contents. For example, the first user can configure the reply content "Which hotel would you like to book?". In one example, the first user may configure the reply content based on the second user's intent and the context of the conversation.

In addition, referring to FIG. 5 , the predetermined form of outputting the reply content to the second user may include pictures, voices, links, menus and rich texts.

In one example, when the second user conducts multiple rounds of man-machine dialogue with the dialogue robot program through the display screen of the robot to provide information query services, the dialogue robot program can display pictures, links, menus and rich text through the display screen. The reply content of the dialogue robot program is output to the second user. In another example, when the second user conducts multiple rounds of man-machine dialogue with the dialogue robot program through the voice dialogue module of the dialogue robot program to provide customer service, the dialogue robot program can use the voice dialogue module to speak to the second user in the form of voice. The user outputs the content of the reply from the dialog bot.

However, the above examples are non-limiting and any other predetermined form is possible.

Among them, the text box in the default state of the content editor of the output node also supports adding the obtained task parameters and the output parameters of the interface node, that is, the return value of the interface, and replying to more personalized words. For example, reply: OK, {time}, {departure airport} to {destination airport} {flight} has been booked for you. (The content in the symbol {} represents the parameter value. In this example, {time}, {departure airport}, {destination airport} are task parameters, and {flight} is the return value of the interface.)

In other words, the output node can be configured to provide multi-form replies based on task parameters and interface return values to meet various needs of the second user.

FIG. 6 shows a schematic diagram of the configuration of a value node according to an embodiment of the present disclosure.

Here, please note that the configuration of the assignment node shown in FIG. 6 is schematic, and the present disclosure is not limited thereto. The value node of the present disclosure may include one or more of the configuration items shown in FIG. 6 , or may include other configuration items.

According to an embodiment of the present disclosure, the configuration item of the assignment node may include at least one of parameter reset and self-defined assignment, wherein the parameter reset includes changing the slot information based on the data received from the previous node of the assignment node, and automatically Definition assignments are used to support the user tag functionality. Since the custom assignment is configured to support the user tag function, the dialogue robot program supports the business tag function in the scene node while supporting the business scenario, so as to realize the data precipitation and feedback of the user portrait.

Referring to Figure 6, an assignment node is configured to perform the task "client type". Here, the slot can be obtained by the previous node. In Figure 6, the parameter reset may include hotel name or check-in time.

Furthermore, in FIG. 6, as an example, a custom assignment may be configured as a customer type of the second user, wherein the customer type is shown to include intended users, potential users, and invalid customers. Therefore, in the multi-round man-machine dialogue of the second user, the effective classification of customers can be realized.

FIG. 7 shows a schematic diagram of the configuration of an interface node according to an embodiment of the present disclosure.

Here, please note that the configuration of the interface node shown in FIG. 7 is schematic, and the present disclosure is not limited thereto. The interface node of the present disclosure may include one or more of the configuration items shown in FIG. 7 , or may include other configuration items.

According to an embodiment of the present disclosure, the configuration item of the interface node includes a configuration item configured to support interface information retrieval configuration.

Referring to FIG. 7, the interface node may be configured to correspond to the task "vacancy query". Here, the configuration item "interface name" of the interface node can be configured to call the interface corresponding to the "interface name", and here, the configuration item "URL" of the interface node can be configured to call the uniform resource locator corresponding to the "URL" . The configuration item "request type" of the interface node can be selected as different types according to the configuration of the first user. Currently, HTTP_GET and HTTP_POST are provided.

In addition, the first user can better provide the interface return value by setting input parameters and output parameters according to the regulations corresponding to the intention of the second user. For example, when the second user intends to book a hotel, the first user can configure the parameter values obtained from the user (for example, information about hotel name, check-in time, room type, etc.), and the configuration items about the output parameters can be based on Enter the configuration item of the parameter to configure. (For example, based on the hotel name, check-in time, check-in room type and other information, the vacancy information that meets the corresponding conditions is called from the interface)

FIG. 8 shows an apparatus for realizing multi-turn dialogue between humans and machines according to an embodiment of the present disclosure.

Referring to FIG. 8 , the apparatus 800 for implementing a multi-turn human-machine dialogue may include a user configuration interface providing unit 810 , an intention unit constructing unit 820 and a human-machine multi-turn dialogue implementing unit 830 .

The user configuration interface providing unit 810 may provide a user configuration interface, wherein the user configuration interface providing unit 810 provides various types of intent nodes in the user configuration interface. The intent unit construction unit 820 may construct the intent unit based on the operations of the first user to select the intent node, configure the intent node, and connect the intent node on the user configuration interface. The man-machine multi-turn dialogue realization unit 830 may implement the man-machine multi-turn dialogue with the second user based on the intention unit, so as to obtain explicit instructions of the second user.

In other words, the user configuration interface providing unit 810 may perform step 110 described with reference to FIG. 1 , the intent unit building unit 820 may perform step 120 described with reference to FIG. 1 , and the human-machine multi-turn dialogue implementing unit 830 may perform step 130 described with reference to FIG. 1 . . Therefore, the user configuration interface and various types of nodes described with reference to FIG. 2 to FIG. 7 are also applicable to the apparatus 800 of FIG. 8 for realizing multi-round dialogue between humans and machines. For brevity, the same description is not repeated here.

FIG. 9 shows an example of a human-machine multi-turn dialogue according to an embodiment of the present disclosure.

Referring to FIG. 9 , an example of a human-machine multi-turn dialogue (or a target dialogue) is shown as flight booking, however, the present disclosure is not limited thereto, and the example of the human-machine multi-turn dialogue may also include hotel reservation, train ticket reservation, information consultation, etc. .

In FIG. 9 , USER may represent a second user who conducts multiple rounds of man-machine dialogue with the dialogue robot program ROBT.

An example of the configuration intent unit based on the example of the man-machine multi-turn dialogue shown in FIG. 9 will be described later with reference to FIG. 10A . Note that the examples shown in FIGS. 9 to 9 are for illustrative purposes, and the configurations of intent units and intent nodes of the present disclosure are not limited to the above examples. Furthermore, it should be understood that the configurations of the intent nodes described with reference to FIGS. 10A to 7 are examples of the intent nodes described with reference to FIGS. 3 to 7 , and thus, the descriptions of the configurations of the intent nodes described with reference to FIGS. 3 to 7 also apply to the graphs 10A to Figures describe the configuration of the intent node. Therefore, for the sake of brevity of description, repeated descriptions will be omitted.

FIG. 10A shows the construction intent unit for the first round of the example of the man-machine multi-round dialogue shown in FIG. 9 ; FIG. 10B shows the configuration of the input node and the configuration of the judgment node with the construction intent unit shown in 10A ; FIG. 10C shows the configuration of the output node corresponding to the parameter logic 1 and the configuration of the output node corresponding to the parameter logic 2 and the configuration of adding task parameters of the construction intent unit shown in 10A.

The dialogue content of the second user in the first round of dialogue in the example of the man-machine multi-round dialogue shown in FIG. 9 is "please book a flight ticket for me". The first user can construct the intent unit shown in FIG. 10A based on the dialogue content of the second user "please book a flight ticket for me".

Referring to Figure 10B, the input node is configured to perform the task "book a flight". The second user may configure the configuration item "Trigger Question" of the input node to include "Please book a ticket for me" and "Book a ticket for me", and configure the configuration item "Similar Question" of the input node to include "book a flight". Here, the second user can also configure the slot "votes". According to the dialogue content of the second user "please book a ticket for me", the information of the slot may be "one". In the case where the second user's dialogue content is "please book a ticket for me", the jump question method is not triggered. After configuring the input node, the first user can click OK to keep the corresponding configuration.

In Figure 10B, the decision node may be configured to perform the task "vote count decision". The second user can configure the parameter logic to include parameter logic 1 and parameter logic 2, where parameter logic 1 is configured to indicate the "votes are empty" condition, and parameter logic 2 is configured to indicate "the votes belong to (parameter dictionary) numbers. Dictionary" case. Here, for example, a digital dictionary as an example of a parameter dictionary may record information on the number of votes.

Referring to FIGS. 10A and 10C , the intent unit may be configured to include two output nodes (ie, an output node corresponding to parameter logic 1 and an output node corresponding to parameter logic 2).

In Figure 1OC, the output node corresponding to parameter logic 1 may be configured to perform the task "parameter logic 1 votes". The output node corresponding to parameter logic 1 may be configured to output "Please confirm the number of votes you need" to the second user. For example, the reply content may be output using one or more of the output forms shown in FIG. 10C. Furthermore, the output node corresponding to parameter logic 2 may be configured to perform the task "parameter logic 2 origin". The output node corresponding to parameter logic 2 may be configured to output "Where is your origin?" to the second user.

With reference to the example of FIG. 9 , when the dialogue content of the second user is “please book a ticket for me”, the dialogue robot program can output the content to the first user based on the intent units configured in FIGS. 10A to 10C , “May I ask your departure Where is it?".

FIG. 10D illustrates a configuration interface for adding task parameters according to an embodiment of the present disclosure.

Referring to FIG. 10D , when the second user clicks on “Task Parameters”, “Add Task Parameters” may pop up. In "Add Task Parameters", parameter names can be configured, and parameter dictionaries (eg, system numbers) can be configured from built-in dictionary 1, built-in dictionary 2, user dictionary 1, and user dictionary 2. In addition, parameter dictionaries can be added via "Parameter Dictionary+".

Fig. 11A shows a further constructed intent unit for the second-round dialogue of the example of the man-machine multi-round dialogue shown in Fig. 9; Fig. 11B illustrates the configuration and judgment node of the input node with the constructing intent unit shown in 11A 11C shows the configuration of the output node corresponding to the parameter logic 1 and the configuration of the output node corresponding to the parameter logic 2 of the build intent unit shown in 11A.

The dialogue content of the second user in the second round of dialogue in the example of the man-machine multi-round dialogue shown in FIG. 9 is "Beijing". The first user may construct the intent unit shown in FIG. 11A based on the conversation content "Beijing" of the second user. Specifically, the portion of the newly constructed intent unit of FIG. 11A compared to FIG. 10A may correspond to the portion in the block in FIG. 11A .

Referring to Figure 11B, the input node is configured to perform the task "departure city". The second user can configure the configuration item "trigger question method" of the input node to include "Beijing" and "book air tickets for me", and configure the configuration item "similar question method" of the input node to include "book air tickets". Here, the second user can also configure the "departure place" of the slot, and the information of the slot can be the system location (eg, Beijing). In addition, according to the previous dialogue content of the second user "please book a ticket for me", the slot may further include the number of tickets, and the information of the slot may be "one ticket". In the case that the content of the dialogue of the second user is content related to booking an air ticket, the jumping question method is triggered, and the user jumps to the previous input node corresponding to the air ticket booking. After configuring the input node, the first user can click OK to keep the corresponding configuration.

In FIG. 11B, the decision node may be configured to perform the task "judgment at origin". The second user can configure the parameter logic to include parameter logic 1 and parameter logic 2, where parameter logic 1 is configured to indicate the "Origin is empty" condition, and parameter logic 2 is configured to indicate "The departure belongs to (parameter dictionary) ) dictionary of the place of departure". Here, for example, the departure place dictionary may record information on the departure place.

Referring to FIGS. 11A and 11C , the intent unit may be configured to include two output nodes (ie, an output node corresponding to parameter logic 1 and an output node corresponding to parameter logic 2 ).

In Figure 11C, the output node corresponding to parameter logic 1 may be configured to perform the task "parameter logic 1 output". The output node corresponding to the parameter logic 1 may be configured to output "Please confirm your origin and tell me again" to the second user. For example, the reply content may be output using one or more of the output forms shown in FIG. 11C. Additionally, the output node corresponding to parameter logic 2 may be configured to perform the task "parameter logic 2 output". The output node corresponding to parameter logic 2 may be configured to output "Which airport are you departing from, Beijing, please?" to the second user.

With reference to the example of FIG. 9 , when the dialogue content of the second round of dialogue of the second user is “Beijing”, the dialogue robot program can output the content to the first user based on the intent unit configured in FIG. 11A to FIG. Departing from the airport?".

Fig. 12A shows a further constructed intent unit for the third round of the dialogue example shown in Fig. 9; 12C shows the configuration of the output node corresponding to the parameter logic 1 and the configuration of the output node corresponding to the parameter logic 2 of the build intent unit shown in 12A.

The dialogue content of the second user in the third round of dialogue of the man-machine multi-round dialogue example shown in FIG. 9 is "Capital International Airport". The first user may construct the intent unit shown in FIG. 12A based on the dialogue content "capital international airport" of the second user. In particular, the portion of the intent unit newly constructed in FIG. 12A compared to FIG. 11A may correspond to the portion in the block in FIG. 12A .

Referring to Figure 12B, the input node is configured to perform the task "Departure Airport". The second user can configure the configuration item "trigger question method" of the input node to include "capital international airport" and "I want to go to the capital international airport", and configure the configuration item "similar question method" of the input node to include "I The departure airport is the Capital International Airport”. Here, the second user can also configure the slot "Beijing Airport". For example, according to the dialogue content of the second user, the information of the slot may be Capital International Airport. In the case where the dialogue content of the second user is the content related to the flight booking or the content related to the departure city, the jumping question method is triggered, and jumps to the previous input node corresponding to the flight booking or related to the departure city input node. After configuring the input node, the first user can click OK to keep the corresponding configuration.

In Figure 12B, the decision node may be configured to perform the task "departure airport decision". The second user may configure the parameter logic to include parameter logic 1 and parameter logic 2, wherein parameter logic 1 is configured to indicate a "departure is empty" situation or a "Beijing airport does not belong to the Beijing airport dictionary" situation. Parameter logic 2 is configured to indicate the case of "Beijing airport belongs to Beijing airport dictionary". Here, for example, the Beijing airport dictionary may record information on airports in Beijing.

12A and 12C, the intent unit may be configured to include two output nodes (ie, an output node corresponding to parameter logic 1 and an output node corresponding to parameter logic 2).

In Figure 12C, the output node corresponding to parameter logic 1 may be configured to perform the task "parameter logic 1 output". The output node corresponding to parameter logic 1 may be configured to output "Please confirm your departure airport and tell me again" to the second user. For example, the reply content may be output in one or more of the output forms shown in Figure 12C. Additionally, the output node corresponding to parameter logic 2 may be configured to perform the task "parameter logic 2 output". The output node corresponding to parameter logic 2 may be configured to output "What is your destination?" to the second user.

With reference to the example of FIG. 9 , when the dialogue content of the second user’s third round of dialogue is “Capital International Airport”, the dialogue robot program can output the content “May I ask your What is the destination?".

Fig. 13A shows a further constructed intent unit for the fourth round of the dialogue example shown in Fig. 9; 13C shows the configuration of the output node corresponding to the parameter logic 1 and the configuration of the output node corresponding to the parameter logic 2 of the build intent unit shown in 13A.

The dialogue content of the second user in the fourth round of dialogue in the example of the man-machine multi-round dialogue shown in FIG. 9 is "Hangzhou". The first user may construct the intent unit shown in FIG. 13A based on the dialogue content "Hangzhou" of the second user. Specifically, the portion of the newly constructed intent unit of FIG. 13A compared to FIG. 12A may correspond to the portion in the block in FIG. 13A .

Referring to Figure 13B, the input node is configured to perform the task "Destination City". The second user can configure the configuration item "trigger question method" of the input node to include "Hangzhou" and "I want to go to Hangzhou", and configure the configuration item "similar question method" of the input node to include "my destination is Hangzhou". Here, the second user can also configure the slot "destination". For example, according to the dialogue content of the second user, the information of the slot may be Hangzhou. When the content of the dialogue of the second user is the content related to the flight booking, the content related to the departure city or the content related to the departure airport, the jumping question method is triggered, and jumps to the previous corresponding to the flight booking. An input node, an input node associated with the departure city, or an input node associated with the departure airport. After configuring the input node, the first user can click OK to keep the corresponding configuration.

In Figure 13B, the decision node may be configured to perform the task "Destination City Decision". The second user may configure the parameter logic to include parameter logic 1 and parameter logic 2, where parameter logic 1 is configured to indicate a "destination is empty" condition or a "destination does not belong to the destination dictionary" condition. Parameter logic 2 is configured to indicate the "destination belongs to the destination dictionary" case. Here, for example, the destination dictionary may record information on the destination.

13A and 13C, the intent unit may be configured to include two output nodes (ie, an output node corresponding to parameter logic 1 and an output node corresponding to parameter logic 2).

In Figure 13C, the output node corresponding to parameter logic 1 may be configured to perform the task "parameter logic 1 output". The output node corresponding to parameter logic 1 may be configured to output "Please confirm your destination and tell me again" to the second user. For example, the reply content may be output using one or more of the output forms shown in Figure 13C. Additionally, the output node corresponding to parameter logic 2 may be configured to perform the task "parameter logic 2 output". The output node corresponding to parameter logic 2 may be configured to output "Which airport in Hangzhou are you going to?" to the second user.

With reference to the example of FIG. 9 , when the dialogue content of the fourth round of dialogue of the second user is “Hangzhou”, the dialogue robot program can output the content to the first user based on the intent units configured in FIG. 13A to FIG. which airport?".

Fig. 14A shows a further constructed intent unit for the fifth round of dialogue and sixth round of the example of the man-machine multi-round dialogue shown in Fig. 9; The configuration of the input node, the assignment node and the judgment node of the five-round dialogue; FIG. 14C shows the configuration of the output node corresponding to the parameter logic 1 based on the fifth round of dialogue with the construction intent unit shown in 14A, and the parameter The configuration of the output node corresponding to logic 2.

The dialogue content of the second user in the fifth round of dialogue in the example of the multi-round dialogue between humans and machines shown in FIG. 9 is "I remembered it wrongly, instead of going to Hangzhou, I will go to Shanghai instead". The first user may construct the intent unit shown in FIG. 14A based on the dialogue content of the second user, "I remembered wrong, I will not go to Hangzhou, but go to Shanghai instead." Specifically, since the intent of the second user has changed, the intent node is reset, and processing proceeds to the input node in FIG. 14B that is connected to the reset intent.

Referring to Figure 14B, the input node is configured to perform the task "Destination City". The second user can configure the configuration item "triggering method" of the input node to include "go to Shanghai instead of Hangzhou", and configure the configuration item "similar method" of the input node to include "my destination is Shanghai" ". Here, the second user can also configure the slot "destination". For example, according to the conversation content of the second user, the information of the slot may be Shanghai. When the content of the dialogue of the second user is the content related to the flight booking, the content related to the departure city or the content related to the departure airport, the jumping question method is triggered, and jumps to the previous corresponding to the flight booking. An input node, an input node associated with the departure city, or an input node associated with the departure airport. After configuring the input node, the first user can click OK to keep the corresponding configuration.

In FIG. 14B, the value node after the input node may be configured to perform the task "parameter reset". For example, the first user can reset the parameter to the destination based on the dialogue content of the second user, "I remembered it wrong, I will not go to Hangzhou, but go to Shanghai instead." In addition, according to the conversation content of the second user, it can be determined by selecting which of the intended customers, potential users and invalid users the second user belongs to in the "customer type". Since the user's classification function can be implemented by configuring the assignment node, more refined services can be provided to the user.

In Figure 14B, the decision node may be configured to perform the task "destination city decision". The second user may configure the parameter logic to include parameter logic 1 and parameter logic 2, where parameter logic 1 is configured to indicate a "destination is empty" condition or a "destination does not belong to the destination dictionary" condition. Parameter logic 2 is configured to indicate the "destination belongs to the destination dictionary" case. Here, for example, the destination dictionary may record information on the destination.

14A and 14C, the intent unit may be configured to include two output nodes (ie, an output node corresponding to parameter logic 1 and an output node corresponding to parameter logic 2).

In Figure 14C, the output node corresponding to parameter logic 1 may be configured to perform the task "parameter logic 1 output". The output node corresponding to parameter logic 1 may be configured to output "Please confirm your destination and tell me again" to the second user. For example, the reply content may be output in one or more of the output forms shown in Figure 14C. Additionally, the output node corresponding to parameter logic 2 may be configured to perform the task "parameter logic 2 output". The output node corresponding to parameter logic 2 may be configured to output "Which airport in Shanghai are you going to?" to the second user.

With reference to the example of FIG. 9 , when the dialogue content of the fifth-round dialogue of the second user is “I remember wrong, I will not go to Hangzhou, but go to Shanghai instead”, the dialogue robot program can be based on the intent units configured in FIGS. 14A to 14C . Output the content "Which airport in Hangzhou are you going to?" to the first user.

Therefore, even if the intention of the second user changes, the dialogue robot program can ensure that the human-machine dialogue can continue flexibly and efficiently.

Fig. 14D shows the configuration of the sixth-round dialogue-based input nodes and judgment nodes of the construction intent unit shown in 14A; Fig. 14E illustrates the sixth-round dialogue-based AND parameter logic 1 of the construction intent unit shown in 14A The configuration of the corresponding output node and the configuration of the output node corresponding to parameter logic 2.

The dialogue content of the second user in the sixth round of dialogue in the example of the multi-round dialogue between humans and machines shown in FIG. 9 is "Shanghai Tianhe Airport". The first user may further construct the intent unit shown in FIG. 14A based on the dialogue content "Shanghai Tianhe Airport" of the second user. Specifically, since the second user's intent has not changed, processing proceeds to an input node parallel to the input node connected to the reset intent in FIG. 14B .

Referring to Figure 14D, the input node is configured to perform the task "Destination Airport". The second user can configure the configuration item "Trigger Question Method" of the input node to include "I want to go to Shanghai Tianhe Airport" and "Shanghai Tianhe", and configure the configuration item "Similar Question Method" of the input node to include "Shanghai Tianhe Airport" Airport". Here, the second user can also configure the slot "Shanghai Airport". For example, according to the dialogue content of the second user, the information of the slot may be Shanghai Tianhe Airport. In the case where the dialogue content of the second user is the content related to the flight booking, the content related to the departure city, the content related to the departure airport, or the content related to the destination, the jump query method is triggered, and the jump to The previous input node corresponding to the flight booking, the input node related to the departure city, the input node related to the departure airport, or the input node related to the destination. After configuring the input node, the first user can click OK to keep the corresponding configuration.

In Figure 14D, the decision node following the input node may be configured to perform the task "destination airport decision". The second user may configure the parameter logic to include parameter logic 1 and parameter logic 2, wherein parameter logic 1 is configured to indicate the case of "Shanghai Airport is empty" or the case of "Shanghai Airport does not belong to the Shanghai Airport Dictionary". Parameter logic 2 is configured to indicate the case of "Shanghai Airport belongs to the Shanghai Airport Dictionary". Here, for example, the Shanghai Airport Dictionary may record information on Shanghai Airport.

14A and 14E, the intent unit may be configured to include two output nodes (ie, an output node corresponding to parameter logic 1 and an output node corresponding to parameter logic 2) following the judgment node.

In Figure 14E, the output node corresponding to parameter logic 1 may be configured to perform the task "parameter logic 1 output". The output node corresponding to parameter logic 1 may be configured to output "Sorry, Shanghai Tianhe was not found, please confirm your destination airport" to the second user. For example, the reply content may be output using one or more of the output forms shown in Figure 14E. Additionally, the output node corresponding to parameter logic 2 may be configured to perform the task "parameter logic 2 output". The output node corresponding to parameter logic 2 may be configured to output "What is your departure time?" to the second user.

With reference to the example of FIG. 9 , when the dialogue content of the sixth-round dialogue of the second user is “Shanghai Tianhe Airport”, the dialogue robot program can be based on the output corresponding to parameter logic 1 of the intent unit configured in FIGS. 14A , 14D and 14E The node outputs the content "Sorry, Shanghai Tianhe was not found, please confirm your destination airport" to the second user.

In addition, in conjunction with the example of FIG. 9 , when the dialogue content of the seventh round of dialogue of the second user is “Shanghai Hongqiao Airport”, the dialogue robot program may correspond to parameter logic 2 based on the intent units configured in FIGS. 14A , 14D and 14E The output node outputs the content "What is your departure time?" to the second user.

Fig. 15A shows a further constructed intent unit for the eighth-round dialogue of the example of the man-machine multi-round dialogue shown in Fig. 9; Fig. 15B illustrates the configuration and judgment node of the input node with the construct intent unit shown in 15A 15C shows the configuration of the output node corresponding to the parameter logic 1 and the configuration of the output node corresponding to the parameter logic 2 and the assignment node of the build intent unit shown in 15A.

The dialogue content of the second user in the eighth dialogue round of the example of the man-machine multi-round dialogue shown in FIG. 9 is "tomorrow at 10 am". The first user may construct the intent unit shown in FIG. 15A based on the second user's dialogue content "tomorrow at 10 am". In particular, the portion of the intent unit newly constructed in FIG. 15A compared to FIG. 14A may correspond to the portion in the block in FIG. 15A .

Referring to Figure 15B, the input node is configured to perform the task "Departure Time". The second user may configure the configuration item "trigger question method" of the input node to include "let's start at 10 am tomorrow". Here, the second user can also configure the "departure time" of the slot. For example, according to the conversation content of the second user, the information of the slot may be 10 am tomorrow. In the case where the dialogue content of the second user is the content related to the flight booking, the content related to the departure city, the content related to the departure airport, or the content related to the destination, the jump query method is triggered, and the jump to The previous input node corresponding to the flight booking, the input node related to the departure city, the input node related to the departure airport, or the input node related to the destination. After configuring the input node, the first user can click OK to keep the corresponding configuration.

In Figure 15B, the decision node may be configured to perform the task "Departure Time Decision". The second user may configure the parameter logic to include parameter logic 1 and parameter logic 2, wherein parameter logic 1 is configured to indicate a "departure time is empty" condition or a "departure time does not belong to the time dictionary" condition. Parameter logic 2 is configured to indicate the "departure time belongs to the time dictionary" case. Here, for example, a time dictionary may record information on time.

15A and 15C, the intent unit may be configured to include two output nodes (ie, an output node corresponding to parameter logic 1 and an output node corresponding to parameter logic 2) and an interface node.

In Figure 15C, the output node corresponding to parameter logic 1 may be configured to perform the task "parameter logic 1 output". The output node corresponding to the parameter logic 1 may be configured to output "Sorry your trigger time cannot be determined, please confirm your departure time" to the second user. For example, the reply content may be output in one or more of the output forms shown in Figure 15C. Additionally, the output node corresponding to parameter logic 2 may be configured to perform the task "parameter logic 2 output". The output node corresponding to parameter logic 2 can be configured to output to the second user "find the following flight from Capital International Airport to Shanghai Hongqiao Airport at 10 am tomorrow, which one do you choose" and the interface return value returned from the interface node "China Southern Airlines AH123456". The interface node has been specifically described above with reference to FIG. 7 , and the description will not be repeated here.

With reference to the example of FIG. 9 , when the dialogue content of the eighth-round dialogue of the second user is "tomorrow at 10 am", the dialogue robot program can output the content "find tomorrow morning" to the first user based on the intent unit configured in FIG. 15A to FIG. 15C At 10:00 am, the following flights from Capital International Airport to Shanghai Hongqiao Airport, which one do you choose?” and the return value of the interface returned from the interface node is “China Southern Airlines AH123456”.

Fig. 16A shows a further constructed intent unit for the ninth round of the example of the man-machine multi-round dialog shown in Fig. 9; Fig. 16B shows the configuration and judgment node of the input node with the construct intent unit shown in 16A 16C shows the configuration of the output node corresponding to the parameter logic 1 and the configuration of the output node corresponding to the parameter logic 2 and the assignment node of the build intent unit shown in 16A.

The dialogue content of the second user in the ninth round of dialogue of the man-machine multi-round dialogue example shown in FIG. 9 is "China Southern Airlines AH123456". The first user can construct the intent unit shown in FIG. 16A based on the dialogue content of the second user "China Southern Airlines AH123456 Bar". In particular, the portion of the newly constructed intent unit in FIG. 16A compared to FIG. 15A may correspond to the portion in the block in FIG. 16A .

Referring to Figure 16B, the input node is configured to perform the task "flight". The second user can configure the configuration item "trigger question method" of the input node to include "China Southern Airlines AH123456". Here, the second user may also configure the slot "flight". For example, according to the conversation content of the second user, the information of the slot may be China Southern Airlines AH123456. In the case where the dialogue content of the second user is the content related to the flight booking, the content related to the departure city, the content related to the departure airport, or the content related to the destination, the jump query method is triggered, and the jump to The previous input node corresponding to the flight booking, the input node related to the departure city, the input node related to the departure airport, or the input node related to the destination. After configuring the input node, the first user can click OK to keep the corresponding configuration.

In Figure 16B, the decision node may be configured to perform the task "flight decision". The second user may configure the parameter logic to include parameter logic 1 and parameter logic 2, wherein parameter logic 1 is configured to indicate a "flight is empty" condition or a "flight is not part of the interface return value" condition. Parameter logic 2 is configured to indicate the "flight belongs to interface return value" case. Here, for example, the interface return value may record information about the flight.

16A and 16C, an intent cell may be configured to include two output nodes (ie, an output node corresponding to parameter logic 1 and an output node corresponding to parameter logic 2) and an assignment node.

In Figure 16C, the output node corresponding to parameter logic 1 may be configured to perform the task "parameter logic 1 output". The output node corresponding to the parameter logic 1 may be configured to output "Sorry, please select among optional flights, reconfirm your flight" to the second user. For example, the reply content may be output using one or more of the output forms shown in Figure 16C. Additionally, the output node corresponding to parameter logic 2 may be configured to perform the task "parameter logic 2 output". The output node corresponding to parameter logic 2 can be configured to output to the second user "Ok, China Southern Airlines AH123456 from Capital International Airport to Shanghai Hongqiao Airport has been booked for you on July 14, 2020, at 10:32 a.m." .

With reference to the example of FIG. 9 , when the dialogue content of the ninth round of dialogue of the second user is "China Southern Airlines AH123456", the dialogue robot program can output the content "Okay, it has been We book China Southern Airlines AH123456 from Capital International Airport to Shanghai Hongqiao Airport on July 14, 2020, at 10:32 am.”

Additionally, the first user may configure the assignment node to perform the task "Airline Intent Marking". Assignment nodes can include parameter resets and client flags. The second user can configure the parameter reset as the destination, and configure the user's customer mark as the intended user when the second customer selects China Southern Airlines AH123456.

9 to 16C are only examples shown for illustrative purposes, and the content of the human-machine dialogue of the present disclosure is not limited to the specific content in the above examples. Referring to FIG. 9 to FIG. 16C , the first user can flexibly select, configure, and connect intent nodes on the user configuration interface according to his own business needs and business experience, so as to construct a second user-oriented user interface for the dialog robot. Intent unit.

The method and apparatus for implementing a human-machine multi-turn dialogue according to an exemplary embodiment of the present disclosure have been described above with reference to FIGS. 1 to 16C . However, it should be understood that the methods used in FIGS. 1 to 7 and FIGS. 9 to 16C can be implemented by software, hardware, firmware or any combination of the above items that perform specific functions, and the methods and apparatuses used in FIG. 8 , systems, units, etc., may be configured as software, hardware, firmware, or any combination of the foregoing, respectively, to perform the specified functions. For example, these systems, devices or units, etc. may correspond to dedicated integrated circuits, may also correspond to pure software codes, or may correspond to units in which software and hardware are combined. In addition, one or more functions implemented by these systems, apparatuses or units, etc., may also be uniformly performed by components in a physical entity device (eg, a processor, a client or a server, etc.).

Furthermore, the above-described method can be implemented by a computer program recorded on a computer-readable storage medium. For example, according to exemplary embodiments of the present disclosure, a computer-readable storage medium may be provided having stored thereon a computer program that, when executed by one or more computing devices, causes all The one or more computing devices implement any of the methods disclosed in this application.

For example, the computer program, when executed by one or more computing devices, causes the one or more computing devices to perform the steps of: providing a user configuration interface in which multiple types of intent nodes are provided ; Build an intent unit based on the first user's operations of selecting an intent node, configuring an intent node, and connecting an intent node on the user configuration interface; based on the intent unit, implement multiple rounds of human-machine dialogue with the second user to obtain the first Second, the key intentions of users in multi-round human-machine dialogues to realize the judgment and prediction of dialogues.

The computer program in the above-mentioned computer-readable storage medium can be executed in an environment deployed in computer equipment such as a client, a host, an agent device, a server, etc. It should be noted that the computer program, when executed, can also be used to perform steps other than those described above. Additional steps other than those described above are performed or more specific processing is performed when the above steps are performed. The contents of these additional steps and further processing have been mentioned in the description of the related methods and apparatuses with reference to FIG. 1 to FIG. 8 , so to avoid repetitions here No further description will be given.

It should be noted that the method and apparatus for implementing multi-turn human-machine dialogue according to the exemplary embodiments of the present disclosure can completely rely on the running of a computer program to implement corresponding functions, wherein each unit of the apparatus or system is in the functional architecture of the computer program Corresponding to each step, the entire device or system is invoked through a special software package (eg, lib library) to implement corresponding functions.

For example, according to an embodiment of the present disclosure, there is provided a system for implementing a multi-turn human-machine dialogue including one or more computing devices and one or more storage devices, wherein the one or more storage devices store a computer program, The computer program, when executed by the one or more computing devices, causes the one or more computing devices to implement any of the methods disclosed in this application. For example, the one or more computing devices are caused to perform the steps of: providing a user configuration interface in which multiple types of intent nodes are provided; selecting an intent on the user configuration interface based on a first user Nodes, configuration intent nodes, and operations of connecting intent nodes to construct an intent unit; based on the intent unit, a man-machine multi-round dialogue with the second user is implemented to obtain the second user's key intent in the man-machine multi-round dialogue, and the realization of Dialogue judgment prediction.

Specifically, the above computing device may be deployed in a server, or may be deployed on a node device in a distributed network environment. In addition, the computing device apparatus may also include a video display (such as a liquid crystal display) and a user interaction interface (such as a keyboard, mouse, touch input device, etc.). All components of the computing device apparatus may be connected to each other via a bus and/or network.

Here, the computing device does not have to be a single device, but can also be any device or a collection of circuits capable of executing the above-mentioned instructions (or instruction sets) individually or jointly. The computing device may also be part of an integrated control computing device or computing device manager, or may be configured as a portable electronic device that interfaces locally or remotely (eg, via wireless transmission).

The computing device for performing the neural network training method or the named entity recognition method according to the exemplary embodiment of the present disclosure may be a processor, and such a processor may include a central processing unit (CPU), a graphics processing unit (GPU), a Programmable logic device, special purpose processor, microcontroller or microprocessor. By way of example and not limitation, the processors may also include analog processors, digital processors, microprocessors, multi-core processors, processor arrays, network processors, and the like. The processor may execute instructions or code stored in one of the storage devices, which may also store data. Instructions and data may also be sent and received over a network via a network interface device, which may employ any known transport protocol.

The storage device may be integrated with the processor, eg, RAM or flash memory arranged within an integrated circuit microprocessor or the like. Additionally, the storage device may include a stand-alone device such as an external disk drive, a storage array, or any other storage device that may be used by a database computing device. The storage device and the processor may be operatively coupled, or may communicate with each other, eg, through I/O ports, network connections, etc., to enable the processor to read files stored in the storage device.

It should be noted that the exemplary implementation of the present disclosure focuses on solving the complex dialogue situation facing the user at present, and the dialogue robot program is often difficult to deal with flexibly and cannot effectively solve the problem corresponding to the user's intention. Specifically, the present disclosure can flexibly select intent nodes, configure intent nodes, and connect intent nodes on the user configuration interface according to the first user's own business needs and business experience, so as to build a first-oriented first-level user interface for conversational robot programs. Two user intent units. Therefore, the dialogue robot program can flexibly respond to the user's complex dialogue situation based on the intent unit oriented to the second user, and effectively obtain the user's final instruction, thereby providing accurate information and services.

Various exemplary embodiments of the present application have been described above, and it should be understood that the above description is only exemplary and not exhaustive, and the present application is not limited to the disclosed exemplary embodiments. Numerous modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of this application. Therefore, the protection scope of the present application should be subject to the scope of the claims.

Industrial Applicability

The system, method and device for realizing human-machine multi-round dialogue provided by the present disclosure can flexibly select intent nodes, configure intent nodes, and connect intent nodes on the user configuration interface according to the first user's own business needs and business experience, thereby Built as a second user-facing intent unit for use by conversational bots. Therefore, the dialog robot program can flexibly respond to the user's complex dialog situation based on the intent unit oriented to the second user, and effectively acquire the user's key intent, thereby providing accurate information and services.

Claims (24)

1. A system for implementing multi-turn human-machine dialogue comprising at least one computing device and at least one storage device, the at least one storage device having instructions stored thereon, the instructions, when executed by the at least one computing device, cause the At least one computing device performs the following steps of a method for implementing a multi-turn human-machine dialogue:

   providing a user configuration interface, wherein multiple types of intent nodes are provided in the user configuration interface;

   constructing an intent unit based on the operations of the first user selecting an intent node, configuring an intent node, and connecting an intent node on the user configuration interface;

   Based on the intention unit, a multi-round dialogue with the second user is implemented, so as to obtain the key intention of the second user in the multi-round dialogue between the human and the computer, so as to realize the judgment and prediction of the dialogue.
2. The system of claim 1, wherein the multiple types of intent nodes include one or more of the following: input nodes, judgment nodes, output nodes, assignment nodes, and interface nodes.
3. The system of claim 2, wherein the configuration items of the input node include:

   Triggering methods for triggering tasks and similar methods to triggering questions;

   Slots, wherein the slots are set based on the trigger method and the trigger method similar to the trigger method;

   Task parameters, wherein the task parameters are set based on at least one of a trigger method and a method similar to the trigger method.
4. The system of claim 3, wherein the configuration item of the input node further includes a jumping method for jumping to other input nodes.
5. The system according to claim 2, wherein the configuration item of the judgment node includes parameter logic, wherein the parameter logic performs logical condition judgment on the value obtained by the previous node and the known value or the user-defined value, and judges based on mutual exclusion As a result, the next node of the judgment node is determined.
6. The system according to claim 5, wherein the configuration items of the parameter logic include a self-defined judgment value, a self-defined logical relationship and a self-defined reference value,

   The parameter logic is configured to determine whether the user-defined judgment value and the user-defined reference value satisfy the user-defined logical relationship, and to determine the next node of the judgment node based on the determined result.
7. The system according to claim 6, wherein the self-defined judgment value includes at least one of the task parameter value in the data received from the previous node, the parameter value in the parameter dictionary and the interface return value,

   Among them, the custom reference value includes at least one of the set value in the parameter dictionary, the interface return value, and the custom value,

   The custom logical relationship is a logical relationship between the custom judgment value selected by the first user according to the intent node and the custom reference value.
8. The system of claim 2, wherein the configuration item of the output node includes a configuration item of outputting the reply content to the second user in a predetermined form.
9. The system according to claim 8, wherein the text box in the default state of the content editor of the output node supports adding the obtained task parameters and the output parameters of the interface node.
10. The system according to claim 2, wherein the configuration item of the assignment node includes at least one of parameter reset and self-defined assignment,

    Wherein, the parameter reset includes the change of slot information based on the data received from the previous node of the assignment node, and the custom assignment is used to support the user tag function.
11. The system of claim 2, wherein the configuration item of the interface node includes a configuration item configured to support interface information retrieval configuration.
12. A method for realizing multi-round dialogue between humans and machines, the method comprising:

    providing a user configuration interface, wherein multiple types of intent nodes are provided in the user configuration interface;

    constructing an intent unit based on the operations of the first user selecting an intent node, configuring an intent node, and connecting an intent node on the user configuration interface;

    Based on the intention unit, a multi-round dialogue with the second user is implemented, so as to obtain the key intention of the second user in the multi-round dialogue between the human and the computer, so as to realize the judgment and prediction of the dialogue.
13. 13. The method of claim 12, wherein the multiple types of intent nodes include one or more of the following: input nodes, judgment nodes, output nodes, assignment nodes, and interface nodes.
14. The method of claim 13, wherein the configuration items of the input node comprise:

    Triggering methods for triggering tasks and similar methods to triggering questions;

    Slots, wherein the slots are set based on the trigger method and the trigger method similar to the trigger method;

    Task parameters, wherein the task parameters are set based on at least one of a trigger method and a method similar to the trigger method.
15. The method of claim 14, wherein the configuration item of the input node further comprises a jumping method for jumping to other input nodes.
16. The method according to claim 13, wherein the configuration item of the judgment node includes parameter logic, wherein the parameter logic performs logical condition judgment on the value obtained by the previous node and the known value or the user-defined value, and judges based on mutual exclusion As a result, the next node of the judgment node is determined.
17. The method according to claim 16, wherein the configuration items of the parameter logic include a self-defined judgment value, a self-defined logical relationship and a self-defined reference value,

    The parameter logic is configured to determine whether the user-defined judgment value and the user-defined reference value satisfy the user-defined logical relationship, and to determine the next node of the judgment node based on the determined result.
18. The method according to claim 17, wherein the self-defined judgment value includes at least one of a task parameter value in the data received from the previous node, a parameter value in a parameter dictionary and an interface return value,

    Among them, the custom reference value includes at least one of the set value in the parameter dictionary, the interface return value, and the custom value,

    The custom logical relationship is a logical relationship between the custom judgment value selected by the first user according to the intent node and the custom reference value.
19. The method of claim 13, wherein the configuration item of the output node includes a configuration item of outputting the reply content to the second user in a predetermined form.
20. The method according to claim 19, wherein the text box in the default state of the content editor of the output node supports adding the obtained task parameters and the output parameters of the interface node.
21. The method of claim 13, wherein the configuration item of the assignment node includes at least one of parameter reset and self-defined assignment,

    Wherein, the parameter reset includes the change of slot information based on the data received from the previous node of the assignment node, and the custom assignment is used to support the user tag function.
22. 14. The method of claim 13, wherein the configuration item of the interface node includes a configuration item configured to support interface information retrieval configuration.
23. A device for realizing man-machine multi-round dialogue, the device comprising:

    a user configuration interface providing unit that provides a user configuration interface, wherein the user configuration interface providing unit provides multiple types of intent nodes in the user configuration interface;

    an intent unit construction unit, which constructs an intent unit based on the operations of the first user selecting an intent node, configuring an intent node, and connecting an intent node on the user configuration interface;

    The human-machine multi-round dialogue realization unit realizes the human-machine multi-round dialogue with the second user based on the intention unit, so as to obtain the key intention of the second user in the human-machine multi-round dialogue, and realize the judgment and prediction of the dialogue.
24. A computer-readable storage medium having stored thereon a computer program that, when executed by one or more computing apparatuses, causes the one or more computing apparatuses to implement the methods of claim 12- The method of any one of 22.

Images