KR20190064313A - Method for providing personalized chatbot service and apparatus thereof - Google Patents

Abstract

Provided are a method for determining a preference for a topic for a user and a method for providing a personalized chatbot service based on the determined preference. According to an embodiment of the present invention, the method for determining a preference by each topic performed by a computing device comprises the steps of: classifying utterances constituting a dialog between a first user and a chatbot by each topic; and determining a preference for a first topic of the first user using the total amount of utterances pertaining to the first topic.

Description

[0001] METHOD FOR PROVIDING PERSONALIZED CHATBOT SERVICE AND APPARATUS THEREOF [0002]

The present invention relates to a personalized chatbot service providing method and apparatus. More particularly, the present invention relates to a method of providing a personalized chat service on the basis of a topic preference of a user of a chat service and a device performing the method.

A chatbot is an artificial intelligence (AI) -based communication software that provides answers to questions and various related information through dialogue with people. Chattob is evaluated as an area with excellent growth potential because it has the convenience of updating the service, which is the advantage of the web, and accessibility and mobility, which are advantages of the app.

Recently, artificial intelligence technology has been developed and natural language processing capability has been added, so that application services combined with chatbots are emerging in various fields. For example, chatbots have been used in various fields such as providing an automatic response service by replacing a contact center agent with a chatbot, or providing a product or content recommendation service based on a chatbot.

Most of the chaotic services proposed so far are able to identify the user's intent or topic from the utterance of the user and select a predefined dialog flow or a predefined conversation pair Provide appropriate conversation services. Accordingly, only a uniform chat service is provided regardless of the characteristics of the user. For example, regardless of the characteristics of the user, the dialogue service is provided according to the same flow of conversation on the same subject, and the same response is provided for the same question.

Accordingly, there is a need for a method of providing a personalized chatbot service reflecting the characteristics of a user.

Korean Patent Publication No. 10-2013-0124799 (published Nov. 15, 2013)

SUMMARY OF THE INVENTION It is an object of the present invention to provide a personalized chatbot service providing method and a device for performing the personalized chatbot service reflecting user's preference by theme.

It is another object of the present invention to provide a method for determining a preference degree for a user for a personalized chatbot service and an apparatus for performing the method.

The technical objects of the present invention are not limited to the above-mentioned technical problems, and other technical subjects not mentioned can be clearly understood by those skilled in the art from the following description.

According to an aspect of the present invention, there is provided a method for determining a topic preference according to an embodiment of the present invention, the method comprising the steps of: generating a utterance ) Of the first subject and the amount of total utterance belonging to the first subject may be used to determine the preference of the first user for the first subject.

In one embodiment, the step of determining preferences for the first subject comprises extracting first feature data for a speech of the first user during a whole speech belonging to the first subject, Extracting the second characteristic data with respect to the ignition of the chatbot during the entire utterance, and calculating the quantity using the first characteristic data and the second characteristic data. At this time, the first feature data and the second feature data may be selected by different criteria.

According to another aspect of the present invention, there is provided a method for determining a topic preference according to another embodiment of the present invention, the method comprising the steps of: And determining a preference of the first user for the first subject using the number of times the predetermined speech-response relationship is extracted in the conversation belonging to the first topic.

According to another aspect of the present invention, there is provided a method of providing personalized chatbot services, the method comprising the steps of: obtaining a preference for a theme of a first user; Without the topic presentation, the chatbot may generate the topic presentation using the subject-specific preferences of the first user.

According to another aspect of the present invention, there is provided a method of providing a personalized chatbot service, the method comprising the steps of: retrieving a user model of a first user, The method comprising the steps of: including a speech-response relationship, adjusting a standard dialog flow using a spoken-response relationship of the searched user model, And providing a chat service.

According to an aspect of the present invention, there is provided a server for providing a service to a chattcom, including a processor, a network interface, a memory for loading a computer program executed by the processor, Wherein the computer program comprises instructions for dividing each utterance constituting a dialog between a first user and a chatbot by topic and an amount of total utterance belonging to the first topic, And may include instructions for determining a preference of the first user for the first subject.

According to another aspect of the present invention, there is provided a server for providing a chat service, including a processor, a network interface, a memory for loading a computer program executed by the processor, and a storage for storing the computer program. Wherein the computer program further comprises instructions for distinguishing each utterance constituting a conversation between the first user and the chatbot by a subject and an instruction for identifying the utterance of the chatbot using the number of times the predetermined utterance- And may include instructions for determining a preference of the first user for the first subject.

According to another aspect of the present invention, there is provided a server for providing a chat service including a processor, a network interface, a memory for loading a computer program executed by the processor, Wherein the computer program includes instructions for obtaining a subject matter preference of a first user and instructions for generating a topic presentation utterance using the subject matter preference of the first user without the subject presenting presentation of the first user .

According to another aspect of the present invention, there is provided a server for providing a chat service including a processor, a network interface, a memory for loading a computer program executed by the processor, Wherein the computer program comprises: querying a user model of a first user, the user model comprising instructions comprising one or more spoken-response relationships, a standard dialogue using the spoken- Instructions for adjusting the dialog flow, and instructions for providing a chat service to the first user based on the adjusted standard chat flow.

According to an aspect of the present invention, there is provided a computer program for causing a computer to function as a topic, comprising a utterance of utterances constituting a dialog between a first user and a chatbot, And to determine the preference of the first user for the first subject using the amount of total speech pertaining to the first subject. ≪ RTI ID = 0.0 > [0002] < / RTI >

According to another aspect of the present invention, there is provided a computer program product comprising a computer readable medium storing a computer program for causing a computer to function as: Readable recording medium so as to execute the step of determining the preference of the first user for the first subject by using the number of times the predetermined speech-response relationship is extracted in the conversation to which it belongs.

According to another aspect of the present invention, there is provided a computer program product, which is combined with a computing device to obtain a preference for a subject of a first user, And may be stored in a computer-readable recording medium so as to execute the step of generating a topic presentation utterance using the first user's subject preference.

According to another aspect of the present invention, there is provided a computer program product, coupled to a computing device, for querying a user model of a first user, the user model including one or more spoken- Coordinating a standard dialog flow using the spoken-response relationship of the searched user model, and providing a chattob service to the first user based on the coordinated standard conversation flow Readable < / RTI > recordable medium.

FIG. 1 is a block diagram of an exemplary chatbot service providing system to which the present invention can be applied.
FIGS. 2 to 4 are views for explaining a method of performing a method for determining a subject preference according to a first embodiment of the present invention and a configuration and operation of a chattb service server for performing the method.
5 to 8 are views for explaining a method and a method for determining a preference degree according to a second embodiment of the present invention and a configuration and operation of a chattb service server for performing the method.
FIG. 9 and FIG. 10 are views for explaining a method and a method for determining a topic preference according to a third embodiment of the present invention, and a configuration and operation of a chat service server performing the method.
11 is a diagram illustrating an exemplary method for providing a personalized chatbot service according to the first embodiment of the present invention.
12 and 13 are diagrams for explaining a personalized chatbot service providing method according to a second embodiment of the present invention.
FIG. 14 is a flowchart illustrating a method for determining a theme preference according to the first embodiment of the present invention.
FIG. 15 is a flowchart illustrating a method for determining a theme preference according to a second embodiment of the present invention.
16 is a flowchart illustrating a method of providing a personalized chatbot service according to the first embodiment of the present invention.
17 is a flowchart illustrating a method for providing a personalized chatbot service according to a second embodiment of the present invention.
FIG. 18 is a hardware configuration diagram of a chattop service server according to an embodiment of the present invention.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS The advantages and features of the present invention and the manner of achieving them will become apparent with reference to the embodiments described in detail below with reference to the accompanying drawings. The present invention may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art. To fully disclose the scope of the invention to those skilled in the art, and the invention is only defined by the scope of the claims. Like reference numerals refer to like elements throughout the specification.

Unless defined otherwise, all terms (including technical and scientific terms) used herein may be used in a sense commonly understood by one of ordinary skill in the art to which this invention belongs. Also, commonly used predefined terms are not ideally or excessively interpreted unless explicitly defined otherwise. The terminology used herein is for the purpose of illustrating embodiments and is not intended to be limiting of the present invention. In the present specification, the singular form includes plural forms unless otherwise specified in the specification.

It is noted that the terms "comprises" and / or "comprising" used in the specification are intended to be inclusive in a manner similar to the components, steps, operations, and / Or additions.

Prior to the description of the present specification, some terms used in this specification will be clarified.

In the present specification, utterance refers to an element constituting a dialog between a user and a chatbot. When a conversation between a user and a chatbot is composed of a user-side message and an action (or a response) provided by the chatbot in response to the message, the speech may be a concept including both the message and the action.

In this specification, a dialog act or speech act means a user's general utterance intention inherent in utterance. For example, the type of the speech may include, but is not limited to, a request for requesting processing of a behavior, a notification for providing information, a query for requesting information, and the like. It can be varied.

In this specification, a topic refers to a user's specific utterance intention. That is, the subject differs from the above-mentioned speech in that it is a specific utterance object that the user wants to achieve through utterance. The utterance may be used interchangeably with terms such as intent, topic, main act, and the like in the art. According to an embodiment, one theme may correspond to one user intention, and a plurality of user intentions may be included in one theme.

In the present specification, a chatbot refers to an intelligent agent that provides a conversation service without user intervention. The chatbot can be used in the art in combination with terms such as an intelligent dialog agent, an artificial assistant, a virtual assistant, and the like.

In this specification, an instruction is a set of instructions, which are grouped by function, indicating that they are executed by a processor and a component of a computer program.

Hereinafter, some embodiments of the present invention will be described in detail with reference to the accompanying drawings.

FIG. 1 is a block diagram of an exemplary chatbot service providing system to which the present invention can be applied.

Referring to FIG. 1, the chat service providing system may be configured to include user terminals 10a, 10b, and 10c and a chat service server 100. However, it should be understood that the present invention is not limited to the above-described embodiments, and that various changes and modifications may be made without departing from the scope of the present invention. It should be noted that each component of the chattbool service providing system shown in FIG. 1 shows functional elements that are functionally distinguished, and that at least one component may be implemented in a form that they are integrated with each other in an actual physical environment. Hereinafter, each component of the chattop service providing system will be described.

In the chattop service providing system, the user terminals 10a, 10b, and 10c are terminals used by a user to use a chat service. 1, the user terminal may be implemented as a laptop, a desktop, a laptop, a smart phone, a wearable device, and the like, but is not limited thereto, And any computing device with communication capabilities.

The user can be provided with a chat service provided by the chat service server 100 on the basis of an app or web installed in the user terminals 10a, 10b, and 10c.

In the chat service providing system, the chat service server 100 is a computing device that provides a chat service. The computing device may be a notebook computer, a desktop computer, a laptop computer, etc., but is not limited thereto and may be implemented in any computing device having a computing function and a communication function. However, in order to provide a smooth chat service to a large number of users, the chat service server 100 may be preferably implemented as a high performance server computing device.

According to the embodiment of the present invention, the chat service service server 100 can provide personalized chat service to each user. For example, the chatbot service server 100 may present a conversation topic not presented by a user, or may provide a conversation service specialized for a specific user by adjusting a predefined standard conversation flow. According to the present embodiment, since the conversation service specialized for each user is provided, the user's satisfaction with the chat service can be enhanced. A detailed description of this embodiment will be given later with reference to Figs. 11 to 13, Fig. 16, and Fig.

In addition, the chattop service server 100 can determine a topic preference for each user in a dialog log of the user. Subject preferences can be used to provide personalized chatbot services. A detailed description of this embodiment will be given later with reference to Figs. 2 to 10, Fig. 14, and Fig.

In the chattop service providing system, the chattcom service server 100 and the user terminals 10a, 10b, and 10c can communicate through a network. The network may be implemented in any type of wired / wireless network such as a local area network (LAN), a wide area network (WAN), a mobile radio communication network, a Wibro (Wireless Broadband Internet) .

Up to now, an exemplary chat service providing system to which the present invention can be applied has been described with reference to FIG. Next, the configuration and operation of the chattcom service server 100 will be described with reference to FIG. 2 to FIG.

First, with reference to FIG. 2 to FIG. 4, a method of determining a theme preference according to a first embodiment of the present invention and a configuration and operation of a chat service server 100 for performing the method will be described.

In the first embodiment, the chattob service server 100 determines the theme preference for each user based on the volume of speech. For example, suppose that the first user's conversation log includes five topics (T1, T2, T3, T4, T5), as shown in FIG. In this case, the chat service service server 100 calculates the utterance amounts 20, 9, 17, 7 and 5 for the respective conversation subjects T1, T2, T3, T4 and T5, (T1> T3> T2> T4> T5) of the first user. The thematic preference may be determined to be a relative value or an absolute value according to the embodiment. 2, the original figure corresponds to the conversation subject, and the size of the original figure is shown to be proportional to the speaking amount.

In the first embodiment, the chattop service server 100 extracts predetermined feature data from the topic-specific utterance contents, and calculates the utterance amount for each topic based on the extracted feature data. More specifically, the first feature data is extracted from the utterance of the user during the entire utterance belonging to the first conversation subject, the second feature data is extracted from the utterance of the chatbot, and at least one of the first feature data and the second feature data The ignition amount is calculated based on one piece of data. At this time, the first feature data and the second feature data may be selected by different criteria. Hereinafter, the first characteristic data and the second characteristic data will be described.

In one embodiment, the first feature data may include a sum of the number of utterances of the question type, the number of utterances of the request type, and the number of utterances of the notification type. The type of the utterance can be determined through an analysis of the sentence. The above-described analysis is obvious to those skilled in the art, and a description thereof will be omitted.

In one embodiment, the first feature data is a sum of the number of entities included in the query type's utterance, the number of entity names included in the request type's utterance, and the number of entity names included in the utterance of the notification type . The entity name may be determined, for example, by a named entity recognition (NER) algorithm. The identification of the entity name is already obvious to those skilled in the art, and a description thereof will be omitted.

In one embodiment, the first feature data may comprise the number of emotion indicator words. The emotional display word may be extracted from the corresponding utterance or extracted through morphological analysis based on a predefined emotional dictionary, but may be extracted in any manner.

In one embodiment, the second feature data may include the number of entity names included in the chatbot's speech in which the user's answer was positive. Whether the user's answer is positive can be determined through sentiment analysis. The emotional analysis is obvious to those skilled in the art, and a description thereof will be omitted.

In one embodiment, the second feature data may include the number of entity names included in the notification type utterance.

The chatbot service server 100 may extract feature data from a speech between a user and a chatbot according to a combination of the above-described embodiments, and may calculate a volume of a topic by using the feature data.

In order to provide a more convenient understanding, a method for determining the subject preference according to the first embodiment will be described in detail with reference to the example shown in FIG.

3 shows an example of determining preferences for a second utterance 20 for a first utterance 10 and a second subject ("personal information change") for a first subject ("card recommendation") . Particularly, FIG. 3 shows an example of determining the preference by theme using both the first feature data determined in the user-side utterance and the second feature data determined in the chattobook utterance.

Referring to FIG. 3, the feature data extracted in the first speech 10 is displayed in the table 30a. In the feature data extraction process for the first speech 10, since the user wants to receive a recommendation for a card, a discount for a shopping card, and a general card for a request, a notification, and a notification type, U1) is extracted as "3 ". Similarly, the feature data U3 is extracted as "3" according to the entity name recognition result for the user side utterance, and the feature data C2 is extracted as "1 " When summed, the ignition quantity for the first ignition 10 can be calculated as "7 ".

The feature data extracted from the second utterance 20 is displayed in the table 40a. The feature data extraction method is the same as that of the first speech 10, and a description thereof will be omitted. When the extracted feature data is added, the amount of utterance for the second utterance 20 can be calculated as "3 ".

Accordingly, in the case of the first user, the preference ("7") of the first subject ("card recommendation") is determined to be higher than the preference ("3") of the second subject ("personal information change").

FIG. 4 shows a configuration of a chat service server 100 that performs a method for determining a subject preference according to the first embodiment.

4 is a block diagram showing a chat service service server 100 according to the first embodiment.

4, the chattop service server 100 includes a topic switching determination unit 104, a topic-based speech grouping unit 106, an analysis unit 108, an emotion analyzing unit 110, an entity name recognizing unit 112, A subject-specific amount determining unit 114, a subject-specific preference determining unit 116, and a chat bot engine 120. 4, only the components related to the embodiment of the present invention are shown. Accordingly, those skilled in the art will recognize that other general-purpose components may be included in addition to those shown in FIG. It should be noted that each component of the chat service service server shown in FIG. 4 shows functional elements that are functionally distinguished, and that at least one component may be integrated in a physical environment.

Referring to each component, the conversation switching determination unit 104 determines whether to switch conversions from conversation data for each user stored in the conversation log 102. [

The topic-based grouping unit 106 groups each utterance according to a theme according to the determination result of the subject-matter switching determination unit 104. [

The transposition analyzing unit 108 performs transcript analysis on each utterance. The type of each utterance (eg question, request, notification, etc.) can be determined through an analysis of the legend. Any leg analysis algorithm may be used for the above analysis.

The emotion analyzing unit 110 performs emotion analysis for each utterance. In the emotional analysis process, the emotional analysis unit 110 may extract emotional display words included in each utterance. Also, the emotional analysis unit 110 may determine emotional labels (e.g., affirmative, negative) for each utterance. Any emotion analysis algorithm may be used for the emotional analysis.

The entity name recognizing unit 112 recognizes the entity names included in each utterance. Any entity name recognition algorithm may be used for entity name recognition.

The subject-specific amount-of-speech calculating unit 114 calculates a speaking amount for each conversation topic using at least one result of the results of the speech analyzing unit 108, the emotion analyzing unit 110, and the entity name recognizing unit 112. Specifically, the subject-specific amount calculating unit 1140 extracts feature data associated with the utterance amount for each subject using the results of the transcription analyzing unit 108, the emotion analyzing unit 110, and the entity name recognizing unit 112, And calculates the amount of utterance based on the feature data. An example of the feature data is as described above, and a description thereof will be omitted.

The thematic preference determination unit 116 determines a preference for each user based on the subject's utterance amount. Since the amount of utterance reflects a user's interest in the subject, the higher the value of the utterance, the higher the preference for the subject.

The theme preference determination unit 116 may determine a theme preference for each user, and build a user model 118 based on the preference preference. The user model 118 is a model that includes the topic preference information for each user, and can be used by the chatbot engine 120 to provide a personalized chatbot service.

The chatbot engine 120 provides an overall dialog processing function.

The chatbot engine 120 may provide a user-specific chatbot service based on the user model 118. A detailed description thereof will be given later with reference to Fig.

Each component of the chattop service server 100 shown in FIG. 4 and FIG. 8 and FIG. 10 to be described later may be implemented by software or hardware such as an FPGA (Field Programmable Gate Array) or ASIC (Application-Specific Integrated Circuit) Hardware). However, the components are not limited to software or hardware, and may be configured to be addressable storage media, and configured to execute one or more processors. The functions provided in the components may be implemented by a more detailed component, or may be implemented by a single component that performs a specific function by combining a plurality of components.

2 to 4, a description has been given of a method for determining a topic preference according to the first embodiment of the present invention and a configuration and operation of the chattcom service server 100 for performing the method. According to the first embodiment, the preference according to the subject can be determined based on the volume of the user's utterance. According to the present embodiment, the subject matter preference can be accurately determined without human intervention, because the amount of utterance is a measure indicating the user's interest in a specific subject.

Hereinafter, a method and a method for determining a theme preference according to a second embodiment of the present invention will be described with reference to FIGS. 5 to 8. FIG.

The method for determining the subject preference according to the second embodiment determines the preference for the subject based on the utterance-response relationship. For example, suppose that the conversation log of the first user includes five conversation subjects T1, T2, T3, T4, and T5, as shown in Fig. In this case, the chattop service server 100 extracts a predetermined speech-response relationship from the utterances of the respective conversation subjects T1, T2, T3, T4, and T5 and extracts the extracted number (10, 4, 8, 2 , 1). Also, the chatbot service server 100 can determine the preference degree (T1> T3> T2> T4> T5) according to the conversation subject based on the calculated number.

An example of a predetermined speech-response relationship is shown in Fig. As in the utterance-response relationship table 50 shown in Fig. 6, the utterance-response relationship may be preset for each subject. In this case, it is preferable that the ignition-response relationship is determined as a pair of utterances closely related to the subject.

According to embodiments of the present invention, the utterance-response relationship table can be automatically constructed by the chatbot service server 100.

In one embodiment, the chatbot service server 100 may extract the topic-specific utterance-response relationship from the conversation log of the entire user and construct the extracted utterance-response relationship as a table as shown in FIG. At this time, the chatab service server 100 can extract the utterance-response relationship based on the frequency of the utterance-response relationship, the emotion of the user (e.g., positive emotion) expressed in the utterance-response relationship, and the like.

In one embodiment, the chatbot service server 100 may extract the subject-specific utterance-response relationship from the conversation log of a specific user and construct the extracted utterance-response relationship as a table as shown in FIG. At this time, the chatab service server 100 can extract the utterance-response relationship based on the frequency of the utterance-response relationship, the emotion of the user (e.g., positive emotion) expressed in the utterance-response relationship, and the like. According to the present embodiment, an utterance-response relation table is constructed for each user, and the chatbot service server 100 refers to the utterance-response relationship table of the user and extracts a predetermined utterance-response relationship from the utterance of the user have.

In order to provide more convenience of understanding, the method for determining the subject preference according to the second embodiment will be described in detail with reference to the example shown in FIG.

FIG. 7 shows an example of determining preferences by subject based on the number of utterance-response relationships extracted from the same utterance contents 10 and 20 as in FIG.

Referring to Fig. 7, the speech-response relationship detected in the first speech 10 is indicated in the table 30b. Shopping relationship is extracted "once" in the first utterance content 10 for the first subject ("card recommendation") and the card type- general relationship is " As extracted, the total number of utterance-response relationships extracted from the first utterance content 10 is calculated "twice ".

When the utterance-response relationship is extracted in a similar manner as described above, the total number of utterance-response relationships extracted from the second utterance content 20 for the second subject ("personal information change") is calculated as "zero".

Accordingly, the preference ("2") for the first theme ("card recommendation") can be determined to be a value higher than the preference ("0") for the second theme ("personal information change").

FIG. 8 shows a configuration of a chat service server 100 that performs a method for determining a subject preference according to the second embodiment. In order to avoid redundant description, description of the same or similar components as those of the chat service service server 100 shown in FIG. 4 among the components shown in FIG. 8 will be omitted.

FIG. 8 is a block diagram illustrating a chat service service server 100 according to a second embodiment of the present invention.

Referring to FIG. 8, the chattop service server 100 includes a topic switching determination unit 104, a topic-based speech grouping unit 106, an entity name recognition unit 112, a speech-response relationship extraction unit 122, Unit 124 and a chat bot engine 120. [

The utterance-response relation extraction unit 122 extracts the utterance-response relation using the entity name recognized in the dialog between the user and the chopsticks. For example, the utterance-response relation extraction unit 122 may extract the utterance-response relation from the topic utterance contents using the pre-set utterance-response relationship table and the entity name recognition result.

The subject-specific preference determining unit 124 determines the preference of the subject based on the number of the utterance-response relationships. The description is omitted since it is the same as described above. As in the first embodiment, the theme preference determined by the subject preference determination unit 124 is stored in the user model 118. [ The stored subject-specific preference information may be used by the chatbot engine 120 to provide a personalized chatbot service.

The chatbot engine 120 may provide a personalized chatbot service for the user by adjusting the standard chat flow using the predetermined user-specific chat-response relationship. This will be described later with reference to FIG. 12 and FIG.

5 to 8, a description has been given of a method for determining a topic preference according to the second embodiment of the present invention and a configuration and operation of the chattcom service server 100 for performing the method. According to the second embodiment, the theme preference is determined based on the detected number of the predetermined speech-response relationship. Since the utterance-response relationship will be set as an utterance-response pair closely related to the subject, the subject matter preference can be more accurately determined based on the relation between the subject and utterance. For example, according to the first embodiment, subject-related preferences may be determined to be inaccurate by contributing utterances not related to the subject to the utterance amount. However, according to the second embodiment, the above-described problems can be prevented in advance.

Hereinafter, with reference to FIG. 9 and FIG. 10, a description will be made of a method of determining a motif according to a third embodiment of the present invention and a configuration and operation of a chattcom service server 100 performing the method.

The subject matter determination method according to the third embodiment determines the subject matter preference using the number of utterances and the number of utterance-response relationships. An example of this is shown in Fig. The description of FIG. 9 will be made with reference to the explanations of FIGS. 3 and 7. FIG. For reference, FIG. 9 shows an example in which the final subject preference is determined by the sum of the first subject preference determined based on the utterance amount and the second subject preference determined based on the number of utterance-response relationships. However, The weighted sum or the like may be used to determine the preference according to the final theme.

FIG. 10 shows a configuration of a chat service server 100 that performs a method for determining a subject preference according to the third embodiment. The description of each component will be made with reference to the explanatory portions of FIGS. 4 and 8. FIG.

Up to now, a method for determining a topic preference according to some embodiments of the present invention and a configuration of a chat service server 100 for performing the method have been described with reference to FIG. 2 to FIG. Hereinafter, with reference to FIG. 11 to FIG. 13, a description will be given of a method by which the chat service server 100 provides the personalized chat service in detail.

11 is a diagram illustrating an exemplary method for providing a personalized chatbot service according to the first embodiment of the present invention.

The personalized chatbot service providing method according to the first embodiment provides the user with a topic presentation utterance using the calculated topic preference according to the above-described embodiments. An example of this is shown in Fig.

Referring to the conversation example 60 shown in FIG. 11, the chattop service server 100 identifies the first user in response to a chat request of the first user. Next, the chat service service server 100 acquires the preference of the first user from the user model 118 for the first user, and before the first user presents the topic on the basis of the acquired preference for the theme, It can provide a topic presentation. In particular, FIG. 11 illustrates by way of example the provision of a topic presentation on "nightly orders" via chatbots.

On the other hand, there may be a case where a user's negative reaction is detected with respect to the topic presentation utterance about the first subject provided by the chattob service server 100. [ The negative reaction can be detected, for example, through emotional analysis of the user's utterance corresponding to the topic presentation utterance.

In such a case, in response to the negative response, the chatbot service server 100 may be operable to provide a topic presentation speech for a second subject other than the first subject. Here, the second subject may be a subject having a lower preference than the first subject. Or the second subject may be a randomly selected topic. In accordance with an embodiment, the chattob service server 100 may generate and provide a user with a topic query query. For example, if a strong negative reaction of the user is detected, the chatbot service server 100 may stop serving further topics and act to query the user for a subject of interest.

In the above-described example, the chat service service server 100 may adjust the preference for the first topic to a lower value in response to the user's negative statement. Also, when a positive response to the first subject or the second subject is sensed, the chattob service server 100 may adjust the preferences for the first subject or the second subject to a higher value.

On the other hand, in the first embodiment, there may be a case where a clue about the subject of interest of the user is detected in the user utterance. For example, when a user speaks "hungry", the topic of food (eg, food recommendation, restaurant recommendation, etc.) is likely to be a topic of interest to the user. As another example, when a user utters "boredom", the topic of play (eg, movie recommendation inquiry) is likely to be a topic of interest to the user. In this case, the service server 100 selects a plurality of candidate topics by using the detected clues, and provides a topic presentation utterance about the selected topic using the preference of the user among the plurality of candidate topics can do.

The personalized chatbot service providing method according to the first embodiment of the present invention has been described with reference to FIG. According to the above-described method, an active and more intelligent chatbot service can be provided based on the user's preference according to the topic, and conversation about a topic of interest can be performed without presenting the topic of the user, The satisfaction and the convenience of the user can be improved.

Hereinafter, a personalized chat service providing method according to a second embodiment of the present invention will be described with reference to FIGS. 12 and 13. FIG.

The personalized chatbot service providing method according to the second embodiment provides a personalized chatbot service based on the speech-response relationship included in the model of the user. An example of this is shown in Figs. 12 and 13. Fig.

12 shows an example in which a chat service is provided according to a predefined standard conversation flow. Specifically, FIG. 12 illustrates a dialog example 70a in which a first user is provided with a chat service in accordance with a standard conversation flow 80a for "card recommendation ".

Referring to the conversation example 70a shown in FIG. 12, it can be seen that the conversation proceeds according to the defined flow based on the standard conversation flow 80a for "card recommendation ".

FIG. 13 shows an example in which conversation proceeds according to a standard conversation flow that has been adjusted for the same subject as in FIG.

Referring to the dialog example 70b shown in FIG. 13, the chattv service server 100 can adjust the standard conversation flow using the speech-response relationship 71a, 73a stored in the user model 118 of the first user have. Specifically, the standard conversation flow 80b is switched from a first intention ("card recommendation") to a second intention ("benefit") and from a second intention ("benefit") to a third intention ("shopping | Quot; coffee ") 71b, and during the speech-response relationship 71a, 73a of the user model 118, the utterance is the second intention (" benefit ") and the response is the third intention ), The chattop service server 100 adjusts so that the standard conversation flow is immediately switched 72 from the first intention ("card recommendation") to the third intention ("shopping") . Accordingly, as illustrated in the conversation example 70b, the chattob service server 100 can provide the first user with a speech "just want a shopping benefit".

Up to now, an example of a method for providing personalized chatbot service according to the second embodiment of the present invention has been described with reference to FIG. 12 and FIG. According to the above description, based on the utterance-response relationship defined for each user, a user's intention can be quickly grasped and a customized chatbot service can be provided. Accordingly, efficient conversation processing is possible, and the user's satisfaction with the chat service can be enhanced.

Hereinafter, a method for determining a subject preference and a method for providing a personalized chat service according to some embodiments of the present invention will be briefly described with reference to flowcharts shown in FIGS. 14 to 17 .

Each step of the method described below may be performed by a computing device. For example, the computing device may be a chatbot service server 100. For the sake of convenience of explanation, the description of the operation subject of each step included in the above method may be omitted. In addition, each step of the method described below may be implemented as an instruction of a computer program executed by a processor.

14 and 15, a method for determining a theme preference according to some embodiments of the present invention will be described.

FIG. 14 is a flowchart illustrating a method for determining a theme preference according to the first embodiment of the present invention. However, it should be understood that the present invention is not limited thereto and that some steps may be added or deleted as needed.

Referring to FIG. 14, in step S101, the chattob service server 100 groups the utterance contents included in the conversation log by topic.

In step S103, the chattop service server 100 calculates the volume related characteristic data from the subject utterance contents (S103). For the example of the feature data, the contents of the above description are referred to.

In step S105, the chattop service server 100 determines the preference by subject based on the calculated volume related characteristic data. For example, as described with reference to FIG. 3, the preference of a topic having a large amount of speech can be determined to be a high value.

Up to now, a method for determining a theme preference according to the first embodiment has been described with reference to FIG. According to the above, the topic preference is determined in consideration of the quantitative aspect of the utterance of the subject. The amount of utterance is a key measure that reflects the user's interest in the subject, so subject matter preferences can be accurately determined without human intervention.

FIG. 15 is a flowchart illustrating a method for determining a theme preference according to a second embodiment of the present invention. However, it should be understood that the present invention is not limited thereto and that some steps may be added or deleted as needed.

Referring to FIG. 15, in step S101, the chattob service server 100 groups the utterance contents included in the conversation log by topic.

In step S104, the chattob service server 100 extracts a pre-set speech-response relationship from the topic utterance contents. According to an embodiment, the preset speech-response relationship may be set based on the conversation log of the user, and may be set based on a plurality of conversation logs of the user.

In step S105, the chattob service server 100 determines the preference by topic based on the number of extracted speech-response relationships.

Up to now, a method for determining a theme preference according to the second embodiment has been described with reference to FIG. According to the above, the topic preference is determined considering the qualitative aspect of the utterance about the subject. Accordingly, it is possible to prevent the subject matter preference from being determined incorrectly.

On the other hand, as described above, the theme preference may be determined using both the number of utterances and the number of utterance-response relationships in order to consider both quantitative and qualitative aspects of utterance on a particular subject. An example of this is shown in Fig.

Hereinafter, a personalized chatbot service providing method according to some embodiments of the present invention will be described with reference to FIG. 16 and FIG.

16 is a flowchart illustrating a method of providing a personalized chatbot service according to the first embodiment of the present invention. However, it should be understood that the present invention is not limited thereto and that some steps may be added or deleted as needed.

Referring to FIG. 16, in step S201, the chattob service server 100 inquires about the preference information of the first user. For example, the chatbot service server 100 may inquire the user's preference information of the first user in the user model in response to the chat request of the first user.

In step S203, the chattob service server 100 determines a candidate topic using the utterance of the first user (S203). For example, when a clue to the candidate topic is detected in the utterance of the first user, the chatbot service server 100 may determine the candidate topic using the clue. The relationship between the clue and the candidate topic may be predetermined according to a predetermined rule.

However, as described above, even when a specific clue is not detected in the utterance of the first user, the chattob service server 100 may provide the utterance utterance using the utterance preference information of the first user. An example of this is shown in Fig.

In step S205, the chattob service server 100 determines a topic to be proposed among candidate topics using the preference information of the first user.

In step S207, the chattop service server 100 generates and transmits the topic presentation speech of the chatbot on the proposed topic.

Up to now, a personalized chatbot service providing method according to the first embodiment has been described with reference to FIG. As described above, the chatbots can provide a topic of interest to the user with a high probability before presenting the user's topic. Accordingly, a more intelligent chatbot service can be provided, and the satisfaction of the user using the chatbot service can be greatly improved.

17 is a flowchart illustrating a method of providing a personalized chatbot service according to a second embodiment of the present invention. However, it should be understood that the present invention is not limited thereto and that some steps may be added or deleted as needed.

In step S301, the chat bot service server 100 inquires about the speech-response relationship of the first user. For example, the chatbot service server 100 may inquire the user's speech-response relationship in the user model of the first user in response to the chat request of the first user.

In step S303, the chattop service server 100 adjusts the standard conversation flow of the current topic using the inquired-response relationship.

In step S305, the chattop service server 100 proceeds with the conversation using the adjusted standard conversation flow. The description of steps S303 and S305 will be made with reference to the example shown in FIG. 13 and the description of FIG.

Up to now, a method of providing personalized chatbot service according to the second embodiment has been described with reference to FIG. In accordance with the above, customized chatbot services can be provided for each user based on the adjusted standard conversation flow. In addition, since the conversation process can be performed quickly, the satisfaction of the user using the chat service can be further improved.

Lastly, a hardware configuration diagram of the chattop service server 100 according to an embodiment of the present invention will be described with reference to FIG.

18, the chattop service server 100 includes one or more processors 101, a memory 102 for loading a computer program executed by the processor 101, a storage 103 bus 104, And a network interface 105. Fig. 18 shows only the components related to the embodiment of the present invention. Therefore, it will be understood by those skilled in the art that other general-purpose components other than the components shown in FIG. 18 may be further included.

The processor 101 controls the overall operation of each configuration of the chatbot service server 100. The processor 101 includes a central processing unit (CPU), a microprocessor unit (MPU), a microcontroller unit (MCU), a graphics processing unit (GPU), or any type of processor well known in the art . The processor 101 may also perform operations on at least one application or program to perform the method according to embodiments of the present invention. The chatbot service server 100 may include one or more processors.

The memory 102 stores various data, steps and / or information. The memory 102 may load a computer program comprised of a series of instructions from the storage 103 to perform a method for determining a subject preference and / or a method for providing a personalized chat service, in accordance with embodiments of the present invention.

In a first example, a first computer program may be loaded into the memory 102 to cause the processor 101 to execute a method of determining a subject preference according to some embodiments of the present invention. To this end, the first computer program may consist of the following series of instructions.

In the example 1-1, the series of instructions includes instructions 130 and 132 for distinguishing utterances constituting a dialog between the first user and the chatbot by topic, (140, 144) for determining the preference of the first user for the first subject using the total amount of speech to which the first user belongs. In addition, the series of instructions may further include an analysis analysis instruction 134, an emotion analysis instruction 136, and / or an entity name recognition instruction 138 to estimate the subject's utterance volume.

In the first to second examples, the series of instructions includes instructions (130, 132) for distinguishing each utterance constituting a conversation between a first user and a chatbot, and a predefined utterance-response relationship (142, 144) for determining a preference of the first user for the first subject using the number of times the first user has been extracted. Further, for the above-described ignition-response relationship extraction, the series of instructions may further include an entity name recognition instruction 138. [

In a second example, the second computer program may be loaded into the memory 102 to cause the processor 101 to execute a personalized chatbot service provisioning method according to some embodiments of the present invention. To this end, the second computer program may consist of the following sequence of instructions.

In the example 2-1, the series of instructions may include instructions to obtain a subject-specific preference of a first user from the subject-specific preferences DB 118 and instructions to use the subject-specific preference of the first user To generate a topic presentation utterance (146).

In the example 2-2, the series of instructions includes an instruction to inquire the first user's speech-response relationship from the subject preference DB 118, a standard conversation flow (" and instructions for providing a chat service to the first user based on the coordinated standard chat flow.

The description of each instruction shown in FIG. 18 is omitted in order to avoid redundant description.

The storage 103 may non-provisionally store the conversation log 102, the subject matter preference DB 118 and one or more programs (not shown).

The storage 103 may be a nonvolatile memory such as a ROM (Read Only Memory), an EPROM (Erasable Programmable ROM), an EEPROM (Electrically Erasable Programmable ROM), a flash memory or the like, a hard disk, a removable disk, And any form of computer-readable recording medium known in the art.

The bus 104 provides communication functions between the components of the chattcom service server 100. The bus 104 may be implemented as various types of buses such as an address bus, a data bus, and a control bus.

The network interface 105 supports wired / wireless Internet communication of the chat service server 100. In addition, the network interface 105 may support various communication methods other than Internet communication. To this end, the network interface 105 may comprise a communication module well known in the art.

Up to now, a hardware configuration diagram of the chattcom service server 100 according to the embodiment of the present invention has been described with reference to FIG.

Up to now, some embodiments of the present invention have been described with reference to Figs. 1 to 18, and effects have been described according to some of the above embodiments. However, the effects of the present invention are not limited to the effects mentioned above, and other effects not mentioned can be clearly understood to those of ordinary skill in the art from the following description.

Up to this point, the concepts of the present invention described with reference to Figs. 1 to 18 can be embodied in computer readable code on a computer readable medium. The computer readable recording medium may be, for example, a removable recording medium (CD, DVD, Blu-ray disk, USB storage device, removable hard disk) . The computer program recorded on the computer-readable recording medium may be transmitted to another computing device via a network such as the Internet and installed in the other computing device, thereby being used in the other computing device.

Although the operations are shown in the specific order in the figures, it should be understood that the operations need not necessarily be performed in the particular order shown or in a sequential order, or that all of the illustrated operations must be performed to achieve the desired result. In certain situations, multitasking and parallel processing may be advantageous. Moreover, the separation of the various configurations in the above-described embodiments should not be understood as such a separation being necessary, and the described program components and systems may generally be integrated together into a single software product or packaged into multiple software products .

While the present invention has been described in connection with what is presently considered to be practical exemplary embodiments, it is to be understood that the invention is not limited to the disclosed embodiments, but, on the contrary, I can understand that. It is therefore to be understood that the above-described embodiments are illustrative in all aspects and not restrictive.

Claims (15)

A method performed by a computing device,
Dividing each utterance constituting a dialog between the first user and the chatbot by topic; And
Determining a preference for the first subject by the first user using the total amount of speech pertaining to the first subject,
How to determine preferences by subject. The method according to claim 1,
Wherein the determining the preference for the first subject comprises:
Extracting first feature data for utterances of the first user during a total utterance belonging to the first subject;
Extracting second characteristic data for utterance of the chatbot during the entire utterance belonging to the first subject; And
Calculating the quantity using the first feature data and the second feature data,
Wherein the first feature data and the second feature data are selected by different criteria,
How to determine preferences by subject. 3. The method of claim 2,
Wherein the first feature data comprises:
The number of utterances of the inquiry type, the number of utterances of the request type, and the sum of the number of utterances of the notification type,
How to determine preferences by subject. 3. The method of claim 2,
Wherein the first feature data comprises:
The number of entity names included in the query type's utterance, the number of entity names included in the request type's utterance, and the sum of the number of entity names included in the utterance of the notification type,
How to determine preferences by subject. 3. The method of claim 2,
Wherein the first feature data comprises:
The number of emotion display words,
How to determine preferences by subject. 3. The method of claim 2,
The second feature data may include:
And the number of entity names included in the speech of the chatbot to which the answer of the first user is affirmative,
How to determine preferences by subject. 3. The method of claim 2,
The second feature data may include:
The number of object names included in the utterance of the notification type,
How to determine preferences by subject. A method performed by a computing device,
Dividing each utterance constituting a dialog between the first user and the chatbot by subject; And
Determining a preference for the first subject by the first user using the number of times the predetermined speech-response relationship is extracted in the conversation pertaining to the first subject,
How to determine preferences by subject. 9. The method of claim 8,
Wherein the predetermined speech-response relationship is extracted from a conversation log with a plurality of users of the chatbot,
How to determine preferences by subject. 9. The method of claim 8,
Wherein the predetermined speech-response relationship is extracted from a chat log with the first user of the chatbot,
How to determine preferences by subject. A method performed by a computing device,
Obtaining a preference of the first user based on the theme; And
Wherein the chatbot generates a topic presentation using the subject matter preference of the first user without the subject presentation of the first user.
A method of providing a personalized chatbot service. 12. The method of claim 11,
Wherein the generating the topic presentation comprises:
Generating a first subject presentation of the first subject; And
Generating a second topic presentation speech for a second subject in response to the first user's negative speech for the first topic presentation speech,
Wherein the second subject is lower than the preference of the first subject,
A method of providing a personalized chatbot service. 12. The method of claim 11,
Wherein the generating the topic presentation comprises:
Selecting a plurality of candidate topics using utterances of the first user;
And generating a topic presentation utterance for the selected topic using the preference of the first user among the plurality of candidate topics.
A method of providing a personalized chatbot service. A method performed by a computing device,
Querying a user model of a first user, the user model comprising one or more speech-response relationships;
Adjusting a standard dialog flow using the spoken-response relationship of the searched user model; And
And providing a chattb service for the first user based on the coordinated standard conversation flow.
A method of providing a personalized chatbot service. 15. The method of claim 14,
Wherein the standard conversation flow includes a transition from a first intent to a second intent and a transition from the second intent to a third intent,
In the spoken-to-answer relationship of the searched user model, if the relationship is that the utterance is a second intention and the response is a third intention, then the coordinated standard conversation flow is a transition from the first intention to the second intention Wherein the second intention is replaced with a transition from the first intent to the third intent.
A method of providing a personalized chatbot service.

Images