KR20230169681A - Method and apparatus for outputting user-specific responses by reflecting relationship with artificial intelligence - Google Patents

Abstract

A method and device for outputting user-specific responses reflecting the relationship with artificial intelligence are presented. A method of outputting a user-specific response by reflecting the relationship between artificial intelligence and a user in artificial intelligence according to an embodiment includes the steps of determining intimacy between the user and artificial intelligence; When a user's speech is input, selecting a response database according to the intimacy level; and outputting a response message based on the response database selected according to the intimacy.

Description

Method and device for outputting user-specific responses by reflecting the relationship with artificial intelligence {METHOD AND APPARATUS FOR OUTPUTTING USER-SPECIFIC RESPONSES BY REFLECTING RELATIONSHIP WITH ARTIFICIAL INTELLIGENCE}

The following embodiments relate to a method and device for outputting user-specific responses by reflecting the relationship with artificial intelligence. More specifically, artificial intelligence that generates speech by referring to conditions external to the conversation, such as the relationship with the user and the environment. It relates to a method and device for outputting a user-specific response by reflecting the relationship with the user.

With the advancement of technology, much of the physical labor performed by humans has been replaced by mechanical devices. Recently, with the development of artificial intelligence technology, it is expected that many fields that are considered human mental areas can be replaced by artificial intelligence.

In particular, research is being actively conducted on artificial intelligence conversation systems that recognize and analyze human language and provide appropriate answers that are difficult to distinguish from humans. To this end, an artificial intelligence conversation system using machine learning techniques is being discussed.

Machine learning is a technique that increases the accuracy of classification by classifying and patterning a large amount of data and changing/modifying the classification and patterning through continuous input data. As one of the machine learning algorithms for classifying data, the Neural Network algorithm, based on the operating principles of neurons in the human brain, is receiving particular attention.

Korean Patent Publication No. 10-2022-0070826 describes a speech manipulation device and method for selecting a response for the next speech in such a search-based dialogue system.

Korean Patent Publication No. 10-2022-0070826

Embodiments describe a method and device for outputting user-specific responses by reflecting the relationship with artificial intelligence, and more specifically, a search-based chatbot ( chatbot) provides technology to output user-specific responses by reflecting the relationship with the chatbot.

Embodiments include a method of determining the intimacy between a user and artificial intelligence, selecting a response database according to the intimacy, and outputting a response specialized for the user by reflecting the relationship with artificial intelligence, which searches for and outputs a response message based on the response of the selected response database. and providing devices.

A method of outputting a user-specific response by reflecting the relationship between artificial intelligence and a user in artificial intelligence according to an embodiment includes the steps of determining intimacy between the user and artificial intelligence; When a user's speech is input, selecting a response database according to the intimacy level; and outputting a response message based on the response database selected according to the intimacy.

In the step of determining the intimacy, the intimacy may be calculated for each turn by analyzing the conversation between the user and the artificial intelligence.

In the step of determining the intimacy, the intimacy can be calculated by reflecting the number of conversation turns with the artificial intelligence as a variable.

In the step of determining the intimacy, the intimacy can be calculated by reflecting the presence or absence of an abusing speech in a conversation with artificial intelligence as a variable.

In the step of selecting a response database according to the intimacy, one response database among a plurality of response databases pre-built according to the calculated intimacy may be selected.

In the step of outputting a response message based on the response database, a response message may be searched and output as a response from the selected response database.

The artificial intelligence includes a search-based chatbot, and can rank response candidate sentences in the response database selected by the search-based chatbot based on the user's last utterance message and select the highest-ranked sentence.

In artificial intelligence according to another embodiment, a device that reflects the relationship between artificial intelligence and a user and outputs a user-specific response includes: an intimacy determination unit that determines the intimacy between the user and artificial intelligence; a response database selection unit that selects a response database according to the level of intimacy when a user's utterance is input; and a response message output unit that outputs a response message based on the response database selected according to the intimacy.

The intimacy determination unit may calculate intimacy for each turn by analyzing the conversation between the user and the artificial intelligence.

The intimacy determination unit may calculate intimacy by reflecting the number of conversation turns with the artificial intelligence as a variable.

The intimacy determination unit may calculate intimacy by reflecting the presence or absence of an abusing utterance in a conversation with artificial intelligence as a variable.

The response database selection unit may select one response database from a plurality of response databases built in advance according to the calculated intimacy.

The response message output unit may search for and output a response message as a response from the selected response database.

The artificial intelligence includes a search-based chatbot, and can rank response candidate sentences in the response database selected by the search-based chatbot based on the user's last utterance message and select the highest-ranked sentence.

According to embodiments, the intimacy between the user and the artificial intelligence is determined and a response database is selected according to the intimacy, and a response specialized for the user is output by reflecting the relationship with the artificial intelligence, which searches and outputs a response message with the response of the selected response database. A method and device for doing so can be provided.

1 is a diagram showing an example of a system for explaining the present embodiments.
FIG. 2 is a diagram showing the configuration of an embodiment to explain the internal configuration of the user terminal and server shown in FIG. 1.
Figure 3 shows a flowchart showing a method of outputting a user-specific response by reflecting the relationship with artificial intelligence according to an embodiment.
Figure 4 is a block diagram showing a device that outputs a user-specific response by reflecting the relationship with artificial intelligence according to an embodiment.
Figure 5 is a diagram for explaining a search-based chatbot and a creation-based chatbot according to an embodiment.
Figure 6 is a diagram for explaining an utterance score labeling method according to an embodiment.
FIG. 7 is a diagram illustrating a method of automatically obtaining a response selection model labeling score from a user's utterance according to an embodiment.
Figure 8 is a diagram illustrating a method of outputting a user-specific response by reflecting the relationship with artificial intelligence according to an embodiment.
FIG. 9 is a diagram illustrating a method of outputting a user-specific response by reflecting the relationship with the chatbot in a search-based chatbot according to an embodiment.

Hereinafter, embodiments will be described with reference to the attached drawings. However, the described embodiments may be modified into various other forms, and the scope of the present invention is not limited to the embodiments described below. In addition, various embodiments are provided to more completely explain the present invention to those with average knowledge in the art. The shapes and sizes of elements in the drawings may be exaggerated for clearer explanation.

The following embodiments provide technology for outputting user-specific responses by reflecting the relationship between artificial intelligence and artificial intelligence. Accordingly, artificial intelligence can generate speech by referring to conditions external to the conversation, such as the relationship with the user and the environment.

The method and device for outputting a user-specific response reflecting the relationship with artificial intelligence according to this embodiment recognizes the user's voice or recognizes the user's chat input through a terminal owned by the user to establish intimacy with the user. A response message can be sent based on .

Here, the terminal is a PC, smartphone, tablet PC, etc., and a method for outputting a user-specific response reflecting the relationship with artificial intelligence according to this embodiment and an application dedicated to access or service to a web/mobile site related to the device ( It can refer to any terminal device that can install and run an application. At this time, the terminal can perform overall service operations, such as service screen configuration, data input, data transmission and reception, and data storage, under the control of a web/mobile site or dedicated application.

Furthermore, these embodiments are not limited to conversations and chats online or through applications through a terminal, and can be applied to various products or services that can provide response data to conversation type data recognized from the user's voice, gestures, and facial expressions. do.

Hereinafter, with reference to FIGS. 1 to 9, a method and device for outputting a user-specific response reflecting the relationship with artificial intelligence according to the present embodiments will be described in detail.

1 is a diagram showing an example of a system for explaining the present embodiments. The system environment in FIG. 1 shows an example including a user terminal 110, a server 120, and a network 130. Figure 1 is an example for explaining the invention, and the number of user terminals or servers is not limited as in Figure 1.

Here, the user terminal 110 is a terminal installed with software or an application that outputs a user-specific response by reflecting the relationship with artificial intelligence according to the present embodiments, and includes a touch screen for the user 101's touch input. It may include a mobile terminal. The user terminal 110 may include any terminal capable of outputting a response message in artificial intelligence, such as a smart phone, mobile phone, computer, VR device, AR device, etc. For example, the user terminal 110 may communicate with other mobile terminals and/or the server 120 through the network 130 using wireless or wired communication methods.

The communication method is not limited, and may include not only a communication method utilizing a communication network that the network 130 may include (for example, a mobile communication network, wired Internet, wireless Internet, and a broadcast network), but also short-range wireless communication between devices. For example, the network 130 may include a personal area network (PAN), a local area network (LAN), a campus area network (CAN), a metropolitan area network (MAN), a wide area network (WAN), and a broadband network (BBN). , may include one or more arbitrary networks such as the Internet. Additionally, the network 130 may include any one or more of network topologies including a bus network, star network, ring network, mesh network, star-bus network, tree or hierarchical network, etc. Not limited.

The server 120 may be implemented as a computer device or a plurality of computer devices that communicate with the user terminal 110 and the network 130 to provide commands, codes, files, content, services, etc.

For example, the server 120 is a user terminal 110 connected through the network 130. For example, the server 120 is a file for installation of a service providing application that outputs a user-specific response by reflecting the relationship with artificial intelligence. can be provided. In this case, the user terminal 110 can install the application using a file provided from the server 120. In addition, the service provided by the server 120 is accessed by accessing the server 120 under the control of an operating system (OS) included in the user terminal 110 or at least one program (for example, a browser or the installed application). I can receive content. For example, when the user terminal 110 transmits a service request message to the server 120 through the network 130 under the control of an application, the server 120 sends a code corresponding to the service request message to the user terminal 110. ), and the user terminal 110 can provide content to the user by configuring and displaying a screen according to the code under the control of the application.

FIG. 2 is a diagram showing the configuration of an embodiment to explain the internal configuration of the user terminal and server shown in FIG. 1. In FIG. 2 , the internal configuration of the user terminal 110 as an example of one user terminal and the server 120 as an example of a single server are explained. If another user terminal or another server is provided, the other server may also have the same or similar internal configuration.

The user terminal 110 and the server 120 may include memories 211 and 221, processors 212 and 222, communication modules 213 and 223, and input/output interfaces 214 and 224. The memory 211, 221 is a computer-readable recording medium, and is a non-permanent mass recording device such as random access memory (RAM), read only memory (ROM), PROM, flash memory, and disk drive. storage device). Additionally, an operating system or at least one program code (for example, code for an application installed and running on the user terminal 110) may be stored in the memories 211 and 221. These software components may be loaded from a computer-readable recording medium separate from the memories 211 and 221. Such separate computer-readable recording media may include computer-readable recording media such as floppy drives, disks, tapes, DVD/CD-ROM drives, and memory cards. In another embodiment, software components may be loaded into the memories 211 and 221 through the communication modules 213 and 223 rather than a computer-readable recording medium. For example, at least one program is installed by files provided through the network 130 by developers or a file distribution system that distributes installation files for applications (e.g., the server described above). may be loaded into the memories 211 and 221 based on the application).

The processors 212 and 222 may be configured to process instructions of a computer program by performing basic arithmetic, logic, and input/output operations. Commands may be provided to the processors 212 and 222 by the memories 211 and 221 or the communication modules 213 and 223. For example, the processors 212 and 222 may be configured to execute received instructions according to program codes stored in a recording device such as the memory 211 and 221.

The communication modules 213 and 223 may provide a function for the user terminal 110 and the server 120 to communicate with each other through the network 130, and can be connected to another mobile terminal (for example, a second user terminal or another server). It can provide functions for communicating with. For example, a request (for example, a search request) generated by the processor 212 of the first user terminal 110 according to a program code stored in a recording device such as the memory 211 is transmitted to the network under the control of the communication module 213. It can be delivered to the server 120 through 130. Conversely, control signals, commands, content, files, etc. provided under the control of the processor 222 of the server 120 pass through the communication module 223 and the network 130 to the communication module of the first user terminal 110. It can be received by the first user terminal 110 through (213). For example, control signals or commands of the server 120 received through the communication module 213 may be transmitted to the processor 212 or memory 211, and content or files may be transmitted to the first user terminal 110. This may be stored in a storage medium that may further include.

The input/output interface 214 may be a means for interfacing with the input/output device 215. For example, an input device may include a device such as a keyboard or mouse, and an output device may include a device such as a display for displaying an application's communication session. As another example, the input/output interface 214 may be a means for interfacing with a device that integrates input and output functions, such as a touch screen. As a more specific example, the processor 212 of the first user terminal 110 is configured using data provided by the server 120 or the second user terminal in processing instructions of the computer program loaded in the memory 211. Service screens or content may be displayed on the display through the input/output interface 214. Likewise, the input/output interface 224 can output information composed using data provided by the server 120 when the processor 222 of the server 120 processes commands of a computer program loaded in the memory 221. there is.

Additionally, in other embodiments, the user terminal 110 and the server 120 may include more components than those shown in FIG. 2 . However, there is no need to clearly show most prior art components. For example, the user terminal 110 is implemented to include at least some of the above-described input/output devices 215 or other components such as a transceiver, a Global Positioning System (GPS) module, a camera, various sensors, a database, etc. More may be included. As a more specific example, when the user terminal 110 is a smartphone, the smartphone generally includes an acceleration sensor, a gyro sensor, a camera, various physical buttons, a button using a touch panel, an input/output port, a vibrator for vibration, etc. It can be seen that various components of can be implemented to be further included in the user terminal 110.

In this embodiment, the server 120 uses artificial intelligence of a learning model learned by a preset response message data set in an application that outputs a user-specific response by reflecting the relationship with artificial intelligence, for example, through a neural network. It learns the contents of the previous conversation between the intelligence and the user, and based on the learned conversation contents, it is possible to determine the transmission point and content of the response message to the user, and the content of the response message determined in this way to be sent to the user at the time of the response message. You can. At this time, the response message may be generated based on a response database selected according to intimacy with the user, and the transmission time may be when the user sends the preemptive message. Here, the present embodiments can transmit the contents of a response message to the user through an application that can exchange conversations between artificial intelligence and the user, and the present invention is applicable to all types of applications that provide not only text-based conversations but also voice-based conversations. It can be applied. Of course, an application that provides voice function conversation may include voice recognition and a function to convert the recognized voice into text in order to check the contents of the previous conversation, and after analyzing the converted text, a response message may be sent by learning the analyzed text. The content can be determined, and the function to convert the response message content into voice can also be included.

The user terminal in this embodiment may refer to a user terminal such as a smartphone that can download and install the application of this embodiment from the server shown in FIGS. 1 and 2.

Figure 3 is a flowchart showing a method of outputting a user-specific response by reflecting the relationship with artificial intelligence according to an embodiment, and shows a flowchart of operations in the server 120 shown in Figures 1 and 2.

Referring to FIG. 3, the method of outputting a user-specific response by reflecting the relationship between artificial intelligence and the user in artificial intelligence according to an embodiment includes determining the intimacy between the user and artificial intelligence (S110), and the user's When an utterance is input, it may include selecting a response database according to intimacy (S120) and outputting a response message based on the response database selected according to intimacy (S130).

The method of outputting a user-specific response reflecting the relationship with artificial intelligence according to an embodiment will be described in more detail by taking as an example a device that outputs a user-specific response reflecting the relationship with artificial intelligence according to an embodiment. can do.

Figure 4 is a block diagram showing a device that outputs a user-specific response by reflecting the relationship with artificial intelligence according to an embodiment.

Referring to FIG. 4, the device 400 that outputs a user-specific response by reflecting the relationship with artificial intelligence according to one embodiment includes an intimacy determination unit 410, a response database selection unit 420, and a response message output unit. It can be done including (430).

In step S110, the intimacy determination unit 410 may determine the intimacy between the user and the artificial intelligence. The intimacy determination unit 410 can calculate intimacy by analyzing the conversation between the user and artificial intelligence at each turn.

For example, the intimacy determination unit 410 may calculate intimacy by reflecting the number of conversation turns with the artificial intelligence as a variable. For example, the intimacy determination unit 410 may determine high intimacy if the number of conversation turns with the artificial intelligence is more than a preset number, and may judge it as low intimacy if the number of conversation turns with the artificial intelligence is less than the preset number. As another example, the intimacy determination unit 410 may calculate intimacy by reflecting the presence or absence of an abusing utterance as a variable in a conversation with artificial intelligence. For example, the intimacy determination unit 410 may determine high intimacy if there is an abusing utterance allowed in a conversation with artificial intelligence, and may judge it as low intimacy if there is no abusing utterance allowed. In other words, it is judged that there are some abusing speeches that are acceptable in intimate relationships. As another example, the intimacy determination unit 410 may calculate intimacy by reflecting both the number of conversation turns with the artificial intelligence and the presence or absence of abusing utterances as variables. For example, the intimacy determination unit 410 may determine high intimacy if the number of conversation turns with the artificial intelligence is more than a preset number and there is an allowed abusing utterance, and if the number of conversation turns with the artificial intelligence is less than a preset number and there is an allowed abusing utterance, the intimacy determination unit 410 may determine it to be high intimacy. Intimacy can be judged in cases where there are abusing utterances, and intimacy can be judged in cases where the number of conversation turns with artificial intelligence is more than a preset number and there are no permitted abusing utterances, and the number of conversation turns with artificial intelligence is a preset number. If it is less than that and there are no permitted abusing utterances, it can be judged as low intimacy.

Meanwhile, the intimacy determination unit 410 may determine a response message based on not only the intimacy with the user, but also the relationship with the user and the user's information. Here, the relationship with the user may include, for example, a friend relationship, a romantic relationship, a senior/junior relationship, etc., and the user's information may include, for example, gender, age, occupation, place of residence, etc.

For example, the intimacy determination unit 410 may determine not only the intimacy with the user but also the relationship with the user and calculate the intimacy based on this or select a response database. For example, if the relationship between the user and the artificial intelligence is that of same-sex friends in their 20s, a response message can be sent by selecting a response database that includes conversations that form a consensus between same-sex people, conversations of people in their 20s, etc. For example, if the intimacy with the user is judged to be high intimacy and the friend is in his 20s, you can select a high intimacy response database of a group of friends in their 20s that reflects this, and if the intimacy with the user is judged to be low intimacy and the friend is in his 20s, you can select it. You can select a friend group in your 20s from a low intimacy response database.

In this way, the response database can be divided and configured according to intimacy and user information.

In step S120, when a user's utterance is input, the response database selection unit 420 may select a response database according to intimacy. The response database selection unit 420 may select one response database from a plurality of response databases built in advance according to the calculated intimacy.

Here, as described above, the response database may be selected considering not only the intimacy with the user, but also the relationship with the user and the user's information.

In step S130, the response message output unit 430 may output a response message based on a response database selected according to intimacy.

The response message output unit 430 may search for and output a response message as a response from the selected response database. Here, the artificial intelligence may include a search-based chatbot, and the search-based chatbot may rank response candidate sentences in a response database selected based on the user's last utterance message and select the highest ranked sentence. Meanwhile, depending on the embodiment, artificial intelligence may include not only search-based chatbots but also creation-based chatbots.

Figure 5 is a diagram for explaining a search-based chatbot and a creation-based chatbot according to an embodiment.

The act of conveying words or writing in sentence units, or the sentence that is the object of the act, is called an utterance. When several turns of utterance come together, a conversation on one topic is established, which is called a conversation session or context. Referring to FIG. 5, when the context 510 and the user's utterance are given, the chatbot outputs a response (chatbot response) corresponding to the context.

Chatbots are broadly classified into two types according to the principle of outputting responses, and are classified into search-based chatbots (520) and generation-based chatbots (530).

The search-based chatbot 520 selects the most appropriate sentence for the given context 510 from the response candidate sentences in the pre-built response database 521 and outputs it as the chatbot's response. A model that selects the highest-ranked sentence by ranking response candidates for the input context is called a “response selection model 522”. At this time, the highest-ranked sentence means the most appropriate sentence.

The generation-based chatbot 530 receives the given context 510 as input, generates a response sentence by sequentially selecting words that appear next, and outputs it as the chatbot's response. The model that calculates the probability of occurrence of each word in the input context and outputs the word with the highest probability is called the “generative model 531”. Here, the word with the highest probability for the input context refers to the most likely word for the input context.

Figure 6 is a diagram for explaining an utterance score labeling method according to an embodiment.

Referring to Figure 6, in order to learn a response selection model, a dataset in which the scores of response candidates (b) for an arbitrary context (a) are marked (labeled) is required. Accordingly, it is possible to implement an interface that can easily label scores for context-response candidate pairs.

The labeling tool for learning the response selection model randomly selects a context from the context database and secures n response candidates for the selected context by requesting them from the chatbot server (n=10 in the example). Then, the previously selected context and response sentences are provided in a screen-type interface as shown in FIG. 6. Accordingly, the labeler (eg, user) can move to the next context by pressing the check request button and the skip button. Additionally, the labeler can write itself by modifying the context as you click on sentences within the context. In this case, the response candidates are newly requested. Additionally, the labeler can score the answers by clicking on the answer candidates or pressing a shortcut key. For example, pressing number 1 can give 1 point, and pressing number 2 can give 2 points.

Accordingly, the response selection model can be learned through a labeling tool, and more appropriate responses can be output through learning.

FIG. 7 is a diagram illustrating a method of automatically obtaining a response selection model labeling score from a user's utterance according to an embodiment.

Referring to Figure 7, the response score of a chatbot can be automatically measured against a context database built from chatbot logs without the intervention of a human labeler.

After selecting a chatbot response and a user utterance pair corresponding to the response in the context 710, the score calculation model 720 can calculate the score of the corresponding chatbot response through the utterance pair. The score calculation model 720 may operate by deducting points for inappropriate attitudes (e.g., dissatisfaction with chatbot speech, sudden change of topic, etc.) and giving additional points to responsive attitudes (e.g., empathy for chatbot speech, etc.). You can. At this time, the score calculation model 720 has data on preset inappropriate attitudes and responsive attitudes, and according to this, the user's utterance corresponding to the chatbot response can be judged as either an inappropriate attitude or a responsive attitude.

Figure 8 is a diagram illustrating a method of outputting a user-specific response by reflecting the relationship with artificial intelligence according to an embodiment.

Referring to FIG. 8, a method of outputting a user-specific response by reflecting the relationship with artificial intelligence according to an embodiment is a method of outputting a user-specific response through the intimacy measurement module 810 when the user's utterance 801 is input. Intimacy can be judged.

At this time, the intimacy measurement module 810 may determine the intimacy based on the number of conversation turns with the user, the presence or absence of an abusing speech, etc. For example, if the number of conversation turns with the user is more than a predetermined number of turns, the user and artificial intelligence can be judged to have high intimacy. As another example, if the number of abusing utterances between the user and artificial intelligence is more than a certain level, and the artificial intelligence's response message according to the abusing utterance is not positive, it may be judged as low intimacy. As another example, if some abusing utterances are allowed between the user and artificial intelligence due to their existing high intimacy, the user and artificial intelligence may be judged to have high intimacy.

When the intimacy between the user and the artificial intelligence is determined, the response selection model 820 may select at least one response database 830 among a plurality of response databases according to the intimacy, and a response message based on the selected response database 830 ( 802) can be output.

FIG. 9 is a diagram illustrating a method of outputting a user-specific response by reflecting the relationship with the chatbot in a search-based chatbot according to an embodiment.

Referring to FIG. 9, a method of outputting a user-specific response by reflecting the relationship with the chatbot in the search-based chatbot 900 according to an embodiment is a method of outputting a response specialized for the user by reflecting the relationship between the user and the search-based chatbot 900. The output of the chatbot 900 can be controlled.

When the user's speech 901 is input, the intimacy measurement module 910 can calculate the conversation between the user and the chatbot at each turn. At this time, the intimacy measurement module 910 may consider variables such as the number of conversation turns between chatbots and the presence or absence of abusing speech.

When a user inputs an utterance, the response databases 931 and 932 can be selected according to familiarity. Here, the response database can be divided into a high intimacy response database 931 and a low intimacy response database 932. Additionally, it may be divided into a high intimacy response database 931, a medium intimacy response database, and a low intimacy response database 932. In this way, the response databases 931 and 932 may be composed of multiple response databases, and there is no limit to their number.

The response selection model 920 can search for and output the chatbot response message 921 only as responses from the corresponding intimacy response databases 931 and 932.

As described above, according to embodiments, by determining the intimacy between the user and the artificial intelligence and selecting a response database according to the intimacy, a response message can be searched and output as a response from the selected response database.

The device described above may be implemented with hardware components, software components, and/or a combination of hardware components and software components. For example, devices and components described in embodiments may include, for example, a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable array (FPA), It may be implemented using one or more general-purpose or special-purpose computers, such as a programmable logic unit (PLU), microprocessor, or any other device capable of executing and responding to instructions. A processing device may execute an operating system (OS) and one or more software applications that run on the operating system. Additionally, a processing device may access, store, manipulate, process, and generate data in response to the execution of software. For ease of understanding, a single processing device may be described as being used; however, those skilled in the art will understand that a processing device includes multiple processing elements and/or multiple types of processing elements. It can be seen that it may include. For example, a processing device may include multiple processors or one processor and one controller. Additionally, other processing configurations, such as parallel processors, are possible.

Software may include a computer program, code, instructions, or a combination of one or more of these, which may configure a processing unit to operate as desired, or may be processed independently or collectively. You can command the device. Software and/or data may be used on any type of machine, component, physical device, virtual equipment, computer storage medium or device to be interpreted by or to provide instructions or data to a processing device. It can be embodied in . Software may be distributed over networked computer systems and stored or executed in a distributed manner. Software and data may be stored on one or more computer-readable recording media.

The method according to the embodiment may be implemented in the form of program instructions that can be executed through various computer means and recorded on a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, etc., singly or in combination. Program instructions recorded on the medium may be specially designed and configured for the embodiment or may be known and available to those skilled in the art of computer software. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tapes, optical media such as CD-ROMs and DVDs, and magnetic media such as floptical disks. -Includes optical media (magneto-optical media) and hardware devices specifically configured to store and execute program instructions, such as ROM, RAM, flash memory, etc. Examples of program instructions include machine language code, such as that produced by a compiler, as well as high-level language code that can be executed by a computer using an interpreter, etc.

As described above, although the embodiments have been described with limited examples and drawings, various modifications and variations can be made by those skilled in the art from the above description. For example, the described techniques are performed in a different order than the described method, and/or components of the described system, structure, device, circuit, etc. are combined or combined in a different form than the described method, or other components are used. Alternatively, appropriate results may be achieved even if substituted or substituted by an equivalent.

Therefore, other implementations, other embodiments, and equivalents of the claims also fall within the scope of the claims described below.

Claims (14)

In artificial intelligence, a method of outputting a user-specific response by reflecting the relationship between artificial intelligence and the user,
Determining the intimacy between the user and artificial intelligence;
When a user's speech is input, selecting a response database according to the intimacy level; and
Outputting a response message based on the response database selected according to the intimacy
A method of outputting a user-specific response, including. According to paragraph 1,
The step of determining the intimacy is,
Analyzing the conversation between the user and the artificial intelligence to calculate intimacy at each turn
A method of outputting a user-specific response, characterized by: According to paragraph 2,
The step of determining the intimacy is,
Calculating intimacy by reflecting the number of conversation turns with artificial intelligence as a variable
A method of outputting a user-specific response, characterized by: According to paragraph 2,
The step of determining the intimacy is,
Calculating intimacy by reflecting the presence or absence of abusing speech as a variable in a conversation with artificial intelligence.
A method of outputting a user-specific response, characterized by: According to paragraph 1,
The step of selecting a response database according to the intimacy is,
Selecting one response database among a plurality of response databases built in advance according to the calculated intimacy.
A method of outputting a user-specific response, characterized by: According to paragraph 1,
The step of outputting a response message based on the response database is:
Searching for and outputting a response message based on the response from the selected response database
A method of outputting a user-specific response, characterized by: According to paragraph 1,
The artificial intelligence is,
A search-based chatbot comprising: ranking response candidate sentences in the response database selected by the search-based chatbot based on the user's last utterance message and selecting the highest-ranking sentence;
A method of outputting a user-specific response, characterized by: In artificial intelligence, a device that outputs a user-specific response by reflecting the relationship between artificial intelligence and the user,
An intimacy determination unit that determines the intimacy between the user and artificial intelligence;
a response database selection unit that selects a response database according to the level of intimacy when a user's utterance is input; and
A response message output unit that outputs a response message based on the response database selected according to the intimacy.
A device that outputs a user-specific response, including a. According to clause 8,
The intimacy determination unit,
Analyzing the conversation between the user and the artificial intelligence to calculate intimacy at each turn
A device that outputs a user-specific response, characterized by: According to clause 9,
The intimacy determination unit,
Calculating intimacy by reflecting the number of conversation turns with artificial intelligence as a variable
A device that outputs a user-specific response, characterized by: According to clause 9,
The intimacy determination unit,
Calculating intimacy by reflecting the presence or absence of abusing speech as a variable in a conversation with artificial intelligence.
A device that outputs a user-specific response, characterized by: According to clause 9,
The response database selection unit,
Selecting one response database among a plurality of response databases built in advance according to the calculated intimacy.
A device that outputs a user-specific response, characterized by: According to clause 8,
The response message output unit,
Searching for and outputting a response message based on the response from the selected response database
A device that outputs a user-specific response, characterized by: According to clause 8,
The artificial intelligence is,
A search-based chatbot comprising: ranking response candidate sentences in the response database selected by the search-based chatbot based on the user's last utterance message and selecting the highest-ranking sentence;
A device that outputs a user-specific response, characterized by:

Images