JP2021022928A - Artificial intelligence-based automatic response method and system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an automatic response method and a system of an artificial intelligence base.
An artificial intelligence-based automatic response method analyzes an intention by using a stage of receiving a utterance voice of the user by connecting a call with the user and an intermediate result (mid results) of the utterance voice. This includes a step of generating a reply and a step of sending the reply to the spoken voice.
[Selection diagram] Fig. 5

Description

The following description relates to an automatic response system (ARS) based on artificial intelligence (AI).

With the development of information and communication technology, the information society is advancing, and the Internet is becoming an important and indispensable medium in all fields such as society, culture, and economy.

In order to make inquiries, make reservations, deliveries, etc. to various companies, there is an offline method in which a user calls the company directly or conveys the request through an agency company.

Improvements over the offline method include connecting from a website to a customer center of a related company based on the development of Internet technology, and using the ARS method based on the development of mobile technology such as smartphones. , By installing and running the customer center application (App, Application), it has become possible to process necessary items.

For example, Patent Document 1 (publication date: March 29, 2019) discloses a technique for providing a customer center connection service on a screen to a customer based on artificial intelligence.

Korean Publication No. 10-2019-0033138

By grasping the meaning of the utterance from the intermediate results (mid results) for the user utterance and preparing the reply in advance, a method and a system capable of promptly providing the reply are provided.

Provided are a method and a system capable of interrupting the sending of a reply when a voice signal uttered by a user is received while the reply is being sent.

Provided are a method and a system capable of providing different response speeds according to the speed of user utterance.

Provide methods and systems that can use real-time translation to translate into the language spoken by the user and provide a response.

An artificial intelligence-based automatic response method performed by a computer system, wherein the computer system comprises at least one processor configured to execute a computer-readable instruction contained in the memory of the artificial intelligence infrastructure. The automatic response method is a stage in which the user is connected to a call by the at least one processor to receive the spoken voice of the user, and the at least one processor utilizes the intermediate results (mid results) of the spoken voice. Provided is an artificial intelligence-based automatic response method including a step of generating a response after analyzing an intention and a step of sending the response to the spoken voice by the at least one processor.

According to one aspect, the generation step is the step of extracting a sample sentence from the conversation log of the automatic response service, and the conversation learning using the sentence excluding the ending from the sample sentence as learning data. It may include a step of analyzing the intent.

According to another aspect, the generation step may include a step of analyzing the intent on a syllable basis obtained as the intermediate result.

According to another aspect, the generation step may include a step of determining the generation time of the response based on the confidence of the intention analysis result using the intermediate result.

According to another aspect, in the generation stage, the response is made in advance when the confidence for the intention analysis result using the intermediate result reaches the threshold value determined by the conversation learning using the sentence excluding the ending. It may include a step of generation.

According to still another aspect, the sending step may send the response as a voice signal if an endpoint is detected from the spoken voice.

An artificial intelligence-based automatic response method performed by a computer system, wherein the computer system comprises at least one processor configured to execute a computer-readable instruction contained in the memory of the artificial intelligence infrastructure. The automatic response method includes a step of receiving the spoken voice of the user by connecting a call with the user by the at least one processor, a step of sending a reply to the spoken voice by the at least one processor, and the step of sending a reply to the spoken voice. Provided is an artificial intelligence-based automatic response method including a step of interrupting the transmission of the response if the spoken voice of the user is received while the response is being transmitted by at least one processor.

According to one aspect, the suspending step may be interrupted immediately after the user's spoken voice is received or after applying fade-out.

According to another aspect, the interrupting step is a step of classifying and learning non-verbal expressions, and the learned non-verbal speech received while the response is being sent. It may include a step of maintaining the sending of the reply when it corresponds to the target expression and interrupting the sending of the reply when it does not correspond to the learned nonverbal expression.

According to another aspect, the automatic response method of the artificial intelligence base is a question-and-answer method for guiding the information when the information required for the response is insufficient as the recognition result of the spoken voice by the at least one processor. It may further include the step of asking a question.

According to another aspect, when a plurality of intentions are recognized by the recognition result of the spoken voice by the at least one processor, the automatic response method of the artificial intelligence base responds to each intention according to the recognition order of the intentions. It may further include the steps of providing in sequence.

According to another aspect, the automatic response method of the artificial intelligence base is a step of recognizing the utterance speed of the user by the at least one processor, and the response by the utterance speed of the user by the at least one processor. It may further include a step of determining the speaking speed of.

According to yet another aspect, the artificial intelligence-based automatic response method is a stage of recognizing the user's utterance language by the at least one processor, and a language model for the automatic response service by the at least one processor. May further include the step of switching to a language model corresponding to the user's utterance language.

Provided is a non-temporary computer-readable recording medium in which a program for causing a computer to execute the automatic response method of the artificial intelligence base is recorded.

A computer system comprising at least one processor configured to execute a computer-readable instruction contained in memory, said at least one processor producing the user's uttered voice by connecting the user to a call. Provided is a computer system that processes a process of receiving, a process of generating a response after analyzing an intention using an intermediate result of the spoken voice, and a process of sending the reply to the spoken voice.

A computer system comprising at least one processor configured to execute a computer-readable instruction contained in memory, said at least one processor producing the user's spoken voice by connecting the user to a call. A computer system that processes a process of receiving, a process of sending a reply to the spoken voice, and a process of interrupting the sending of the reply if the user's spoken voice is received while the reply is being sent. I will provide a.

According to the embodiment of the present invention, the response can be provided promptly by grasping the meaning of the utterance from the intermediate result for the user utterance and preparing the response in advance before the end of the utterance.

According to the embodiment of the present invention, it is possible to realize a form in which a person actually talks to a person by interrupting the transmission of the response when a voice signal uttered by the user is received during the transmission of the response. ..

According to the embodiment of the present invention, by providing different response speeds according to the user's utterance speed, it is possible to provide a service suitable for an interaction suitable for the user's utterance speed.

According to an embodiment of the present invention, real-time translation is used to translate into the language spoken by the user and provide a response, thereby improving accessibility and convenience of the service without being restricted by the language. it can.

It is a figure which showed the example of the network environment in one Embodiment of this invention. It is a block diagram for demonstrating the internal structure of an electronic device and a server in one Embodiment of this invention. It is explanatory drawing for demonstrating the AI automatic response system in one Embodiment of this invention. It is a figure which showed the example of the component which can include the AI automatic response system in one Embodiment of this invention. It is a flowchart which showed the example of the AI automatic response method in one Embodiment of this invention. It is a figure for demonstrating the example of the learning sentence in one Embodiment of this invention. It is explanatory drawing for demonstrating the process of analyzing the intention of a user by using the intermediate result with respect to the speech speech of the user in one Embodiment of this invention. It is explanatory drawing for demonstrating the process of analyzing the intention of a user by using the intermediate result with respect to the speech speech of the user in one Embodiment of this invention. It is a figure for demonstrating an example of the process of interrupting the sending of a reply in one Embodiment of this invention. It is a figure for demonstrating another example of the process of interrupting the sending of a reply in one Embodiment of this invention. It is a figure for demonstrating the exceptional situation which does not interrupt sending of a reply in one Embodiment of this invention.

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

Embodiments of the present invention relate to an automatic response system (ARS) based on artificial intelligence (AI).

The embodiments including the matters specifically disclosed in the present specification can provide an automatic response system realized in a form of actually talking with a human being based on artificial intelligence, whereby with a user. Inquiries, reservations, delivery orders, etc. can be processed quickly and conveniently while realizing a more natural call.

FIG. 1 is a diagram showing an example of a network environment according to an embodiment of the present invention. The network environment of FIG. 1 shows an example including a plurality of electronic devices 110, 120, 130, 140, a plurality of servers 150, 160, and a network 170. Such FIG. 1 is merely an example for explaining the invention, and the number of electronic devices and the number of servers are not limited as in FIG.

The plurality of electronic devices 110, 120, 130, 140 may be fixed terminals or mobile terminals realized by a computer system. Examples of a plurality of electronic devices 110, 120, 130, 140 include AI speakers, smartphones, mobile phones, navigation systems, PCs (personal computers), notebook PCs, digital broadcasting terminals, PDAs (Personal Digital Assistants), and PMPs (PMPs (Personal Digital Assistants). There are Portable Multimedia Players, tablets, game consoles, wearable devices, IoT (internet of things) devices, VR (virtual reality) devices, AR (augmented reality) devices, and the like. As an example, in FIG. 1, an AI speaker is shown as an electronic device 110, but in the embodiment of the present invention, the electronic device 110 substantially utilizes a wireless or wired communication method, and another device 110 is used via a network 170. It may mean one of a variety of physical computer systems capable of communicating with electronics 120, 130, 140 and / or servers 150, 160.

The communication method is not limited, and not only the communication method using the communication network (for example, mobile communication network, wired Internet, wireless Internet, broadcasting network, satellite network, etc.) that can be included in the network 170, but also the device. Short-range wireless communication between them may be included. For example, the network 170 includes a PAN (personal area network), a LAN (local area network), a CAN (campus area network), a MAN (metoropolitan area network), a WAN (wide network), etc. It may include any one or more of the networks. Further, network 170 may include, but is limited to, any one or more of network topologies including bus networks, star networks, ring networks, mesh networks, star-bus networks, tree or hierarchical networks, and the like. Will not be done.

The servers 150, 160 are communicated with a plurality of electronic devices 110, 120, 130, 140 via a network 170, respectively, by one or more computer devices that provide instructions, codes, files, contents, services, and the like. It may be realized. For example, the server 150 may be a system that provides the first service to a plurality of electronic devices 110, 120, 130, 140 connected via the network 170, and the server 160 may also be a plurality of systems connected via the network 170. It may be a system that provides a second service to electronic devices 110, 120, 130, 140. As a more specific example, the server 150 passes through an application which is a computer program installed and executed in a plurality of electronic devices 110, 120, 130, 140, and a service (automatic response as an example) intended by the application. (Service, etc.) may be provided to a plurality of electronic devices 110, 120, 130, 140 as a first service. As another example, the server 160 may provide a service as a second service that distributes files for installing and executing the above-mentioned application to a plurality of electronic devices 110, 120, 130, 140.

FIG. 2 is a block diagram for explaining the internal configurations of the electronic device and the server according to the embodiment of the present invention. FIG. 2 describes the internal configuration of the electronic device 110 and the internal configuration of the server 150 as examples for the electronic device. Further, the other electronic devices 120, 130, 140 and the server 160 may have the same or similar internal configuration as the electronic device 110 or the server 150 described above.

The electronics 110 and server 150 may include memory 211,221, processors 212,222, communication modules 213 and 223, and input / output interfaces 214 and 224. The memories 211 and 221 may be non-temporary computer-readable recording media, and may be a RAM (random access memory), a ROM (read only memory), a disk drive, an SSD (solid state drive), or a flash memory (flash memory). ) Etc. may be included in a non-temporary mass recording device. Here, a non-temporary large-capacity recording device such as a ROM, SSD, flash memory, or disk drive is used in the electronic device 110 or the server 150 as another non-temporary recording device that is separated from the memories 211 and 221. May be included. Further, the memory 211 and 221 are installed in the operating system and at least one program code (for example, a browser installed and executed in the electronic device 110, or installed in the electronic device 110 to provide a specific service. Code for applications etc.) may be recorded. Such software components may be loaded from a computer-readable recording medium separate from the memories 211 and 221. Such other computer-readable recording media may include computer-readable recording media such as floppy® drives, disks, tapes, DVD / CD-ROM drives, memory cards, and the like. In other embodiments, software components may be loaded into memory 211 and 221 through communication modules 213 and 223 that are not computer readable recording media. For example, at least one program is a computer program installed by a file provided via network 170 by a file distribution system (eg, server 160 described above) that distributes developer or application installation files (eg, described above). It may be loaded into the memory 211 or 221 based on the application.

Processors 212 and 222 may be configured to process instructions in a computer program by performing basic arithmetic, logic, and input / output operations. Instructions may be provided to processors 212 and 222 by memory 211, 221 or communication modules 213 and 223. For example, processors 212 and 222 may be configured to execute instructions received according to program code recorded in a recording device such as memory 211 and 221.

The communication modules 213 and 223 may provide a function for the electronic device 110 and the server 150 to communicate with each other via the network 170, and the electronic device 110 and / or the server 150 may provide another electronic device (as an example). , Electronic device 120) or another server (as an example, server 160) may provide a function for communicating. As an example, a request generated by the processor 212 of the electronic device 110 according to a program code recorded in a recording device such as a memory 211 may be transmitted to the server 150 via the network 170 under the control of the communication module 213. On the contrary, control signals, instructions, contents, files, etc. provided under the control of the processor 222 of the server 150 are received by the electronic device 110 through the communication module 213 of the electronic device 110 via the communication module 223 and the network 170. May be done. For example, control signals, instructions, contents, files, etc. of the server 150 received through the communication module 213 may be transmitted to the processor 212 and the memory 211, and the contents, files, etc. may be further included in the electronic device 110. It may be recorded on a medium (the non-temporary recording device described above).

The input / output interface 214 may be a means for an interface with the input / output device 215. For example, an input device may include a device such as a keyboard, mouse, microphone, camera, and an output device may include a device such as a display, speaker, haptic feedback device, and the like. As another example, the input / output interface 214 may be a means for an interface with a device such as a touch screen in which functions for input and output are integrated into one. The input / output device 215 may be composed of an electronic device 110 and one device. Also, the input / output interface 224 of the server 150 may be a means for connecting to the server 150 or for interfacing with a device (not shown) for input or output that the server 150 can include. As a more specific example, when the processor 212 of the electronic device 110 processes an instruction of a computer program loaded in the memory 211, a service screen or content configured by using data provided by the server 150 or the electronic device 120. May be displayed on the display through the input / output interface 214.

Also, in other embodiments, the electronic device 110 and the server 150 may include more components than the components of FIG. However, most prior art components need not be clearly illustrated. For example, the electronic device 110 may be implemented to include at least a portion of the input / output devices 215 described above, and may further include other components such as transceivers, cameras, various sensors, databases, and the like. It may be. As a more specific example, when the electronic device 110 is an AI speaker, various sensors generally included in the AI speaker, a camera module, various physical buttons, buttons using a touch panel, input / output ports, and vibrations. Various components, such as a speaker for, may be realized to be further included in the electronic device 110.

The automatic response system provides an automatic response service platform that transmits various information related to restaurants, accommodation facilities, airline tickets, movies, performances, hospitals (medical treatment), travel, etc. based on conversations with users.

In the following, as an example, processing inquiries, reservations, delivery orders, etc. to the restaurant while naturally talking to the user on behalf of the restaurant clerk will be described, but this is only an example and is limited to this. It should not be, and it is applicable to all companies and fields where the automatic response system can be used.

FIG. 3 is an exemplary diagram for explaining the AI automatic response system according to the embodiment of the present invention.

For example, the AI automatic response system 300 according to the embodiment of the present invention may be realized on the server 150 described with reference to FIGS. 1 and 2.

Referring to FIG. 3, the AI automatic response system 300 may provide an automatic response service for processing inquiries, reservations, delivery orders, and the like to a plurality of companies 31 to 33.

The AI automatic answering system 300 may provide an artificial intelligence conversation with the user 301 by a telephone function (call) or a chatbot (chatbot) call, and provides the information desired by the user 301 based on the conversation with the user 301. Alternatively, the request of the user 301 may be transmitted to the companies 31 to 33.

The AI automatic response system 300 is realized by a database system (not shown) containing company information for each company 31 to 33 and one system in order to provide an automatic response service related to the companies 31 to 33. It may be realized by another system that can be linked. The AI automatic response system 300 may provide the information desired by the user 301 or process the request of the user 301 based on the company information.

For example, the AI automatic response system 300 receives a voice input "Can I order from company A?" By the user 301 from the electronic device 110 including an interface that operates based on the conversation with the user 301. To do. As a result, the AI automatic response system 300 recognizes and analyzes the voice input "Is it possible to order from company A?" Of the user 301 received from the electronic device 110, and then responds based on the company information "Currently, it is possible to order." Please tell me the menu. ”And the generated response may be sent to the electronic device 110 by voice signal. The AI automatic response system 300 summarizes the requests of the user 301, for example, the selected menu and the quantity, based on the conversation with the user 301, and then puts the summarized information into the corresponding company (one of 31 to 33). ) May be transmitted.

In the following, the automatic response method of the artificial intelligence base and the specific embodiment of the system will be described.

FIG. 4 is a diagram showing an example of components that can be included in the AI automatic response system according to the embodiment of the present invention.

The server 150 according to the present embodiment plays the role of a platform for providing an automatic response service related to a company. In particular, the server 150 may include an AI automatic response system 300 realized in a form of actually talking with a human being based on artificial intelligence.

As shown in FIG. 4, the AI automatic response system 300 may include a gateway 401, a voice recognizer 410, a dialog manager 420, a response generator 430, a speech synthesizer 440, a sensing unit 450, and a translator 460.

The gateway 401 has a receiving end (RX) that plays a role of receiving a voice input of a user of the electronic device 110 by a telephone function of an application installed in the electronic device 110 or a chatbot, and a response voice to the electronic device 110 by an ARS response. It may include a transmitting end (TX) that serves to transmit. Further, the gateway 401 may include a socket controller (socket controller) for controlling the transmission of the response voice.

The speech recognizer 410, the dialog manager 420, the response generator 430, the speech synthesizer 440, the sensing unit 450, and the translator 460 may be included as components of the processor 222 of the server 150. Depending on the embodiment, the components of processor 222 may be selectively included or excluded from processor 222. Also, depending on the embodiment, the components of processor 222 may be separated or merged to represent the functionality of processor 222.

Such processors 222 and components of processor 222 may directly process operations by control instructions or control server 150 for execution of the AI autoresponder method described below. .. For example, the processor 222 and the components of the processor 222 may be implemented to execute an instruction by the code of the operating system included in the memory 221 and the code of at least one program.

Here, the components of the processor 222 may be representations of different functions of the processor 222 that are executed by the processor 222 according to the instructions provided by the program code recorded in the server 150. For example, the voice recognizer 410 may be used as a functional representation of the processor 222 that controls the server 150 according to the instructions described above so that the server 150 recognizes the voice input received from the electronic device 110.

The processor 222 may read the necessary instructions from the memory 221 in which the instructions related to the control of the server 150 are loaded. In this case, the read instruction may include an instruction for controlling the processor 222 to execute the AI automatic response method described below.

The following AI auto-responder methods may not occur in the order shown in the figure, and some of the steps may be omitted or additional steps may be included.

FIG. 5 is a flowchart showing an example of the AI automatic response method according to the embodiment of the present invention.

At step 510, processor 222 may perform conversational learning using a learning data set that includes a sentence without endings and a response to the sentence. For conversation between the AI automatic response system 300 and the user, after labeling the sentences excluding the ending to the learning data level, the labeled learning data is learned by deep learning or machine learning to build a conversation learning model. You can. At this time, a response with a high probability may be configured together as a learning data set for each learning sentence and used for conversation learning.

FIG. 6 is a diagram for explaining an example of a learning sentence in one embodiment of the present invention.

For example, referring to FIG. 6, the sample sentence 601 may be extracted from the conversation log appearing in the automatic response service, and the sentence obtained by removing the ending from the extracted sample sentence may be used as the learning sentence 602. For example, using the remaining sentence "Is it possible to reserve a restaurant this Saturday?" Excluding the ending "?" From the sample sentence "Is it possible to reserve a restaurant this Saturday?" As a learning sentence. Good. By learning the rest of the sentences excluding the endings, it is possible to learn in advance a set of correct answers that indicate what kind of response should be made before the end of a word in the middle of a conversation. Excludes not only the closing ending, which means the ending ending of a sentence, but also the connecting ending contained in the connecting clause, the non-ending ending such as the transposed ending contained in the embedded clause, or the remaining components other than the core keywords. It is also possible to utilize sentences as learning sentences.

In other words, the AI automatic response system 300 includes a conversation learning model constructed in advance by learning sentences excluding endings.

Again, in step 520, in FIG. 5, the voice recognizer 410 may receive the user's uttered voice in a real-time stream from the electronic device 110 via the gateway 401 if a call is connected to the user of the electronic device 110. At this time, the voice recognition device 410 may convert the received utterance voice into text by STT (speech to text) in real time and transmit it to the dialog manager 420.

At step 530, the dialog manager 420 may analyze the user's intention on the text transmitted from the speech recognizer 410 based on natural language understanding (NLU) technology and conversation learning. In particular, the dialog manager 420 may grasp the user's intention in units of syllables of text. That is, the dialog manager 420 may grasp the user's intention by dividing it into syllable units acquired as intermediate results by real-time text conversion, instead of sentence units acquired as final results after the user's words are finished. The dialog manager 420 may grasp the meaning of the user's utterance in advance at the stage of the intermediate result before the end of the word.

At step 540, the response generator 430 may pre-generate a response based on the intent analysis result. In particular, the response generator 430 may determine the response generation time point based on the confidence of the intention analysis result using the intermediate result. As an example, the response generator 430 preliminarily sends a response corresponding to the user's intention at the corresponding time point when the confidence of the intention analysis result becomes equal to or higher than a predetermined threshold value at the stage of the intermediate result before the end of the word. May be generated. The confidence for determining the response generation time may be determined by conversation learning using sentences excluding the ending. For example, the process of finding a numerical value whose confidence for the intermediate result is within the error range compared to the final result is repeated. After that, the confidence threshold for determining the response generation time may be determined based on the statistical value of the value obtained by the iterative process.

7 and 8 are illustrations for explaining a process of analyzing a user's intention by using an intermediate result for a user's spoken voice in one embodiment of the present invention.

Referring to FIG. 7, the user's intention is analyzed based on the intermediate result before the end of the word for the user's utterance voice "I am planning to take my parents. Is there a recommended menu for elderly people?" In that case, a threshold confidence that can generate a response when the "parents", "elderly", and "menu" are grasped may appear. It is possible to generate a response in advance before the end of the user's words, that is, at the time when the user issues the ending "is there?".

As another example, referring to FIG. 8, in the case of the automatic response service for reservation, the user's intention is set based on the intermediate result for the user's utterance voice "Please make a reservation with three people this Saturday." It will be analyzed, but at this time, if the intention analysis result meets all the necessary information (information slot) predetermined in relation to the reservation, it has reached the confidence that a reply can be generated. It is possible to generate a reply in advance before the user's words are finished, that is, before the user issues the ending "please."

Again in FIG. 5, at step 550, the voice synthesizer 440 may synthesize the response generated in step 540 as a voice signal and then send it to the electronic device 110 via the gateway 401 when the user's words are finished. .. As an example, the voice synthesizer 440 may determine that the user's words have ended when the endpoint (end point) is detected from the user's spoken voice, and at this time, the state in which the spoken voice is not detected is predetermined. If it is maintained for more than a certain period of time, it may be recognized as an endpoint. The voice synthesizer 440 may synthesize the response generated by the intermediate result as a voice signal in advance before the user's words are finished, and send the reply voice to the electronic device 110 when the user's words are finished.

Therefore, the AI automatic response system 300 grasps the user's intention from the intermediate result before the end of the user's words, generates and synthesizes the response in advance at the time when the user utters the end, and then prepares in advance when the user's words are finished. By providing the prepared response, the response can be provided more quickly.

FIG. 9 is a flowchart showing another example of the AI automatic response method according to the embodiment of the present invention. The following AI automatic response method may be included in step 550 described above.

At step 901, the dialog manager 420 may continually check whether the user's spoken voice is received while the response voice is being sent from the speech synthesizer 440 to the electronic device 110.

In step 902, the voice synthesizer 440 may interrupt the transmission of the response voice if the user's spoken voice is received while the response voice is being transmitted.

The AI automatic response system 300 may basically include a structure in which the receiving end (RX) and the transmitting end (TX) coexist as the gateway 401, and at this time, the gateway 401 receives the voice spoken by the user. A socket controller may be included to control the transmission of the response voice. The AI automatic response system 300 transmits a voice signal from the transmitting end (TX) if a voice signal based on the user's spoken voice is input to the receiving end (RX) while the response voice is being transmitted from the transmitting end (TX). The signal may be interrupted.

As an example, the voice synthesizer 440 may immediately interrupt the response voice that was being transmitted when the user's spoken voice is input. As another example, the speech synthesizer 440, when the user's spoken voice is input, applies a predetermined length of fade-out to the reply voice being transmitted, and then the reply voice. May be interrupted.

Instead of interrupting the response voice for all spoken voices received during the transmission of the response voice, the reply voice may be selectively interrupted by using a non-verbal filter.

FIG. 10 is a diagram for explaining another example of the process of interrupting the transmission of the response in one embodiment of the present invention.

Referring to FIG. 10, in step 1001, the sensing unit 450 expresses the non-verbal form of the spoken voice received during the transmission of the response voice in order to filter the non-verbal expression as the voice that must be ignored. You may judge whether it is. The sensing unit 450 may determine the non-verbal expression as a continuer that does not have to interrupt the response transmission during the response transmission, and further, the non-verbal expression does not grasp the user's intention. You may judge that it is also good.

FIG. 11 is a diagram for explaining an exceptional situation in which response transmission is not interrupted in one embodiment of the present invention.

For example, referring to FIG. 11, non-verbal expressions such as [yes, yeah, yes, yes, ...] or various exclamations such as [yes, ah, ah, ...] You may classify as linguistic expressions and learn such non-verbal expressions. In addition to this, you may learn a voice that includes emotional expressions as an example of nonverbal expressions, and at this time, emotions are separated by analyzing the waveform of the voice and the characteristics of the end of the word and separating the question mark and the end. You may grasp the rise of. The sensing unit 450 may determine whether the spoken voice received during the transmission of the response voice corresponds to the non-verbal expression based on the learning result of the non-verbal expression.

Again, in FIG. 10, at step 1002, the speech synthesizer 440 determines that the received spoken voice is a voice that should be ignored when it corresponds to a non-verbal expression, and sends the reply voice as it is. If the spoken voice that does not correspond to the non-verbal expression is received, it may be judged that the voice is meaningful and should not be ignored, and the transmission of the reply voice may be interrupted.

Therefore, the AI automatic response system 300 realizes a form in which a person actually talks to a person by interrupting the transmission of the response when the voice signal uttered by the user is received while the response is being transmitted. Can be done.

The dialog manager 420 may provide an appropriate response to the failure interval to support natural conversations such as those with humans in the auto attendant service. As an example, the dialog manager 420 is applicable when the amount of result information for voice recognition is insufficient, for example, when the recognized intention (intent) is insufficient or the information of the slot related to the service is insufficient. You may provide a question-and-answer question to guide the speech about the information.

As another example, if the amount of result information for speech recognition is too large, for example, if two or more intents are recognized at once, the dialog manager 420 prior to replying to clarify the intents. You may ask the user a question to confirm the intent. For example, if the intention of the "recommended menu" and the intention of the "reservation" are grasped at the same time from the user utterance "If there is a recommended menu for the elderly, I would like to reserve a restaurant this Saturday", "First May I ask you a question for user confirmation, such as "Can I make a reservation after guiding you to the recommended menu?"

Also, if the question contains different intents, for example, from a user utterance such as "Do you have a chair for children?", Not only will you answer whether you have a chair, but you will also be asked if you are booking with children. You may ask a question.

As another example, the dialog manager 420 may provide replies in order according to intention when the amount of result information for voice recognition is too large, for example, when two or more intentions are recognized at one time. For example, each intent may be given a response in order by applying numbering (first, second, etc.) according to the order of intent recognition. In response to the user's utterance "If there is a recommended menu for elderly people, I would like to reserve a restaurant this Saturday", "First, we have a Korean set meal course as a recommended menu for elderly customers. Second, what time do you want to make a reservation this Saturday? ”, You may provide the responses in order according to your intention.

User utterance In the case of a user utterance such as "I want to make a reservation, do you have a parking lot?", It can be understood as meaning that if there is a parking lot, it means making a reservation, and if there is no parking lot, the car will not be used. However, it can also be understood as meaning that you want to make a reservation. In the present invention, it is possible to provide a response while considering such various user intentions.

As another example, the dialog manager 420 may process by the amount of information as a response to a failure interval when exactly the same or similar responses are repeated within a unit time or continuously more than a predetermined number of times. .. As yet another example, Dialog Manager 420 may provide a response based on a multi-turn conversation technique that reflects the previous conversation context. In addition to this, by providing appropriate responses to system utterance definition problems, natural language understanding (NLU) errors, etc., it is possible to realize a natural conversation form that actually talks to humans in any situation. Can be done.

Further, the dialog manager 420 may recognize the user's utterance speed by the voice recognition device 410, and may apply different response speeds according to the user's utterance speed. As an example, the dialog manager 420 may determine the utterance speed of the response voice in proportion to the utterance speed of the user. For example, a user who asks a fast question may respond quickly according to the question speed.

Therefore, the AI automatic response system 300 can provide a service suitable for interaction according to the user's utterance speed by providing different response speeds according to the user's utterance speed.

Further, the dialog manager 420 may use real-time translation to translate into the language spoken by the user and provide an automatic response service. The dialog manager 420 may include a translator 460 containing at least one language model, or may be configured in an interlocking manner, and may switch to a language model corresponding to the user's utterance language to provide an automatic response service by real-time translation. You can. As an example, the translator 460 may maintain a Korean model, an English model, and a Japanese model, and the sensing unit 450 may be used as a wakeup word for setting a language from among the spoken voices of the user. It may sense the language of the first part. At this time, the dialog manager 420 may recognize the user's utterance language based on the wakeup word, and may branch to the language model corresponding to the user's utterance language to provide the automatic response service. For example, the dialog manager 420 is the first word of the user's utterance (outside 1).

If is recognized, it is a Korean model, if "hello (English corresponding to)" is recognized, it is an English model, and if "Hello" is recognized, it is a Japanese model, and the language model of the translator 460 is branched. And provide an automatic response service. In other words, Dialog Manager 420 senses the language of the first utterance, even when a foreigner is connected to the phone, and uses that language model to automatically respond to restaurant inquiries, reservations, delivery orders, etc. Can provide services.

Therefore, the AI automatic response system 300 enhances accessibility and convenience to the service without being restricted by the language by translating into the language spoken by the user using real-time translation and providing the response. be able to.

The devices described above may be implemented by hardware components, software components, and / or combinations of hardware components and software components. For example, the apparatus and components described in the embodiments include a processor, a controller, an ALU (arithmetic logic unit), a digital signal processor, a microcomputer, an FPGA (field program gate array), a PLU (programmable log unit), a microprocessor, and the like. Alternatively, it may be implemented using one or more general purpose computers or special purpose computers, such as various devices capable of executing and responding to instructions. The processing device may execute an operating system (OS) and one or more software applications running on the OS. The processing device may also respond to the execution of the software, access the data, and record, manipulate, process, and generate the data. For convenience of understanding, one processor may be described as being used, but those skilled in the art may appreciate that the processor may include multiple processing elements and / or multiple types of processing elements. You can understand. For example, a processor may include multiple processors or one processor and one controller. Other processing configurations, such as parallel processors, are also possible.

The software may include computer programs, code, instructions, or a combination of one or more of these, configuring the processing equipment to operate at will, or instructing the processing equipment independently or collectively. You may do it. The software and / or data is embodied in any type of machine, component, physical device, computer recording medium or device to be interpreted based on the processing device or to provide instructions or data to the processing device. Good. The software is distributed on a computer system connected by a network and may be recorded or executed in a distributed state. The software and data may be recorded 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 by various computer means and recorded on a computer-readable medium. Here, the medium may be one that continuously records a computer-executable program, or one that temporarily records it for execution or download. In addition, the medium may be a variety of recording or storage means in the form of a combination of single or multiple hardware, and is not limited to a medium directly connected to a computer system, but is distributed over a network. It may exist. Examples of media include hard disks, floppy (registered trademark) disks, and magnetic media such as magnetic tape, optical media such as CD-ROMs and DVDs, optomagnetic media such as floptic discs, and It may include a ROM, a RAM, a flash memory, and the like, and may be configured to record program instructions. In addition, other examples of media include recording media or storage media managed by application stores that distribute applications, sites that supply or distribute various other software, servers, and the like.

As described above, the embodiments have been described based on the limited embodiments and drawings, but those skilled in the art will be able to make various modifications and modifications from the above description. For example, the techniques described may be performed in a different order than the methods described, and / or components such as the systems, structures, devices, circuits described may be in a form different from the methods described. Appropriate results can be achieved even if they are combined or combined, or confronted or replaced by other components or equivalents.

Therefore, even different embodiments belong to the attached claims as long as they are equivalent to the claims.

300: AI automatic response system 401: Gateway 410: Speech recognizer 420: Dialog manager 430: Response generator 440: Speech synthesizer 450: Sensing unit 460: Translator

Claims (16)

An artificial intelligence-based automatic response method executed by a computer system.
The computer system includes at least one processor configured to execute computer-readable instructions contained in memory.
The automatic response method of the artificial intelligence base is
The stage of receiving the spoken voice of the user by connecting the user with the call by the at least one processor.
This includes a step of generating a response after analyzing the intention by using the intermediate result of the spoken voice by the at least one processor, and a step of sending the response to the spoken voice by the at least one processor. ,
An automatic response method based on artificial intelligence. The generation stage is
This includes a step of extracting a sample sentence from the conversation log of the automatic response service and a step of analyzing the intention from the intermediate result by conversation learning using the sentence excluding the ending from the sample sentence as learning data.
The automatic response method of the artificial intelligence base according to claim 1. The generation stage is
Including the step of analyzing the intention in syllable units acquired as the intermediate result.
The automatic response method of the artificial intelligence base according to claim 1. The generation stage is
Including the step of determining the generation time of the response based on the confidence in the intention analysis result using the intermediate result.
The automatic response method of the artificial intelligence base according to claim 1. The generation stage is
The step of generating the response in advance is included when the confidence for the intention analysis result using the intermediate result reaches the threshold value determined by the conversation learning using the sentence excluding the ending.
The automatic response method of the artificial intelligence base according to claim 1. The sending stage is
When the endpoint is detected from the spoken voice, the response is transmitted as a voice signal.
The automatic response method of the artificial intelligence base according to claim 1. An artificial intelligence-based automatic response method executed by a computer system.
The computer system includes at least one processor configured to execute computer-readable instructions contained in memory.
The automatic response method of the artificial intelligence base is
The stage of receiving the utterance voice of the user by connecting the user's call by the at least one processor.
When the at least one processor sends a response to the spoken voice, and the at least one processor receives the user's spoken voice while the reply is being sent, the reply is sent. Including the stage of interrupting
An automatic response method based on artificial intelligence. The interrupting stage is
When the spoken voice of the user is received, the transmission of the reply is interrupted immediately or after applying the fade-out.
The automatic response method of the artificial intelligence base according to claim 7. The interrupting stage is
The stage of classifying and learning non-verbal expressions, and if the spoken voice received during the transmission of the response corresponds to the learned non-verbal expression, the transmission of the response is maintained, and the above. Including the step of interrupting the sending of the reply when it does not correspond to the learned non-verbal expression.
The automatic response method of the artificial intelligence base according to claim 7. The automatic response method of the artificial intelligence base is
Further comprising providing a question-and-answer question to guide the information if the at least one processor lacks the information required for the response as a result of recognizing the spoken voice.
The method for automatically responding to the artificial intelligence base according to claim 1 or 7. The automatic response method of the artificial intelligence base is
When a plurality of intentions are recognized by the recognition result of the spoken voice by the at least one processor, a step of providing a response to each intention in order of intention recognition is further included.
The method for automatically responding to the artificial intelligence base according to claim 1 or 7. The automatic response method of the artificial intelligence base is
The step further includes a step of recognizing the utterance speed of the user by the at least one processor, and a step of determining the utterance speed of the response by the utterance speed of the user by the at least one processor.
The method for automatically responding to the artificial intelligence base according to claim 1 or 7. The automatic response method of the artificial intelligence base is
It further includes a step of recognizing the user's spoken language by the at least one processor and a step of switching the language model for the automatic response service to the language model corresponding to the user's spoken language by the at least one processor. ,
The method for automatically responding to the artificial intelligence base according to claim 1 or 7. A non-temporary computer-readable recording medium in which a program for causing a computer to execute the automatic response method of the artificial intelligence infrastructure according to any one of claims 1 to 9 is recorded. It ’s a computer system,
Contains at least one processor configured to execute computer-readable instructions contained in memory.
The at least one processor
The process of receiving the user's spoken voice by connecting with the user,
It processes the process of generating a response after analyzing the intention using the intermediate result of the spoken voice, and the process of sending the response to the spoken voice.
Computer system. It ’s a computer system,
Contains at least one processor configured to execute computer-readable instructions contained in memory.
The at least one processor
The process of receiving the user's spoken voice by connecting with the user,
Processes the process of sending a reply to the spoken voice, and the process of interrupting the sending of the reply if the user's spoken voice is received while the reply is being sent.
Computer system.