KR20200073624A - System For Providing Service Of Artificial Intelligence Conversation - Google Patents

Abstract

The present invention relates to a system for providing an artificial intelligence (AI) conversation service to realize a conversation between an AI conversation service device and a user. According to the present invention, the system comprises: the AI conversation service device providing content to a device user through one or more additional modules electrically connected to a terminal and digitizing a speech of the device user inputted or outputted through a microphone and a speaker; and an AI conversation service providing server receiving the digitized speech of the device user from the AI conversation service device to analyze a corpus of the speech in accordance with types, automatically generating a sentence corresponding to context from the analyzed corpus, which is speech data of the device user, through a machine learning algorithm, and providing the generated sentence to the AI conversation service device.

Description

System for Providing Artificial Intelligence Conversation Service

The present invention relates to an artificial intelligence conversation service providing system.

Recently, as various artificial intelligence services have been released, artificial intelligence devices, which are devices for using artificial intelligence conversation services, have also been launched in various ways.

In general, in the case of implementing a function of a conversation or voice in a device capable of an AI conversation service, it is configured to play a voice simply in a predetermined order or recognize a voice set with a simple command and execute a service when a predetermined condition is met.

Accordingly, interaction between the device and the user of the device is impossible, and it is limited to a simple speaking function, so that there is a problem that the user, especially a child or an elderly person, cannot cause further interest.

Against this background, an object of the present invention is to provide an artificial intelligence conversation service providing system in which a user purchases content and maintains a natural conversation in response to a user's speech after viewing the content.

In order to achieve the above object, in one aspect, the present invention, an AI conversation service providing system that enables a user and a device to perform AI conversation through an AI conversation service device, the electronic communication with one or more additional modules Having one or more terminals that can be connected or detached, content can be provided to a device user through one or more additional modules electrically connected to the terminal, and data of a device user input or output in the device through a microphone and speaker is dataized, The AI conversation service device capable of wireless communication with the AI conversation service providing server through the wireless Internet network and the voice of the device user who has been data from the AI conversation service device receive morpheme analysis, keyword recognition, dialogue act analysis, or semantic analysis. Analyzes the corpus of speech through one or more of the same multiple natural language understanding technologies, and analyzes the corpus, i.e., the device user's speech data, through Ridge Regression, Logistic Regression, The AI conversation service device and the device user can use the machine learning algorithms such as Generalized Linear Model, Random Forest, GradientBoostinf Model and Neural Network. AI conversation including an AI conversation service providing server that automatically generates context-sensitive sentences to enable conversation and provides them to an AI conversation service device to enable conversation between the AI conversation service device and a device user Provide a service providing system.

As described above, according to the present invention, the AI conversation service device of the present invention and the user of the device enable conversation between users, thereby retaining the effect of inducing the user's interest.

In addition, according to the present invention, a voice conversation between a device and the user of the device is performed based on content, so that the device has an effect of easily grasping the opinions and emotions of the device user.

In addition, according to the present invention, it has the effect that the child or the elderly can easily perform a conversation with the device without any special usage.

1 is a diagram briefly showing the configuration of an artificial intelligence conversation service providing system according to an embodiment of the present invention;
2 is a diagram briefly showing the configuration of an apparatus for providing AI conversation service according to an embodiment of the present invention;
3 is a diagram briefly showing the configuration of an artificial intelligence dialogue unit according to an embodiment of the present invention;
4 is an exemplary view showing the appearance of an apparatus for providing artificial intelligence conversation service according to an embodiment of the present invention;
5 is a flowchart illustrating a process of providing an AI conversation service in an apparatus for providing an AI conversation service according to an embodiment of the present invention.

The present invention can be applied to various changes and can have various embodiments, and specific embodiments will be illustrated in the drawings and described in detail.

However, this is not intended to limit the present invention to specific embodiments, and should be understood to include all modifications, equivalents, and substitutes included in the spirit and scope of the present invention.

The terms used in the present application are only used to describe specific embodiments, and are not intended to limit the present invention. Singular expressions include plural expressions unless the context clearly indicates otherwise. In this application, terms such as “include” or “have” are intended to indicate that a feature, number, step, operation, component, part, or combination thereof described in the specification exists, one or more other features. It should be understood that the existence or addition possibilities of fields or numbers, steps, operations, components, parts or combinations thereof are not excluded in advance.

Unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by a person skilled in the art to which the present invention pertains. Terms such as those defined in a commonly used dictionary should be interpreted as having meanings consistent with meanings in the context of related technologies, and should not be interpreted as ideal or excessively formal meanings unless explicitly defined in the present application. Does not.

In addition, it should be understood that all detailed descriptions that list the principles, aspects, and embodiments of the invention as well as specific embodiments are intended to include structural and functional equivalents of these matters. It should also be understood that these equivalents include all currently invented equivalents as well as equivalents to be developed in the future, ie, all devices invented to perform the same function regardless of structure.

Thus, for example, it should be understood that the block diagrams herein represent conceptual views of exemplary circuits embodying the principles of the invention. Similarly, all flow charts, state transition diagrams, pseudo-codes, and the like can represent a variety of processes performed by a computer or processor, whether or not the computer or processor is explicitly shown in a substantially readable form on a computer-readable medium. It should be understood as indicating.

In addition, the term'terminal' used in the present invention is a term used for convenience of description, and may be expressed as'device' or'server' and is replaceable.

Also, the clear use of terms presented in terms of processors, controls, or similar concepts should not be interpreted exclusively by reference to hardware capable of executing software, and without limitation, digital signal processor (DSP) hardware, ROM for storing software. It should be understood as implicitly including (ROM), RAM and non-volatile memory. Other hardware for governors may also be included.

The above objects, features, and advantages will become more apparent through the following detailed description in connection with the accompanying drawings, and accordingly, those skilled in the art to which the present invention pertains can easily implement the technical spirit of the present invention. There will be. In addition, in the description of the present invention, when it is determined that the detailed description of the known technology related to the present invention may unnecessarily obscure the subject matter of the present invention, the detailed description will be omitted.

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

1 is a diagram briefly showing the configuration of an artificial intelligence conversation service providing system according to an embodiment of the present invention.

According to FIG. 1, an artificial intelligence conversation service providing system according to an embodiment of the present invention includes an artificial intelligence conversation service device 100, a wireless Internet network 200, and an artificial intelligence conversation service providing server 300.

The AI conversation service device 100 transmits a voice signal of a device user to the AI conversation service providing server 300 through data transmission and reception with the AI conversation service providing server 300 through the wireless Internet network 120, The AI conversation service providing server 300 may receive content data related to the AI conversation service corresponding to a voice signal.

In addition, the AI conversation service device 100, when the device user listens to or views specific content provided by the AI conversation service providing server 300, answers the questions and answers associated with the content to the AI conversation service providing server ( 300) may be designed to enable communication between the device and the user of the device.

In addition, the AI conversation service device 100 may include one or more additional modules for video output and video input related to the AI conversation service, that is, components related to the AI conversation service are coupled to the AI conversation service device 100 or Can be separated.

To this end, the AI conversation service device 100 includes a sound input terminal unit 110, a sound output terminal unit 120, one or more module connection terminal units 130, a wireless communication unit 140, a control unit 150, and a power unit unit as shown in FIG. It may include 160.

The sound input terminal unit 110 is electrically connected to or disconnected from the microphone module 510 that receives a device user's voice. Here, the device user means a user who uses the module-coupled artificial device 100.

In an embodiment of the present invention, the sound input terminal unit 110 receives a voice signal for voice from a microphone module 510 that receives a voice generated by a device user to use an artificial intelligence conversation service. Here, the voice signal may be an analog signal, and the sound input terminal unit 110 may be implemented as an auxiliary (AUX) terminal.

The sound output terminal unit 120 is electrically connected to or disconnected from the speaker module 520 that outputs an audible sound audible to a device user.

In an embodiment of the present invention, the sound output terminal unit 120 may receive the sound signal transmitted from the control unit 150 and transmit it to the speaker module 520.

In particular, the sound output terminal unit 120 may receive an unconnected guidance voice signal from the controller 150 and transmit it to the speaker module 520. Here, the acoustic signal including the module unconnected guidance voice signal may be an analog signal, and the sound output terminal unit 120 may be implemented as an AUX terminal like the sound input terminal unit 110.

Also, the sound output terminal unit 120 may receive the voice signal of the content downloaded from the artificial intelligence conversation service providing server 300 stored in the removable memory device 560, which will be described later, and deliver it to the speaker module 520. Detailed description thereof will be described later in FIG. 3.

The one or more module connection terminal units 130 are electrically connected or disconnected from one or more additional modules for image output, image input, and the like related to artificial intelligence services.

Here, the one or more module connection terminal units 130 may include first module connection terminal units 132 to Nth module connection terminal units 136.

For example, the first module connection terminal 132 may be electrically connected to or disconnected from the display module 530 displaying an image corresponding to image data on the screen as shown in FIG. 4. The first module connection terminal 132 may be a universal serial bus (USB) terminal, a mini USB terminal, a micro USB terminal, or a high-definition multimedia interface (HDMI) terminal.

The second module connection terminal unit 134 may be electrically connected or disconnected from the camera module 540 that captures an object and generates photographing data as illustrated in FIG. 4. The second module connection terminal 134 may be a USB terminal, a mini USB terminal, a micro USB terminal, or a 3.5 pie input terminal.

In addition, the third module connection terminal unit (not shown) may be electrically connected or disconnected from the beam projector module 550 that projects an image corresponding to image data on a screen as shown in FIG. 4. The third module connection terminal (not shown) may be a universal serial bus (USB) terminal, a mini USB terminal, a micro USB terminal, or a high-definition multimedia interface (HDMI) terminal.

The fourth module connection terminal unit (not shown) may be electrically connected to or disconnected from the removable storage device 560 storing content downloaded from the artificial intelligence conversation service providing server 300 as shown in FIG. 4. The fourth module connection terminal (not shown) may be a Universal Serial Bus (USB) terminal, a mini USB terminal, or a micro USB terminal.

The fourth module connection terminal unit (not shown) includes a speaker module 520 connected to one or more module connection terminal units 130 of a corresponding audio signal and image data of contents stored on the removable storage device 560, and a display module 530 and the beam projector module 550 to output audio on the speaker or to output an image on the screen.

At this time, the content is an e-book in the form of an e-book containing fairy tales, novels, poetry, humanities, essays, or music including K-POP, nursery rhymes, classical music, OST, etc., or on a screen such as a movie, drama, animation, etc. It may be composed of a moving image, it can be downloaded from the artificial intelligence conversation service providing server 300 through the voice command of the device user.

Detailed description of this will be described later in the artificial intelligence conversation service providing server 300 of FIG. 3.

The wireless communication unit 140 performs data transmission and reception with the artificial intelligence conversation service providing server 300 through the wireless Internet network 200.

The wireless communication unit 140 may access the wireless Internet network 200 through Bluetooth communication with a Wi-Fi communication module or a tethering terminal that can access the wireless Internet network 120 through an instrument among Wi-Fi (Wireless Fidelity). It can be implemented in any one of the Bluetooth (Bluetooth) module.

The control unit 150 is a module that controls the overall function of the AI conversation service device 100.

In an embodiment of the present invention, when any one of the one or more additional modules is electrically connected to any one of the one or more module connection terminal units 130, the control unit 150 is connected to any one of the additional modules The module connection notification information for notifying the AI conversation service providing server 300 is generated, and the module connection notification information is transmitted to the AI conversation service providing server 300 through the wireless communication unit 140. Here, the control unit 150 may receive any one of the additional module identification information from any one of the additional modules connected to any one of the module connection terminal units, and thereby generate the module connection notification information. In other words, the module connection notification information may include identification information (eg, serial number, model name, etc.) of any one additional module.

After transmitting the module connection notification information to the artificial intelligence conversation service providing server 300, the control unit 150 may receive the voice signal of the device user through the sound input terminal unit 110 electrically connected to the microphone module 510. The audio signal can be converted into audio data that is digital data.

The controller 150 may transmit voice data to the artificial intelligence conversation service providing server 300 through the wireless communication unit 140. Here, the voice signal is an analog signal corresponding to the voice input by the device user to the microphone module 510, and the microphone module 510 may convert the device user's voice into a voice signal.

In the embodiment of the present invention, if the voice data includes a command related to the operation of any one of the additional modules, the control unit 150 may transmit any one of the additional modules from the artificial intelligence conversation service providing server 300 through the wireless communication unit 140. The operation-related data may be received, and the operation-related data may be transmitted to any one additional module through any one module connection terminal.

Meanwhile, when the voice data includes an operation related command of an unconnected additional module, which is an additional module that is not connected to one or more module connection terminal units 130 in the voice data, the controller 150 communicates artificial intelligence through the wireless communication unit 140. The module unconnected guide voice data may be received from the service providing server 300, and the module unconnected guide voice data may be converted into an analog signal module unconnected guide voice signal.

Then, the control unit 150 transmits the module unconnected guidance voice signal to the speaker module 520 through the sound output terminal unit 120 so that the module unconnected guidance voice data is output as an audible sound through the speaker module 520, This allows the device user to recognize that the wrong voice was delivered.

For example, when the display module 530 is electrically connected to the first module connection terminal portion 132, that is, when the connector of the display module 530 is connected to the first module connection terminal portion 132, the control unit 150 The module connection notification information for notifying the AI conversation service providing server 130 that the display module 530 is connected is generated, and the module connection notification information is transmitted to the AI conversation service providing server 300 through the wireless communication unit 140. Can transmit.

Subsequently, the controller 150 may receive a voice signal corresponding to the voice of the device user through the sound input terminal unit 110 and convert it into voice data, and provide artificial intelligence conversation service through the wireless communication unit 140 It can be transmitted to the server 300.

Here, when the voice data includes a command related to the operation of the display module 530 (for example, “show a △△ cartoon movie”, etc.), the AI communication service device 100 is received through reception of module connection notification information. The AI conversation service providing server 300 recognizing that the display module 530 is combined can confirm that the operation-related command of the display module 530 included in the voice data is a valid command, and the The motion-related data (eg, “△△ cartoon movie” video data) of the display module 530 corresponding to the motion-related command may be searched and extracted from the internal database and provided to the AI conversation service device 100. .

In this case, when the video data (ie, content) of the “Δ△ cartoon movie”, for example, is not searched and extracted from the internal database, the control unit 150 may perform a “no corresponding video” through the sound output terminal unit 120. It may be set to output a phrase such as, in this case, the device user can purchase and download the corresponding video from the artificial intelligence conversation service providing server 300 to watch the video.

For example, when a device user receiving a phrase such as “the corresponding video does not exist” generates a new motion-related command (eg, “Please purchase a △△ cartoon movie”, etc.), the controller 150 is artificial. Content corresponding to'△△ cartoon movie' can be downloaded from the intelligent conversation service providing server 300, and it is stored in the removable storage device 260 so that the content can be searched and extracted from the device's internal database. can do.

Through this, the control unit 150 may receive data related to the operation of the display module 530 from the AI communication service providing server 300 through the wireless communication unit 140, and display it through the first module connection terminal unit 132. The operation-related data of the module 530 is transferred to the display module 530 so that the display module 530 can display an image corresponding to the operation-related data.

On the other hand, when the voice data includes an operation-related command (eg, “take a picture”) of the camera module 540 that is not connected to the one or more module connection terminal 130, through reception of module connection notification information The AI conversation service providing server 300, which recognizes that only the display module 530 is coupled to the AI conversation service device 100, indicates that an operation related command of the camera module 540 included in voice data is an invalid command. You can confirm that.

In this case, the AI conversation service providing server 300 may generate module unconnected guidance voice data (eg, “the camera module is not connected”, etc.) and transmit it to the AI conversation service device 100. .

Through this, the control unit 150 receives the module unconnected guidance voice data from the artificial intelligence conversation service providing server 300 through the wireless communication unit 140 to convert the module unconnected guidance voice data into an analog signal module unconnected guidance voice signal. It is possible to convert and transmit the voice signal of the module unconnected to the speaker module 520 through the sound output terminal unit 120 so that the module unconnected guide voice data can be output as an audible sound through the speaker module 520.

The power supply unit 160 supplies power to the artificial intelligence conversation service device 100. The power supply unit 160 may be any one of a rechargeable nickel-cadmium battery, an alkaline battery, a lithium ion battery, and a lithium polymer battery.

As described above, when the AI conversation service apparatus 100 according to an embodiment of the present invention is electrically connected to any one of the one or more additional modules, the AI conversation service is provided to which one of the additional modules is connected. Module connection notification information for notifying the server 300 is generated and transmitted to the artificial intelligence conversation service providing server 300.

Then, the artificial intelligence conversation service device 100 may receive the voice of the device user through the microphone module 510 and transmit it to the artificial intelligence conversation service providing server 300, and the device user voice may include If a motion-related command is included, the motion-related data of any one additional module may be received from the artificial intelligence conversation service providing server 300 and the motion-related data may be delivered to any one of the additional modules.

On the other hand, if the device user's voice does not include an operation related command of any one of the additional modules, the module unconnected guidance voice data can be received from the artificial intelligence conversation service providing server 300, and through the speaker module 520 The module unconnected guidance voice data is output as an audible tone so that the user of the device can recognize that the wrong voice has been delivered.

The wireless Internet network 200 is a means for mediating data transmission and reception between the AI conversation service device 100 and the AI conversation service providing server 300, a wireless LAN network, a WCDMA (Wideband Code Division Multiple Access) network, and an HSDPA (High) Speed Downlink Packet Access (LTE) network, LTE (Long Term Evolution) network, LTE-A (LTE Advanced) network, and may include a 5G mobile communication network.

The AI conversation service providing server 300 receives and recognizes a device user's voice from the AI conversation service device 100, and data (for example, a video) for responding to the device user's voice in its own database or external database. Data, image data, sound data, etc.) can be searched and extracted and transmitted to the AI communication service device 100.

In the embodiment of the present invention, the AI conversation service providing server 300 receives module connection notification information from the AI device 100 and determines the type of any one additional module electrically connected to the AI conversation service device 100. Can be confirmed. Here, the module connection notification information may include identification information (eg, serial number, model name, etc.) of any one of the additional modules.

After receiving the module connection notification information, the AI conversation service providing server 300 receives voice data from the AI conversation service device 100, and can recognize the meaning of voice corresponding to the voice data through the voice recognition function. have.

In this case, the speech recognition function may analyze the corpus of speech according to the form in the corpus analysis unit 330 of the artificial intelligence conversation service providing server 300 of FIG. 3 to be described later to grasp the meaning of the speech.

As a result of grasping the meaning of the voice corresponding to the voice data, if the voice data includes an operation related command of any one of the additional modules, the artificial intelligence conversation service providing server 300 may use any one to respond to the operation related command. The operation-related data (eg, video data, image data, sound data, module driving control data, etc.) of the additional module may be searched and extracted from its own database or an external database, and transmitted to the artificial intelligence conversation service device 100.

Meanwhile, as a result of understanding the meaning of the voice corresponding to the voice data, if the voice data includes an operation related instruction of an unconnected additional module rather than an operation related instruction of any one additional module, the artificial intelligence conversation service providing server 300 Can generate the module unconnected guidance voice data and transmit it to the artificial intelligence conversation service device 100. Here, the unconnected additional module may refer to an additional module that is not connected to the artificial intelligence device 100. In other words, if an operation related command for an additional module different from the type of the additional module identified through the module connection notification information is included in the voice data, the artificial intelligence conversation service providing server 300 generates the module unconnected guidance voice data To the AI conversation service device 100.

As described above, through the module connection notification information received from the AI conversation service providing server 300 from the AI conversation service device 100, the type of the additional module connected to the AI conversation device 100 can be confirmed, and the AI conversation Since the service providing server 300 provides only artificial intelligence services related to an additional module connected to the artificial intelligence conversation device 100, it is possible to provide differentiated artificial intelligence conversation services to device users according to components configured by the device user. .

Hereinafter, the configuration of the artificial intelligence conversation service providing server 300 according to an embodiment of the present invention will be described.

3 is a block diagram schematically showing the configuration of an AI conversation service providing server according to an embodiment of the present invention.

The apparatus for providing an AI conversation according to an embodiment of the present invention, that is, the AI conversation service providing server 300 includes a communication interface unit 310, a data storage unit 320, a corpus analysis unit 330, and an AI conversation generation unit It may be configured to include 340 and the control unit 350.

The communication interface 310 performs data transmission and reception with the artificial intelligence conversation service device 100 connected through the wireless Internet network 200.

Specifically, the communication interface unit 310 receives the module connection notification information and the voice data of the user of the device transmitted from the AI conversation service device 100 through the wireless Internet network 200, and the AI conversation service device 100 ) May transmit the operation related data of any one additional module connected to the module or the voice data of the module not connected to the artificial intelligence conversation service device 100 through the wireless Internet network 200.

The communication interface unit 310 may be implemented with a network interface card or the like.

The data storage unit 320 temporarily stores data processed by the control unit 350 to be described later, or stores a program for performing the function of the AI conversation service providing server 300, and the AI conversation service providing server ( 300).

The data storage unit 320 may include one or more of an internal storage device of the AI conversation service providing server 300 and a storage server separate from the AI conversation service providing server 300.

The corpus analysis unit 330 is one of a number of natural language understanding technologies such as morpheme analysis, keyword recognition, dialogue act analysis, or semantic analysis of the voice of a device user input through the sound input terminal unit 110 of the artificial intelligence conversation service device 100. The voice may be analyzed through one or more and a corresponding service may be delivered to the artificial intelligence conversation service device 100.

For example, when the corpus analysis unit 330 receives voice data of "△△ buy a cartoon movie" from the artificial intelligence conversation service device 100, "△△", "cartoon" through morpheme analysis, "Movie", "Buy me", etc. are extracted.

Next, the corpus analysis unit 330 is a cartoon movie containing the title of "△△" in the storage server separate from the internal storage of the AI conversation service providing server 300 and the AI conversation service providing server 300, etc. And extract it and deliver it to the AI conversation service device 100.

Depending on the design, in the process of extracting the cartoon movie containing the title of "△△", the device provides the user with information on the content price, content summary, and expiration date of the content after the download. ), and when a device user receives a voice signal output, and then transmits a voice command related to the content (for example, "buy content") through the device, the corresponding content is determined to be purchased and artificial intelligence It can be stored in the removable storage device 560 of the conversation service device 100.

In addition, the corpus analysis unit 330, when a device user listens to or views a specific content and then transmits a voice for comments and emotions on the content through the device, analyzes it in the same manner as the above-described method You can do

At this time, the analyzed information may be transmitted to the AI conversation generating unit 340, which will be described later, and may be used to automatically generate context-sensitive sentences so that the device user and the device can communicate smoothly.

For example, when a device user listens to or views a specific content and then transmits voice data such as "The main character is too poor. My heart hurts" through the device, the corpus analysis unit 330 uses the morpheme analysis to "the main character" , "Slight", "Mind", "Sick", etc. are extracted, and "Slight", "Sick", etc. can be categorized into "Sentiment of sadness".

In addition, the corpus analysis unit 330 may determine the emotional state of the device user through utterance sound, which is the voice of the device user, according to the design. It can recognize emotions by extracting the frequency bandwidth, spectral slope, and open ratio of voice gates as parameters of a device user's voice signal, which can be designed in such a way as to extract commonly used voice signal parameters.

The AI conversation generating unit 340 is adapted to the context so that the AI conversation service device 100 and the device user can talk about specific content based on the voice data of the device user analyzed through the corpus analysis unit 330. Performs the function to automatically generate sentences.

At this time, the artificial intelligence conversation generation unit 340 includes a machine learning processing unit (not shown) to automatically generate sentences according to machine learning. The machine learning algorithm in the machine learning processing unit may use one or more processing methods, furthermore, by applying input variables and target variables to a combination of a plurality of machine learning algorithms and a plurality of machine learning algorithms, pre-learning and learning results The optimal machine learning algorithm or the optimal combined machine learning algorithm may be determined in consideration of at least one of accuracy, computation time, performance ratio, and model stability.

The machine learning algorithm includes, for example, Ridge Regression, Logistic Regression, Generalized Linear Model, Random Forest, GradientBoostinf Model and Neural. A network (Neural Network) or the like may be used.

Ridge regression is a basic linear model and provides additional parameters to handle extremes or outliers. Logistic regression is a basic linear model and is effective when the target variable, or the object to be predicted, shows a binomial distribution. The generalized linear model is a basic linear model, and it is effective to use when the target variable shows a Poisson distribution. It is a random forest ensemble model composed of a number of decision trees. Each decision tree considers the relationship between input variables and target variables. Therefore, it is developed individually and has more flexible characteristics than the upper linear model. The stimulus model is also called the Generalized Boosting Model or GBM, and it is often referred to as a model with high accuracy compared to a Random Forest because there is a difference between an ensemble model with a tendency similar to that of a Random Forest, or that each decision tree is opened considering the accuracy of the decision tree that has already been created. It is also considered.

The neural network is a very flexible model that can predict virtually any target and can accommodate both linear and non-linear patterns by adjusting parameters. In particular, neural networks are statistical learning algorithms inspired by the neural networks of biology in machine learning and cognitive science. Deep learning can be implemented by composing such neural networks in a complex way.

For example, the neural network may include overall fully connected (FC), convolutional neural network (CNN) optimized with Max Pooling and convolution, and sequential neural networks (RNN) useful for time series data.

Specifically, the artificial intelligence dialogue generation unit 340 includes Ridge Regression, Logistic Regression, Generalized Linear Model, Random Forest, and GradientBoostinf Model) and a neural network can provide a function for a device user to interact with the AI conversation service device 100 as shown in the following example through a machine learning algorithm.

First, the device user downloads the content for a specific movie through the AI conversation service providing server 300, and after viewing all of them, voices emotions, opinions, and the like felt by the device user to the AI conversation service device 100. Talk through, and the AI conversation service device 100 delivers it to the AI conversation service providing server 300 so that the device can listen to the device user's voice and answer questions or answers. 300) provides the appropriate sentence.

For example, after a device user downloads'Cinderella' from the AI conversation service providing server 300 and views it through the AI conversation service device 100, the device reads, "If Cinderella's glass shoes are different, What would it have been like to be hit by someone?" If said, the device passes this to the AI conversation service providing server 300 to proceed with corpus analysis, and through machine learning, "But the prince remembers Cinderella, so he would have found the real Cinderella" or "Wizard Since it was a glass shoe, wasn't Cinderella right?" To generate the sentence, and transmits it back to the AI conversation service device 100 to perform a function to output to the device user.

At this time, the sentence generated from the artificial intelligence conversation service providing server 300 may be decorated to be associated with the content and the voice delivered by the device user, and according to the design, a sentence such as an expected question and answer related to the content may be embedded in the content itself database Therefore, when a user of a device talks about opinions and emotions, it can perform a function of promptly answering a built-in sentence.

In addition, the AI conversation service providing server 300 allows the device user to store an ongoing conversation in the above-described data storage unit 320, and then the device user views or listens to other content and asks the content related to the previous content. Even if it can be designed to answer appropriately.

For example, after the device user has watched'Cinderella' as the first content, and then watch'Snow White and the seven dwarfs' as the second content, and the device 100 provides AI communication service device 100 "Cinderella hates a lot of stepmother When I said the sentence, "Why are you so sorry for the princesses?", the AI conversation service providing device 100 is connected to the'Cinderella' content previously stored in the AI conversation service providing server 300. It can bring up the conversation history with the user of the Korean device, and answers like, "But Cinderella met the prince by making only one shoe, and Snow White met the prince with the help of the seven dwarves. Can be output.

Although the artificial intelligence conversation generating unit 340 has been described through the above-described examples, it is not limited thereto, and it is needless to say that all sentences that can be communicated with the device user can be generated.

The control unit 350 is a module that controls the overall function of the AI conversation service providing server 300, and may be implemented as a main board including a CPU (Central Processing Unit).

The following is a flow chart showing a process for providing an AI conversation service in an AI conversation service providing server according to an embodiment of the present invention.

First, when the device user requests the AI conversation service device 100 to purchase a specific content, the AI conversation service providing server 300 is a storage separate from the internal storage device or the AI conversation service providing server 300 The content may be extracted from the data storage unit 320 composed of one or more servers and transmitted to the artificial intelligence conversation service device 100, which may receive, download, and store the content.

Subsequently, when a device user views or listens to specific content downloaded to the AI conversation service device 100 and then speaks emotions, opinions, etc., through voice, the AI conversation service device 100 dataizes it to make an AI conversation It transmits to the service providing server 300. After receiving the voice data from the artificial intelligence conversation service providing server 300, the server analyzes the corpus of voice data (S20, S30).

After analyzing the corpus, the AI conversation service providing server 300 grasps the meaning of voice corresponding to the voice data, and generates an AI conversation based on the voice (S50).

For example, after a device user downloads'Cinderella' from the AI conversation service providing server 300 and views it through the AI conversation service device 100, the device reads, "If Cinderella's glass shoes are different, What would it have been like to be hit by someone?" If said, the device passes this to the AI conversation service providing server 300 to proceed with corpus analysis, and through machine learning, "But the prince remembers Cinderella, so he would have found the real Cinderella" or "Wizard Since it was a glass shoe, wasn't Cinderella right?" To generate the sentence, and transmits it back to the AI conversation service device 100 to perform a function to output to the device user.

Through the function of the AI conversation service providing server 300, the device user can communicate with the AI conversation service device 100, and when the device user requests new content, the above-described processes are repeated to accumulate the conversation contents, That is, even if the new content is stored or watched or listened to, the conversation content is set to be continued so that a new sentence can be generated (S60).

For example, after the device user has watched'Cinderella' as the first content, and then watch'Snow White and the seven dwarfs' as the second content, and the device 100 provides AI communication service device 100 "Cinderella hates a lot of stepmother When I said the sentence, "Why are you so sorry for the princesses?", the AI conversation service providing device 100 is connected to the'Cinderella' content previously stored in the AI conversation service providing server 300. It can bring up the conversation history with the user of the Korean device, and answers like, "But Cinderella met the prince by making only one shoe, and Snow White met the prince with the help of the seven dwarves. Can be output.

In the above, it has been described that all the components constituting the embodiment of the present invention are combined and operated as one, but the present invention is not necessarily limited to these embodiments. That is, if it is within the scope of the present invention, all of the components may be selectively combined and operated. The above description is merely illustrative of the technical idea of the present invention, and those skilled in the art to which the present invention pertains may make various modifications and variations without departing from the essential characteristics of the present invention. The scope of protection of the present invention should be interpreted by the claims below, and all technical spirits within the equivalent range should be interpreted as being included in the scope of the present invention.

Claims (5)

In the AI conversation service providing system that enables the user and the device to perform AI conversation through the AI conversation service device,
A device having one or more terminals electrically connected to or detachable from one or more additional modules to provide content to a device user through one or more additional modules electrically connected to the terminals, and devices input or output in a device through a microphone and speakers An artificial intelligence conversation service device capable of wirelessly communicating with an artificial intelligence conversation service providing server through the wireless internet network,
After receiving the voice of the device user who has been data from the artificial intelligence conversation service device, the corpus of speech is analyzed according to the form through one or more of a number of natural language understanding techniques such as morpheme analysis, keyword recognition, speech act analysis, or semantic analysis. , Analyzed corpus, i.e., Ridge Regression, Logistic Regression, Generalized Linear Model, Random Forest, GradientBoostinf Model ) And a neural network such as a machine learning algorithm to automatically generate context-sensitive sentences to enable the AI conversation service device and the device user to talk about the content and provide it to the AI conversation service device. And an artificial intelligence conversation service providing server that functions to enable conversation between the artificial intelligence conversation service device and a device user.
AI conversation service provision system.
According to claim 1,
The AI conversation service providing server,
Based on the voice data of the device user received from the AI conversation service device, the device user's emotions and opinions are identified, and a sentence corresponding thereto is automatically generated.
AI conversation service provision system.
According to claim 1,
The AI conversation service providing server,
Characterized by accumulating and storing the voice data of the device user received from the AI conversation service device and the sentence generated from the AI conversation service provision server.
AI conversation service provision system.
The method of claim 3,
The AI conversation service providing server,
When a device user requests new content, a sentence is generated so that the conversation content continues based on the accumulated voice data of the device user and the sentence history generated from the AI conversation service providing server.
AI conversation service provision system.
According to claim 2,
The AI conversation service providing server,
Characterized by determining the emotional state of the device user through the voice of the device user extracting the frequency bandwidth, spectral slope and the open ratio of the voice gate as parameters as a parameter of the device user's voice signal
AI conversation service provision system.

Images