KR101927050B1 - User terminal and computer readable recorindg medium including a user adaptive learning model to be tranined with user customized data without accessing a server - Google Patents

Abstract

A user terminal is provided. The user terminal of the present disclosure includes a user input module configured to receive natural language input from a user; and an interactive AI agent module configured to provide an interactive AI agent service based on natural language input - an interactive AI agent service is provided by using a user-customized intent decision model constructed according to a machine learning method-. The user-customized intent decision model is pre-learned in accordance with the machine learning method and is provided in advance to the user terminal in an online or offline manner. The user-customized intent decision model is further configured to be learned and updated based on user customized data for the user obtained from interaction with the interactive AI agent module by the user.

Description

USER TERMINAL AND COMPUTER READABLE RECORDENT MEDIUM INCLUDING A USER ADAPTIVE LEARNING MODEL TO BE TRANINED WITH USER CUSTOMIZED WITH A CUSTOMIZED LEARNING MODEL CONFIGURABLE TO LEARN TO USE WITH PERSONALIZED DATA WITHOUT ACCESS TO THE SERVER DATA WITHOUT ACCESSING A SERVER}

[0001] This disclosure relates to user terminals including a user-tailored learning model associated with machine learning based interactive AI agent systems, and more particularly to a user terminal that includes user personalization data accumulated by interaction with an interactive AI agent system A user terminal including a user-customized learning model configured to learn on a user terminal without access to the server, and the like.

Description of the Related Art [0002] In recent years, with the development of artificial intelligence fields, especially natural language understanding fields, it has become possible to move away from the machine operation according to the conventional machine-centered command input / output method and to allow users to use natural language in a more human-friendly manner such as voice and / Interactive AI agent systems are increasingly being developed and utilized to interactively manipulate machines and obtain desired services from machines. Accordingly, in various fields, including but not limited to, online consulting centers, online shopping malls, and the like, a user can provide desired services through an interactive AI agent system that provides natural language conversations in the form of voice and / or text It is possible to receive it.

Recently, the interactive AI agent system that supports free speech communication based on the learning model based on machine learning has been developed over the conventional interactive AI agent system which only provides a simple dialogue type dialog service based on fixed scenarios The demand is increasing. In order to improve the performance of an interactive AI agent system based on a learning model based on machine learning, it is necessary to continuously learn such a learning model even after the learning model is once constructed. In particular, considering that the individual conversation styles and intent may vary, it is necessary to adaptively learn the learning model to the individual user.

Patent Document 1: Korean Patent Application Publication No. 10-2011-0072847

However, up to now, building and updating (updating according to additional learning) and management of a learning model according to a machine learning method for an interactive AI agent system have all been carried out at a server terminal, And the like are not sensitive and quick and adaptable to each user. Therefore, a learning model for the interactive AI agent system is provided to the user terminal, and on the user terminal, the user corresponding to the terminal is added sensitively, promptly, adaptively, and irrespective of the communication state to the server Learning and updating.

According to one aspect of the disclosure, a user terminal is provided. The user terminal of the present disclosure includes a user input module configured to receive natural language input from a user; And an interactive AI agent module configured to provide an interactive AI agent service based on natural language input, the interactive AI agent service being provided using a user-customized learning model built according to a machine learning method, The model is provided in advance to the user terminal in an online or offline manner in a state pre-learned according to a machine learning method, and the user-customized learning model is a model of a user acquired from the interaction with the interactive AI agent module by the user And is further configured to learn and update based on personalization data.

According to one embodiment of the present disclosure, a customized learning model can be configured to additionally learn and update on the user terminal, based on personalization data about the user, without any external server communication.

In accordance with one embodiment of the present disclosure, learning and updating of a customized learning model is performed on a user terminal at predetermined time intervals based on personalization data regarding the user obtained in accordance with the interaction with the interactive AI agent module by the user Can be repeatedly performed.

According to one embodiment of the present disclosure, personalization data relating to a user obtained from an interaction with a user by an interactive AI agent module includes user's utterance record, whether a recognition / processing failure has occurred regarding user utterance, A cause and / or solution for a processing failure, a response by the interactive AI agent module according to a user's utterance record, and / or a task performance record.

According to one embodiment of the present disclosure, the interactive AI agent service may include determining a user's intent corresponding to natural language input and providing an interactive response corresponding to the determined intent.

According to another aspect of the present disclosure, there is provided a computer-readable recording medium on which a customized learning model is recorded. The computer-readable recording medium of the present disclosure is included in a user terminal, and the user-customized learning model is stored in a computer-readable recording medium in a pre-learned state according to a machine learning method, Wherein the user-customized learning model is used by the interactive AI agent module on the user to provide an interactive AI agent service to the user, wherein the user-customized learning model is based on personalization data about the user obtained from interaction with the interactive AI agent module by the user To be additionally learned and updated.

According to the presently disclosed embodiment, a learning model for an interactive AI agent system is provided to a user terminal, and on a user terminal, a user is alerted to the user corresponding to the terminal in a sensitive, quick and adaptive manner, It can be additionally learned and updated irrelevantly. According to the embodiment of the present disclosure, a learning model on a user terminal can be learned almost in real time by data (for example, user history data) based on an interaction between a user corresponding to the terminal and a corresponding interactive AI agent system And can update. Therefore, according to the embodiment of the present disclosure, unlike the conventional case where the interactive AI agent system developer only updates the learning model for the entire user at the server end, The history and characteristics of the individual users can be reflected in the learning model in almost real time regardless of the timing and the like.

1 is a functional block diagram that schematically illustrates the configuration of a user terminal 100, in accordance with one embodiment of the present disclosure.
FIG. 2 is a functional block diagram that schematically illustrates the functional configuration of the interactive AI agent module 110 of FIG. 1, in accordance with one embodiment of the present disclosure.
FIG. 3 is a functional block diagram schematically illustrating the functional configuration of the dialog / task processing unit 204 of FIG. 2, according to one embodiment of the present disclosure.

Hereinafter, embodiments of the present disclosure will be described in detail with reference to the accompanying drawings. Hereinafter, when it is judged that there is a possibility that the gist of the present disclosure may be unnecessarily blurred, a detailed description of known functions and configurations will be omitted. It is also to be understood that the following description is only an example of the present disclosure, and the present disclosure is not limited thereto.

The terminology used in this disclosure is used only to describe a specific embodiment and is not used to limit the present disclosure. For example, an element expressed in singular < Desc / Clms Page number 5 > terms should be understood as including a plurality of elements unless the context clearly dictates a singular value. It is to be understood that the term " and / or " as used in this disclosure encompasses any and all possible combinations of one or more of the listed items. It should be understood that the terms " comprises " or " having ", etc. used in the present disclosure are intended to specify that there exist features, numbers, steps, operations, elements, It is not intended to exclude the presence or addition of one or more other features, numbers, steps, operations, components, parts, or combinations thereof, by use.

In the embodiments of the present disclosure, 'module' or 'sub-module' means a functional part that performs at least one function or operation, and may be implemented in hardware or software, or a combination of hardware and software. Also, a plurality of "modules" or "sub-modules" may be integrated into at least one software module and implemented by at least one processor, except for "module" or "sub-module" have.

In the embodiment of the present disclosure, the 'interactive AI agent system' is a system in which a user interacts with a user via a natural word input (e.g., a natural language) input from a user through interactive interaction mediated by a natural language of voice and / (E.g., instructions, statements, requests, questions, etc. from the user) to find out the intent of the user and perform the necessary actions based on the found user's intent, i.e., Or any other information processing system capable of providing performance of a task, and is not limited to any particular form. In an embodiment of the present disclosure, the actions performed by the 'interactive AI agent system' may be, for example, an interactive response and / or task performance, respectively, depending on the intent of the user.

In an embodiment of the present disclosure, the interactive response provided by the 'interactive AI agent system' may be in the form of a visual, audible and / or tactile (e.g., voice, sound, text, video, image, symbol, emoticon, hyperlink, Animation, various notices, motion, haptic feedback, and the like), and the like. The tasks performed by the 'interactive AI agent system' in the present embodiment include, for example, searching for information, progressing payment, composing a message, creating an email, dialing, playing music, photographing, (E.g., but not limited to) various types of tasks, including, but not limited to, navigation services, and the like.

In the embodiment of the present disclosure, the 'interactive AI agent system' includes a chatbot system based on the messenger platform, for example, a chatbot system for exchanging messages with a user on a messenger, providing various information desired by the user, or performing a task But it should be understood that the disclosure is not so limited.

In addition, all terms used in the present disclosure, including technical or scientific terms, unless otherwise defined, have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure belongs. It should be understood that commonly used predefined terms are to be interpreted as having a meaning consistent with the contextual meaning of the related art and are not to be interpreted excessively or extensively unless explicitly defined otherwise in this disclosure .

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

Figure 1 is a functional block diagram that schematically illustrates the functional configuration of a user terminal 100, in accordance with one embodiment of the present disclosure. The user terminal 100 includes a user input receiving module 102, a sensor module 104, a program memory module 106, a processing module 108, an interactive AI agent module 110, (112), and a response output module (114).

According to one embodiment of the present disclosure, the user input receiving module 102 may be configured to receive various types of input from a user, such as natural language input (such as voice input and / or text input, and additionally other types of input Can be received. According to one embodiment of the present disclosure, the user input receiving module 102 includes, for example, a microphone and an audio circuit, and can acquire a user voice input signal through a microphone and convert the obtained signal into audio data. According to one embodiment of the present disclosure, the user input receiving module 102 may include various types of input devices such as various pointing devices such as a mouse, joystick, trackball, keyboard, touch panel, touch screen, , And can acquire a text input and / or a touch input signal inputted from a user through these input devices. In accordance with one embodiment of the present disclosure, the user input received at the user input receiving module 102 may be associated with performing certain tasks, such as performing certain applications or retrieving certain information, It is not. According to another embodiment of the present disclosure, the user input received at the user input receiving module 102 may require only a simple conversation response, regardless of any application execution or retrieval of information.

According to one embodiment of the present disclosure, the sensor module 104 includes one or more different types of sensors, and through these sensors, status information of the user terminal 100, such as the physical state of the user terminal 100, Software and / or hardware status, or information about the environmental conditions of the user terminal 100, and the like. According to one embodiment of the present disclosure, the sensor module 104 may include an optical sensor, for example, and may sense the ambient light condition of the user terminal 100 through an optical sensor. According to one embodiment of the present disclosure, the sensor module 104 includes, for example, a movement sensor, and can detect whether the corresponding user terminal 100 is moving through the movement sensor. According to one embodiment of the present disclosure, the sensor module 104 includes, for example, a velocity sensor and a GPS sensor, and through these sensors, the position and / or orientation of the user terminal 100 can be sensed. It should be noted that, according to another embodiment of the present disclosure, the sensor module 104 may include other various types of sensors, including temperature sensors, image sensors, pressure sensors, touch sensors, and the like.

According to one embodiment of the present disclosure, the program memory module 106 may be any storage medium that stores various programs that may be executed on the user terminal 100, such as various application programs and related data, and the like. According to one embodiment of the present disclosure, the program memory module 106 may include a memory device such as a telephone dialer application, an email application, an instant messaging application, a camera application, a music playback application, a video playback application, an image management application, , And data related to the execution of these programs. According to one embodiment of the present disclosure, the program memory module 106 may be configured to include volatile or non-volatile memory in various forms such as DRAM, SRAM, DDR RAM, ROM, magnetic disk, optical disk, .

According to one embodiment of the present disclosure, the processing module 108 may communicate with each component module of the user terminal 100 and perform various operations on the user terminal 100. According to one embodiment of the present disclosure, the processing module 108 can drive and execute various application programs on the program memory module 106. According to one embodiment of the present disclosure, the processing module 108 may receive signals obtained from the user input receiving module 102 and the sensor module 104, if necessary, and perform appropriate processing on these signals have.

According to one embodiment of the present disclosure, the interactive AI agent module 110 may comprise a customized learning model. According to one embodiment of the present disclosure, the customized learning model may be pre-learned in accordance with the machine learning method and provided in advance on the user terminal 102 in an online or offline manner. According to one embodiment of the present disclosure, the interactive AI agent module 110 may continue to further learn and update a customized learning model using personalization data for that user collected on the user terminal 102 . According to one embodiment of the present disclosure, the user personalization data used for further learning and updating of the user-customized learning model includes user input received through the user input receiving module 102 described below, can do. According to one embodiment of the present disclosure, the customized learning model may be further learned and updated by the personalization data accumulated by the interaction with the interactive AI agent module 110 of the user.

In accordance with one embodiment of the present disclosure, the interactive AI agent module 110 receives user input in the form of voice, text and / or touch received via the user input receiving module 102, The received user input may be processed based on a user-customized learning model provided on the user terminal to determine the intent of the user. In accordance with one embodiment of the present disclosure, the interactive AI agent module 110 also generates a specific conversation response that matches the user intent based on a customized learning model and outputs it via the response output module 114, can do. In accordance with one embodiment of the present disclosure, the interactive AI agent module 110 may also cause a predetermined task corresponding to the determined user intent to be performed on the user terminal 100. [

In accordance with one embodiment of the present disclosure, the interactive response generated by the interactive AI agent module 110 may include other visual elements such as images, video, symbols, emoticons, etc., along with the natural language responses of the speech and / Or other auditory elements such as sound, or other tactile elements. According to one embodiment of the present disclosure, an interactive AI agent module 110 may be provided that has the same type (e.g., A response may be generated (e.g., a voice response is generated if a voice input is given and a text response is generated if a text input is given), but the invention is not so limited. It should be appreciated that, in accordance with another embodiment of the present disclosure, responses in the form of voice and / or text may be generated and provided regardless of the type of user input.

According to one embodiment of the present disclosure, the interactive AI agent module 110 may communicate with a predetermined external service server (not shown) via the communication module 112 described below. The external service server may be, for example, a server providing an interactive AI agent service, an online consultation center server, a financial service server, a legal service server, a medical service server, an online shopping mall server, an information search server, a map service server, And the disclosure is not so limited.

According to one embodiment of the present disclosure, the communication module 112 enables the user terminal 100 to communicate with a predetermined external service server (not shown). According to one embodiment of the present disclosure, the communication module 112 may be configured to allow signals obtained on the user input receiving module 102 and the sensor module 104 to be transmitted to an external service server according to a predetermined protocol, A response signal or various control signals including natural language responses in the form of voice and / or text, and the like, and perform appropriate processing according to a predetermined protocol on the received signals.

According to one embodiment of the present disclosure, the response output module 114 may output a response corresponding to a user input in various forms, such as a visual, audible, and / or tactile sense. According to one embodiment of the present disclosure, the response output module 114 includes various display devices, such as touch screens, based on techniques such as LCD, LED, OLED, QLED, etc., Such as text, symbols, video, images, hyperlinks, animations, various notices, etc., to the user. According to one embodiment of the present disclosure, the response output module 114 may include, for example, a speaker or a headset, and may provide an audible response, e.g., voice and / or acoustic response, can do. According to one embodiment of the present disclosure, the response output module 114 includes a motion / haptic feedback generator, through which a tactile response, e.g., motion / haptic feedback, can be provided to the user. It should be appreciated that, in accordance with one embodiment of the present disclosure, the response output module 114 may concurrently provide any two or more combinations of text responses, voice responses, and motion / haptic feedback corresponding to user input.

According to one embodiment of the present disclosure, the user terminal 100 may be any user electronic device having wired or wireless communication capabilities. The user terminal 100 may be a variety of wired or wireless communication terminals including, for example, a smart phone, a tablet PC, a music player, a smart speaker, a desktop, a laptop, a PDA, a game console, a digital TV, a set- It should be noted that it is not limited.

FIG. 2 is a functional block diagram that schematically illustrates the functional configuration of the interactive AI agent module 110 of FIG. 1, in accordance with one embodiment of the present disclosure. The interactive AI agent module 110 includes a user-customized learning model 202, a customized learning model update section 204, and a dialog / task processing section 208.

According to one embodiment of the present disclosure, the user-customized learning model 202 may be configured such that the interactive AI agent module 110 processes the input from the user to determine the user intent, as described above with reference to Figure 1 And a dialogue response corresponding to the intent. As described above, the user-customized learning model 202 may be previously provided on the user terminal 102 in an online or offline manner, pre-learned in accordance with a machine learning method by predetermined data. According to one embodiment of the present disclosure, the user-customized learning model 202 can be continually further learned and updated by the user-customized learning model updating section 204 described later.

According to one embodiment of the present disclosure, the user-customized learning model update unit 204 updates the user personalization data related to the user of the user terminal 100, for example, through the user terminal 100, It is possible to continuously collect data generated as the interaction with the mobile terminal 110 proceeds. According to one embodiment of the present disclosure, the user-customized learning model updating unit 204 updates the user-customized learning model updating unit 204 based on, for example, a user's utterance record, whether a recognition / processing failure has occurred with respect to user utterance, Such as (but not limited to) responses and / or task performance records by the interactive AI agent module 110 according to the utterance record as personalization data about the user. According to one embodiment of the present disclosure, the user-customized learning model updating unit 204 further learns the user learning model 202 using the user personalization data continuously collected, ) Can be continuously updated. According to one embodiment of the present disclosure, the user-customized learning model update unit 204 may further learn and update the user learning model 202 at predetermined time intervals in accordance with the collection of the above-mentioned user personalization data. According to one embodiment of the present disclosure, the collection of user personalization data of the user-customized learning model update unit 204 and the additional learning and updating of the user-customized learning model 202 based thereon are performed on the user terminal 100 ≪ / RTI > without the aid of a < / RTI >

According to one embodiment of the present disclosure, the dialog / task processing unit 206 receives the user natural language input, processes the received user natural language input based on the user-customized learning model 202, The intent of the corresponding user can be determined. According to one embodiment of the present disclosure, the dialog / task processing unit 206 may also provide an action consistent with the determined user ' s tent, e.g., an appropriate conversation response and / or task performance.

FIG. 3 is a functional block diagram that schematically illustrates the functional configuration of the dialog / task processing unit 208 of FIG. 2, according to one embodiment of the present disclosure. Task processing unit 204 includes a Speech-To-Text (STT) module 302, a Natural Language Understanding (NLU) module 304, a dialogue management module 306, a dialog generation module 308, and a text-to-speech (TTS) module 310.

According to one embodiment of the present disclosure, the STT module 302 receives speech input from user inputs received via the user input receiving module 102 and converts the received speech input into text data based on pattern matching or the like can do. According to one embodiment of the present disclosure, the STT module 302 may extract a feature from a user's speech input to generate a feature vector sequence. According to one embodiment of the present disclosure, the STT module 302 may be implemented as a dynamic time warping (DTW) method, a HMM model (Hidden Markov Model), a GMM model (Gaussian-Mixture Mode), a deep neural network model, For example, a sequence of words, based on various statistical models of the speech recognition results.

According to one embodiment of the present disclosure, the NLU module 304 may receive text input from the user input receiving module 102 or the STT module 302. In accordance with one embodiment of the present disclosure, the textual input received at the NLU module 304 may be, for example, a textual input received from the user input receiving module 102 or a textual < RTI ID = A result, e.g., a sequence of words. According to one embodiment of the present disclosure, the NLU module 304 is configured to receive status information associated with the user input, such as the status information of the user terminal 100 at the time of the user input, And the like. As described above, the status information may include various status information (e.g., the physical state of the user terminal 100, the software status of the user terminal 100) related to the user terminal 100 at the time of inputting user voice and / And / or hardware state, environmental state information around the user terminal 100, etc.).

According to one embodiment of the present disclosure, the NLU module 304 may map the received text input to one or more user tents based on the user-customized learning model 202 of FIG. Where the user intent can be associated with a series of operations (s) that can be understood and performed by the interactive AI agent module 110 according to the user's tent. According to one embodiment of the present disclosure, the NLU module 304 may refer to the status information described above in mapping the received text input to one or more user intents.

According to one embodiment of the present disclosure, the conversation management module 306 may generate a corresponding series of operational flows according to the user < Desc / Clms Page number 10 > tent determined by the NLU module 304. [ In accordance with one embodiment of the present disclosure, the dialogue management module 306 may perform any action, e.g., any action corresponding to the user tent received from the NLU module 304, based on the customized learning model 202 of FIG. Determine whether to perform conversation responses and / or task performances, and generate a corresponding detailed action flow.

According to one embodiment of the present disclosure, the dialog generation module 308 may generate the necessary dialog response based on the operation flow generated by the dialogue management module 306. [ According to one embodiment of the present disclosure, the dialog generation module 308 may generate an appropriate dialog response based on the customized learning model 202 of FIG.

According to one embodiment of the present disclosure, the TTS module 310 may receive the conversation response generated by the conversation creation module 310. The interactive response received at the TTS module 310 may be a natural word or a sequence of words having a textual form. According to one embodiment of the present disclosure, the TTS module 310 may convert the input of the received text form into a speech form, according to various types of algorithms.

In the embodiment of the present disclosure described above with reference to Figures 1-3, all functional components of the interactive AI agent system are described as being implemented on the user terminal 100, but the present disclosure is not so limited. According to another embodiment of the present disclosure, each functional component of the interactive AI agent system may be implemented and distributed between the user terminal 100 and an external interactive AI agent server (not shown) via a communication network . In a case where the interactive AI agent system distributes the functions between the user terminal and the server according to an embodiment of the present disclosure, the distribution of each function of the interactive AI agent system between the user terminal and the server differs depending on the embodiment It should be noted that the present invention can be implemented.

Further, although the embodiments of the present disclosure described above with reference to Figs. 1 to 3 have been described as if a particular module performs certain operations for convenience, the present disclosure is not limited thereby. It should be noted that, in accordance with another embodiment of the present disclosure, the operations described above as performed by any particular module in the above description can be performed by separate and distinct modules, respectively. Further, in the embodiment of the present disclosure described above with reference to Figs. 1 to 3, although the description has been made on the case where the customized learning model on the user terminal is used for the interactive AI agent service, the present disclosure is not limited thereto. It should be noted that according to another embodiment of the present disclosure, a customized learning model on a user terminal may be extended to be applied to other machine learning based artificial intelligence services, such as image retrieval and / or synthesis, speech recognition and / do.

As will be appreciated by those skilled in the art, the present invention is not limited to the examples described in this disclosure, but may be variously modified, rearranged and replaced within the scope of the present disclosure. It should be understood that the various techniques described herein may be implemented in hardware or software, or a combination of hardware and software.

A computer program according to an embodiment of the present disclosure may be stored in a storage medium readable by a computer processor or the like, for example, a non-volatile memory such as an EPROM, EEPROM, or flash memory device, a magnetic disk such as an internal hard disk and a removable disk, CDROM disks, and the like. Also, the program code (s) may be implemented in assembly language or machine language. All such modifications and variations that fall within the true spirit and scope of this disclosure are intended to be embraced by the following claims.

Claims (6)

As a user terminal,
A user input module configured to receive natural language input from a user;
A sensor module configured to receive status information regarding the user terminal at the time of inputting the natural language; And
An interactive AI agent module,
Wherein the interactive AI agent module comprises:
A user-customized tent determination model previously provided in a pre-learned state in accordance with a machine learning method and provided in advance to the user terminal in an online or offline manner;
The method comprising: based on the user-customized tent decision model on the user terminal, with the status information on the user terminal received via the sensor module, A dialog and task processing module configured to process the natural language input received from the user to determine a predetermined user tent and to perform conversation and task processing based on the determined user tent; And
Collecting personalization data related to the user obtained from an interaction with the interactive AI agent module by the user without communicating with any external server by the user terminal, and using the collected personalization data, And an update module of the user-customized tent decision model configured to learn and update an additional machine learning tent decision model,
Wherein the personalization data includes at least one of a record of the user's utterance, a cause or remedy at the occurrence of the recognition or processing failure, and a determination of whether the recognition or processing failure occurs on the interactive AI agent module of the user utterance, A response by the interactive AI agent module according to the recording of the user's utterance or a record of task performance,
User terminal. delete delete delete delete delete

Images