Electronic device for determining time for which chatbot maintains conversation and operating method thereof

Abstract

An electronic device for determining a session duration according to a response message characteristic in a chatbot and an operating method thereof are provided. The electronic device is configured to determine a default session time based on a difficulty level of a response message, determine an additional session time based on conversation history information of the chatbot before the response message is output, determine a session duration that is a sum of the default session time and the additional session time, and provide the session duration.

Description

TECHNICAL FIELD

[0001] The disclosure relates to an electronic device for determining a length of time for which a chatbot maintains a session in which a response message is provided for a user input message, and an operating method thereof. BACKGROUND

[0002] When a user inputs an inquiry message or asks a question through a voice message in an electronic device, a system for providing an answer to the user's question is used. More specifically, recently, with the development of artificial intelligence and big data technology, intelligent response systems such as chatbots are widely used. A "chatbot" is an abbreviation of a chatting robot, and is an artificial intelligence service configured to perform operations such as providing information on a question or providing a service requested by using a voice signal or text to have a conversation with a user. As a service that interacts through a messenger application, a chatbot is provided through a preset rule or artificial intelligence technology.

[0003] In a session through a chatbot, a session duration is determined according to personal information protection of a user, and the session ends when the session duration elapses. The session duration is updated by the user's input. When the session ends before a question is completely solved as the session duration elapses, in order to return to a previous question solving scenario being in progress, the user has to input a message input at a previous point in time again and be provided with the same response message again, which can be cumbersome and time consuming. In addition, even though a different processing time is required according to a question solving scenario, there is a problem that the session is accidentally terminated due to the time taken by the user to solve the question, since a total session duration is provided.

[0004] Extending the session duration of all question solving scenarios can be considered as a solution. However, this solution can cause additional resources to be used due to an increase in the number of sessions managed at the same time and an increase in the corresponding cost, and there is a high risk of personal information leakage.

[0005] The above information is presented as background information only to assist with an understanding of the present disclosure. No determination has been made, and no assertion is made, as to whether any of the above might be applicable as prior art with regard to the present disclosure. SUMMARY

[0006] [TECHNICAL PROBLEM]

[0007] Aspects of the present disclosure are to address at least the above-mentioned problems and / or disadvantages and to provide at least the advantages described below. Accordingly, an aspect of the present disclosure is to provide an electronic device for providing a conversation duration suitable for a problem-solving scenario by determining a default conversation time and an additional conversation time according to characteristics of a reply message.

[0008] Other aspects will be partially set forth in the description which follows, and, in part, will become apparent to those having ordinary skill in the art by reference to the description, or can be learned by practice of the presented embodiments.

[0009] [Solution]

[0010] According to an aspect of the present disclosure, a method of determining a conversation duration of a chatbot is provided. The method includes outputting a response message for an input message of a user, determining a default conversation time in which a conversation is maintained based on a difficulty level of the outputted response message, determining an additional conversation time based on conversation history information of the chatbot before the response message is outputted, waiting for an additional message of the user to be inputted during a conversation duration which is a sum of the default conversation time and the additional conversation time, and determining an end of the conversation based on whether the additional message is inputted.

[0011] According to an embodiment of the present disclosure, determining the default conversation time includes obtaining a labeled value of the difficulty level by applying the response message as an input to a deep neural network model, and determining a conversation time mapped to correspond to the obtained labeled value as the default conversation time.

[0012] According to an embodiment of the present disclosure, the deep neural network model is trained through supervised learning in which a plurality of training messages are applied as an input and a labeled value of the difficulty level is applied as an output ground truth.

[0013] According to an embodiment of the present disclosure, the conversation history information can include information about at least one of a number of response messages previously outputted in a conversation before the response message is outputted or a conversation usage time in a scenario performed according to the previously outputted response message.

[0014] According to an embodiment of the present disclosure, the additional conversation time can be determined by a preset ratio of a sum of the conversation usage time for each scenario.

[0015] According to an embodiment of the present disclosure, determining the additional conversation time can include calculating a conversation time usage ratio by performing an operation of dividing the conversation usage time for each scenario by the default conversation time, comparing the calculated conversation time usage ratio with a preset threshold, and determining the additional conversation time based on a comparison result.

[0016] According to an embodiment of the disclosure, determining the additional conversation time can include calculating an average of conversation usage times before the response message is output, comparing the calculated average with a preset threshold, and determining the additional conversation time based on a comparison result.

[0017] According to an embodiment of the disclosure, determining the conversation end can include outputting a conversation end notification providing information about the conversation end before the conversation duration elapses.

[0018] According to an embodiment of the disclosure, determining the conversation duration can further include storing conversation usage history information of a plurality of users about the output response message, and adjusting the default conversation time based on the stored conversation usage history information of the plurality of users.

[0019] According to an embodiment of the disclosure, the conversation usage history information can include a plurality of users' statistics about at least one of a number of conversation extension requests at a point in time before the response message is output, a number of scenarios performed before the response message, or a conversation usage time for the response message.

[0020] According to another aspect of the disclosure, an electronic device for determining a conversation duration of a chatbot is provided. The electronic device includes a communication interface configured to perform transmission and reception of data with other devices, a memory, and a processor configured to execute one or more instructions. The processor is further configured to execute one or more instructions to obtain an input message input by a user through the communication interface, output a response message for the input message of the user, determine a default conversation time in which a conversation is maintained based on a difficulty level of the output response message, determine an additional conversation time based on conversation history information of the chatbot before the response message is output, wait for an additional message to be input by the user during a conversation duration that is a sum of the default conversation time and the additional conversation time, and determine an end of the conversation based on whether the additional message is input.

[0021] According to an embodiment of the disclosure, the processor can obtain an output label value of the difficulty level by applying the response message as input to a deep neural network model, and determine a conversation time mapped to correspond to the obtained label value as the default conversation time.

[0022] According to an embodiment of the disclosure, the deep neural network model can be an artificial intelligence model trained through supervised learning in which a plurality of training messages are applied as input and a label value of the difficulty level is applied as an output true value.

[0023] According to an embodiment of the disclosure, the dialogue history information can include information about at least one of a number of response messages previously output in a session before the response message is output, or a session usage time in a scenario executed according to the previously output response message.

[0024] According to an embodiment of the disclosure, the processor can be further configured to execute one or more instructions to determine the additional session time by a preset ratio of a sum of the session usage time for each scenario.

[0025] According to an embodiment of the disclosure, the processor can be further configured to execute one or more instructions to calculate a session time usage ratio by performing an operation of dividing the session usage time for each scenario by a default session time, divide the session usage time for each scenario by the default session time, compare the calculated session time usage ratio with a preset threshold, and determine the additional session time based on a result of the comparison.

[0026] According to an embodiment of the disclosure, the processor can be further configured to execute one or more instructions to output a session end notification providing information about the end of the session before the session duration elapses.

[0027] According to an embodiment of the disclosure, the processor can be further configured to execute one or more instructions to store, in a database of the memory, session usage history information about the output response message for a plurality of users, and adjust the default session time based on the stored session usage history information for the plurality of users.

[0028] According to an embodiment of the disclosure, the session usage history information can include statistics for a plurality of users about at least one of a number of session extension requests at a point in time before the response message is output, a number of scenarios executed before the response message, or a session usage time for the response message.

[0029] According to another aspect of the disclosure, a computer program product is provided. The computer program product includes a computer readable storage medium having stored therein a program to be executed on a computer. BRIEF DESCRIPTION OF DRAWINGS

[0030] The above and other aspects, features and advantages of certain embodiments of the disclosure will be more apparent from the following description taken in conjunction with the accompanying drawings, in which:

[0031] Figure 1 is a conceptual diagram illustrating a method of providing a chatbot service to a client device through a network by an electronic device according to an embodiment of the disclosure;

[0032] Figure 2is a conceptual diagram illustrating a method of determining a session duration, performed by an electronic device, according to an embodiment of the disclosure;

[0033] Figure 3 is a block diagram of components of an electronic device, according to an embodiment of the disclosure;

[0034] Figure 4 is a flowchart of an operating method of an electronic device, according to an embodiment of the disclosure;

[0035] Figure 5a is a diagram illustrating a method of training a deep neural network model for classifying a difficulty level of a response message, performed by an electronic device, according to an embodiment of the disclosure;

[0036] Figure 5b is a diagram illustrating a method of obtaining information about a difficulty level of a response message by using a deep neural network model, performed by an electronic device, according to an embodiment of the disclosure;

[0037] Figure 5c is a diagram illustrating a method of determining a default session time based on a difficulty level, performed by an electronic device, according to an embodiment of the disclosure;

[0038] Figure 6 is a flowchart of an operating method of an electronic device, according to an embodiment of the disclosure;

[0039] Figure 7 is a time table illustrating a relationship between a default session time and a session usage time according to an input message input by a user and a response message provided by a chatbot, respectively, according to an embodiment of the disclosure;

[0040] Figure 8 is a flowchart of a method of determining whether to provide an additional session time, performed by an electronic device, according to an embodiment of the disclosure;

[0041] Figure 9 is a flowchart of a method of determining whether to provide an additional session time, performed by an electronic device, according to an embodiment of the disclosure;

[0042] Figure 10 is a flowchart of a method of determining whether to provide an additional session time, performed by an electronic device, according to an embodiment of the disclosure;

[0043] Figure 11 is a flowchart of a method of outputting a session end notification, performed by an electronic device, according to an embodiment of the disclosure;

[0044] Figure 12 is a flowchart of a method of adjusting a default session time, performed by an electronic device, according to an embodiment of the disclosure; and

[0045] Figure 13 FIG. 1 is a flowchart illustrating a method of extending or shortening a default session time, according to an embodiment of the disclosure.

[0046] In all the drawings, it should be noted that like reference numerals are used to describe the same or similar elements, features, and structures. DETAILED DESCRIPTION

[0047] The following description with reference to the accompanying drawings is provided to assist in a comprehensive understanding of various embodiments of the disclosure as defined by the claims and their equivalents. It includes various specific details to assist in that understanding but these are to be regarded as merely exemplary. Accordingly, those of ordinary skill in the art will recognize that various changes and modifications of the various embodiments described herein can be made without departing from the scope and spirit of the disclosure. In addition, descriptions of well-known functions and constructions can be omitted for clarity and conciseness.

[0048] The terms and words used in the following description and claims are not limited to the bibliographical meanings, but are merely used to enable a clear and consistent understanding of the disclosure. Accordingly, it should be apparent to those skilled in the art that the following description of various embodiments of the present disclosure is provided for illustration purpose only and not for the purpose of limiting the disclosure as defined by the appended claims and their equivalents.

[0049] It should be understood that the singular forms "a," "an," and "the" include plural referents unless the context clearly dictates otherwise. Thus, for example, reference to "a component surface" includes reference to one or more of such surfaces.

[0050] Throughout the disclosure, the expression "at least one of a, b, or c" means only a, only b, only c, both a and b, both a and c, both b and c, all of a, b, and c, or variations thereof.

[0051] Although the terms used in the embodiments of the present disclosure are selected from general terms that are currently and widely used in consideration of functions in the disclosure, these terms can be changed according to the intention of one of ordinary skill in the art, the judicial precedents, or the introduction of new technologies. Also, the applicant can intentionally choose the term, and in this case, the meaning of the term is disclosed in the corresponding description part of the disclosure. Therefore, the terms used in the specification should not be defined by the simple names of the terms, but should be defined by the meanings and the content throughout the disclosure.

[0052] It should be understood that the singular forms "a," "an," and "the" include one or more of the indicated number of entities, unless the relevant context clearly dictates otherwise. As used herein, all terms including technical or scientific terms have the same meaning as commonly understood by one of ordinary skill in the art to which the description pertains.

[0053] Throughout the disclosure, when it is assumed that a certain part includes a certain component, the term "include" means that the corresponding component can further include other components unless a specific meaning contrary to the corresponding component is written. The terms used in the specification, such as "unit" or "module," mean a unit for processing at least one function or operation, and can be implemented in hardware, software, or a combination of hardware and software.

[0054] The expression "configured (or set) to" used in the specification can be replaced with, for example, "adapted to," "have the ability of," "designed to," "made to," or "capable of," depending on the situation. The term "configured (or set) to" does not always mean "designed to" only by hardware implementation. Alternatively, in some cases, the expression "a device configured to" can mean that the device "can" operate with another device or component. For example, the phrase "a processor configured (or set) to perform A, B, and C" can be a dedicated processor (e.g., an embedded processor) for performing the corresponding operations or a general-purpose processor (e.g., a central processing unit (CPU) or an application processor) that can perform the corresponding operations by executing at least one software program stored in a memory.

[0055] In the disclosure, "Chatbot" is an abbreviation of chatting robot, and is an artificial intelligence service configured to perform operations, for example, providing information about a question by using a voice signal or text to have a conversation with a user, or providing a service for a request. As a service that interacts through a messenger application, a Chatbot is provided through a preset rule or an artificial intelligence technology.

[0056] In the disclosure, "conversation" is a set of messages of a user and response messages of a Chatbot that are input through a Chatbot during a certain period.

[0057] In the disclosure, "conversation duration" is a time for which a conversation is maintained without termination. The conversation duration starts from a time at which a response message of a Chatbot to a user input message is output. The conversation duration is updated and initialized when a user input message is input.

[0058] In the disclosure, the "session use time" is the time in which a user uses a session, and refers to a time interval between when a response message of a chatbot is output and when a user inputs a reply message or an additional message to the response message.

[0059] In the disclosure, the "scenario" is a problem solving operation (or process) provided by a chatbot for an input message of a user, such as requesting, inquiring, replying, or providing information through a response message. For example, when the input message is "about an application of a failure", the scenario can refer to an operation in which a chatbot solves a problem through a response message of requesting, inquiring, replying, or providing information until the application of the failure is completed. Reference will be made to Figure 7 The session use time and the scenario are described.

[0060] Figure 1 is a conceptual diagram illustrating a method of providing a chatbot service to a client device through a network, performed by an electronic device, according to an embodiment of the disclosure.

[0061] Reference Figure 1 The electronic device 1000 is a computing device that receives an input message from a user of the client devices 3001, 3002, and 3003 and provides a response message for the input message. According to an embodiment of the disclosure, the electronic device 1000 can be implemented as a chatbot that provides a response message for an input message, but is not limited thereto.

[0062] According to an embodiment of the disclosure, the electronic device 1000 can include a server or a workstation. However, the disclosure is not limited thereto, and the electronic device 1000 can include at least one of, for example, a smart phone, a tablet personal computer (PC), a mobile phone, a video phone, an e-book reader, a desktop PC, a laptop PC, a web book computer, a personal digital assistant (PDA), or a portable multimedia player (PMP).

[0063] The electronic device 1000 can transmit or receive data to or from the client devices 3001, 3002, and 3003 through the network 2000. The network 2000 can be configured not to be limited to a certain communication method, such as wired or wireless communication. The network 2000 can include at least one wired or wireless data communication method, including, for example, Ethernet, wired or wireless local area network (LAN), Wi-Fi, Wi-Fi direct (WFD), and wireless gigabit alliance (WiGig).

[0064] The client devices 3001, 3002, and 3003 are terminals used by a user to receive a chatbot service from the electronic device 1000 through a chatbot application. According to an embodiment of the disclosure, the user can execute the chatbot application through the client devices 3001, 3002, and 3003, input a text message to the chatbot application, and receive a response message from the electronic device 1000. According to an embodiment of the disclosure, the client devices 3001, 3002, and 3003 can receive a voice input from the user, convert the voice input into text and transmit the text to the electronic device 1000, and receive a response message to the text from the electronic device 1000.

[0065] The client devices 3001, 3002, and 3003 can include, for example, a smartphone, a tablet PC, a mobile phone, a video phone, an e-book reader, a desktop PC, a laptop PC, a web book computer, a PDA, or a PMP.

[0066] The electronic device 1000 of the disclosure is not limited to receiving an input message from the client devices 3001, 3002, and 3003 through the network 2000 and providing a response message to the client devices 3001, 3002, and 3003, as Figure 1 shown. According to an embodiment of the disclosure, unlike the embodiment shown in Figure 1 , the electronic device 1000 can directly receive an input message from a user without intervention of the network 2000 and output a response message to the input message.

[0067] Figure 2 is a conceptual diagram illustrating a method of determining a session duration, performed by an electronic device, according to an embodiment of the disclosure.

[0068] Referring to Figure 2 , a chatbot session screen is shown. The electronic device 1000 can receive a first input message 210 "food in the refrigerator is not frozen" from a user and output a first response message 220 for the first input message 210. The first response message 220 can be a message requesting confirmation from the user for the first input message 210. For example, the first response message 220 can be "Have you checked that the power cord of the refrigerator is plugged into the outlet?"

[0069] The electronic device 1000 can initialize a session time when the first response message 220 is output. The electronic device 1000 can wait for a user to input an additional message during a preset default session duration. The user can use the session time due to checking the power cord.

[0070] When the user inputs a second input message 230, the electronic device 1000 can output a second response message 240 for the second input message 230. According to an embodiment of the disclosure, the second input message 230 can include a response of the user for the first response message 220 or an additional question. In Figure 2 In the disclosed embodiment of the disclosure, the second input message 230 can be "Yes, the power cord is plugged in." The second response message 240 can be a message for an additional request, a response, or information provided by the chatbot for the second input message 230. In Figure 2 In the disclosed embodiment of the disclosure, the second response message 240 can be "Have you checked whether the cooling engine of the refrigerator is working?" The electronic device 1000 can initialize the session time when the second response message 240 is output.

[0071] The electronic device 1000 can determine a default session time based on a difficulty level of the second response message 240. The difficulty level of the second response message 240 refers to the degree of difficulty in solving a problem by a response message requesting, inquiring, or providing information. According to an embodiment of the disclosure, the electronic device 1000 can perform training for classifying a response message according to a difficulty level using a deep neural network model, and obtain information about the difficulty level of the second response message 240 through the deep neural network model. The electronic device 1000 can determine a session time mapped to correspond to the difficulty level as the default session time.

[0072] The electronic device 1000 can wait for the user to input an additional message during the default session time. When no additional message of the user is input even though the default session time elapses, the electronic device 1000 can determine whether to provide an additional session time. The electronic device 1000 can determine the additional session time based on the currently output response message (i.e., the conversation history information before the second response message 240 is output). According to an embodiment of the disclosure, the electronic device 1000 can determine the additional session time based on information about at least one of the number of response messages output before the second response message 240 is output, or the session use time used in a scenario performed through the previously output response message.

[0073] In Figure 2 In the disclosed embodiment of the disclosure, the electronic device 1000 can identify the first response message 220 output before the current message (i.e., the second response message 240) is output, obtain information about the session use time used by the user until the second input message 230 is input by the user during the scenario of the first response message 220, and determine the additional session time based on the number of the first response message 220 and the session use time in the previous scenario.

[0074] The electronic device 1000 can determine a certain ratio of the default session time as the additional session time. For example, the electronic device 1000 can determine 20% of the default session time as the additional session time.

[0075] The electronic device 1000 can determine the sum of the default session time and the additional session time as the session duration. The electronic device 1000 can identify whether the session duration has elapsed, and when the session duration has elapsed, can output the session end notification 250.

[0076] According to an embodiment of the disclosure, the electronic device 1000 can extend or shorten the default session time based on statistical data related to the conversation history information of the other user for the current response message.

[0077] In a conversation through a chatbot, the session duration is determined according to the user's personal information protection, and the conversation is terminated when the session duration is terminated. In a chatbot according to the related art, since the conversation is terminated when the session duration elapses, the conversation is terminated before the question is completely solved, and thus, in order to return to the question solving scenario previously in progress, the user has to input the message input at the previous point in time again and be provided with the same response message again, which is cumbersome and time-consuming. In addition, even though different processing times are required for the chatbot according to the question solving scenario according to the related art, since the total session duration is provided, there is a problem in that the conversation is accidentally terminated due to the time taken by the user to solve the question. Extending the session duration for all question solving scenarios can be considered as a solution. However, this solution can cause an increase in the number of conversations simultaneously managed by the chatbot, can correspondingly increase resource usage, and can increase network costs.

[0078] The electronic device 1000 according to an embodiment of the disclosure can provide a chatbot service that adaptively controls the session time according to the question solving scenario by determining the default session time according to the difficulty level of the response message, determining whether to provide the additional session time based on the conversation history information before the currently output response message, and extending or shortening the session time by using statistical data on the conversation usage history information of a plurality of users. Accordingly, the electronic device 1000 of the disclosure can have technical effects of preventing the user from wasting time and inconvenience due to repeated provision of the same response message to the user due to the session duration elapsing, and improving resource usage and network cost efficiency.

[0079] Figure 3 is a block diagram of components of an electronic device according to an embodiment of the disclosure.

[0080] Reference Figure 3The electronic device 1000 can include a communication interface 1100, a processor 1200, and a memory 1300. Figure 3 Only basic components for explaining the functions and / or operations of the electronic device 1000 are illustrated, and components included in the electronic device 1000 are not limited to the components illustrated in Figure 3

[0081] The communication interface 1100 can be configured to perform transmission and reception of data between the electronic device 1000 and other devices (e.g., the client devices 3001, 3002, and 3003). The communication interface 1100 can perform data communication with a server or other devices by using at least one of wired and wireless data communication methods including Ethernet, wired or wireless LAN, Wi-Fi, WFD, or WiGig. According to an embodiment of the disclosure, the electronic device 1000 can include a server, receive text data about an input message from other devices (e.g., the client devices 3001, 3002, and 3003) through the communication interface 1100, and transmit text data about a response message to the other devices.

[0082] Although not illustrated in Figure 3 , the electronic device 1000 can include a voice input interface and an automatic speech recognition (ASR) model. The voice input interface can include a device that receives a voice input (e.g., a user question from a user). The voice input interface can include, for example, a microphone. The voice input interface can receive a voice input (e.g., a user's utterance) from a user through the microphone, and obtain a voice signal from the received voice input. According to an embodiment of the disclosure, the processor 1200 of the electronic device 1000 can receive a voice input of a user through the microphone, convert the received voice input into a sound signal, and obtain a voice signal by removing noise (e.g., a non-voice component) from the sound signal.

[0083] The ASR model is a speech recognition model that recognizes a user's voice, and is a model that is trained to convert a voice input received from a user into text and output the converted text. According to an embodiment of the disclosure, the electronic device 1000 can convert a voice input into text by using the ASR model.

[0084] Although not illustrated in the drawings, the electronic device 1000 can include a voice preprocessing module having a function of detecting a designated voice input (e.g., a wake-up input such as "Hi Bixby," "Ok Google," etc.) or a function of preprocessing a voice signal obtained from some voice input.

[0085] ​The processor 1200 can be configured to control the electronic device 1000 to perform operations and / or functions by reading and executing one or more instructions or program codes. The memory 1300 can store instructions or program codes executed by the processor 1200, and the processor 1200 can execute the instructions or program codes loaded from the memory 1300, but is not limited thereto. The processor 1200 itself can include instructions or program codes.

[0086] The processor 1200 can include at least one of a central processing unit (CPU), a microprocessor, a graphics processing unit (GPU), an application-specific integrated circuit (ASIC), a digital signal processor (DSP), a digital signal processing device (DSPD), a programmable logic device (PLD), or a field programmable gate array (FPGA) that performs arithmetic, logical, and input / output operations and signal processing, but is not limited thereto.

[0087] The memory 1300 can store instructions and program codes readable by the processor 1200. The memory 1300 can include at least one type of storage medium, such as a flash memory type, a hard disk type, a multimedia card micro type, a card type memory (e.g., a secure digital (SD) or an extreme digital (xD)-picture card memory (xD memory)), a random access memory (RAM), a static random access memory (SRAM), a read-only memory (ROM), an electrically erasable programmable read-only memory (EEPROM), a programmable read-only memory (PROM), a magnetic memory, a magnetic disk, or an optical disk.

[0088] The processor 1200 can include a default session time determination module 1210, an additional session time determination module 1220, an end notification module 1230, and a session time adjustment module 1240.

[0089] The processor 1200 can receive a user's input message through the communication interface 1100. The processor 1200 can identify a domain, an intent, and an entity of the input message by analyzing the received user input message, and output a response message for the identified domain, intent, and entity of the input message. According to an embodiment of the disclosure, the processor 1200 can parse text in morphemes, words, or phrases by using a natural language understanding model, and infer a meaning of the parsed text by using linguistic features (e.g., grammatical elements) of the morphemes, words, or phrases. The processor 1200 can identify a domain, an intent, and an entity by comparing the inferred word meaning with each of pre-defined domains, intents, and entities provided by the natural language understanding model. According to an embodiment of the disclosure, the processor 1200 can output a response message set corresponding to the domain, intent, and entity.

[0090] The default session time determination module 1210 is a module configured to obtain information on the difficulty level 540 of the response message by analyzing the output response message, and determine a default session time based on the difficulty level 540 of the response message. According to an embodiment of the disclosure, the default session time determination module 1210 can include a deep neural network model 1214. However, the disclosure is not limited thereto, and the deep neural network model 1214 can be executed by the processor 1200 or stored in the storage 1300 as a model independent of the default session time determination module 1210. The processor 1200 can apply text data of the output response message to the deep neural network model 1214 as input, and obtain an output label value of the difficulty level 540 via the deep neural network model 1214 by executing an instruction or program code on the default session time determination module 1210.

[0091] The deep neural network model 1214 can be a model trained to classify the response message according to the difficulty level 540. The deep neural network model 1214 can be implemented as a trained neural model at a certain time before the response message is output. According to an embodiment of the disclosure, the deep neural network model 1214 can be trained through supervised learning in which a plurality of training messages previously obtained are applied as input, and a label value of the difficulty level 540 set for each of the plurality of training messages is applied as an output true value. Here, the "difficulty level" 540 refers to the degree of difficulty in solving a problem through a request, a question, an answer, or information provided by a message. The label value of the difficulty level 540 can be defined as an integer value such as 0, 1, 2,..., n, but is not limited thereto.

[0092] The deep neural network model 1214 can be implemented as, for example, a convolutional neural network model (CNN), but is not limited thereto. According to an embodiment of the disclosure, the deep neural network model can be implemented as a neural network model based on deep learning such as a recurrent neural network (RNN), a restricted Boltzmann machine (RBM), a deep belief network (DBN), a bidirectional recurrent deep neural network (BRDNN), or a deep Q-network.

[0093] The processor 1200 can input the response message to the deep neural network model 1214, and obtain a label value of the difficulty level 540 by which the response message can be classified through training by using the deep neural network model 1214. According to an embodiment of the disclosure, the processor 1200 can search for a label value for the obtained difficulty level 540 from a session time database (DB) 1310, and obtain information on a session time mapped to correspond to the label value from the session time DB 1310.

[0094] The conversation time DB 1310 is a database that stores conversation times according to the difficulty levels 540. According to an embodiment of the disclosure, the conversation time DB 1310 can store data of a plurality of difficulty marker values and a plurality of conversation times. The plurality of conversation times can be mapped as marker values corresponding to each of the plurality of difficulties. According to an embodiment of the disclosure, the conversation time DB 1310 can store the marker values for each of the plurality of difficulties and data of the plurality of conversation times as a key-value type.

[0095] A detailed embodiment of the disclosure in which the default conversation time determination module 1210 is used to determine a default conversation time for a response message will be described with reference to FIG. 13. Figures 5a to 5c And Figure 6 A detailed embodiment of the disclosure in which the default conversation time determination module 1210 is used to determine a default conversation time for a response message will be described with reference to FIG. 13.

[0096] The additional conversation time determination module 1220 is a module configured to receive information on whether the default conversation time for a response message has elapsed from the default conversation time determination module 1210 and determine an additional conversation time for the response message. The additional conversation time determination module 1220 can determine the additional conversation time based on conversation history information before the response message is output. According to an embodiment of the disclosure, the conversation history information can include information on at least one of the number of response messages output in a previous conversation before the response message is output or the conversation usage time for each scenario according to the previously output response messages.

[0097] The processor 1200 can determine whether to provide an additional conversation time based on the conversation history information by executing an instruction or program code on the additional conversation time determination module 1220. According to an embodiment of the disclosure, the processor 1200 can calculate a usage ratio of the conversation time by dividing the conversation usage time for each scenario performed before the response message is output by the default conversation time and determine whether to provide an additional conversation time based on the calculated usage ratio of the conversation time. According to another embodiment of the disclosure, the processor 1200 can calculate an average value of the conversation usage time for each scenario performed before the response message is output and determine whether to provide an additional conversation time based on the calculated average value.

[0098] The processor 1200 can determine a certain ratio of the sum of the conversation usage times for each scenario as an additional conversation time. A detailed embodiment of the disclosure in which the processor 1200 determines an additional conversation time will be described with reference to FIG. 14. Figures 8 to 10 A detailed embodiment of the disclosure in which the processor 1200 determines an additional conversation time will be described with reference to FIG. 14.

[0099] The end notification module 1230 is a module configured to output a notification message about the end of the session when it is confirmed that the session duration has elapsed. According to an embodiment of the disclosure, the end notification module 1230 can include a natural language generation model. The end notification module 1230 can generate a message notifying the user that the session keeping time has elapsed by using the intent and entity associated with the response message, and output the generated message. The session end message can be generated and output as, for example, "Did you solve the question about the refrigerator? End the chat."

[0100] The session duration can be determined by the sum of the default session time and the additional session time. The processor 1200 can identify whether the session duration has elapsed by executing the instructions or program codes regarding the end notification module 1230, and output a session end notification message according to the identification result. It will be described with reference to Figure 11 An embodiment of the disclosure regarding output of a session end notification message will be described.

[0101] The session time adjustment module 1240 is a module configured to adjust the default session time based on session use history information of a plurality of users regarding the output response message. The session time adjustment module 1240 can receive information about the session duration from the end notification module 1230, and thus can search for statistics data about the session use history information related to the currently output response message from the session use history statistics DB 1320. The session time adjustment module 1240 can obtain the session use history information of a plurality of users according to the search result of the session use history statistics DB 1320, and can lengthen or shorten the default session time based on the session use history information. The session time adjustment module 1240 can provide information about the lengthened or shortened default session time to the default session time determination module 1210.

[0102] The session use history statistics DB 1320 is a database for storing statistics data about the session use history of a plurality of users. According to an embodiment of the disclosure, the session use history information can include statistics data of a plurality of users related to at least one of the number of session extension requests at a point in time before the current response message is output, the number of scenarios executed before the response message, or the session use time for the response message.

[0103] The processor 1200 can search for statistics data about the session use history of a plurality of users from the session use history statistics DB 1320 by executing the instructions or program codes regarding the session time adjustment module 1240, and obtain the session use history information according to the search result. The processor 1200 can lengthen or shorten the default session time based on the session use history information. It will be described with reference to Figure 12 andFigure 13 The detailed embodiments of the disclosure describing the processor 1200 adjusting the default session time based on the session use history information.

[0104] The session time DB 1310 and the session use history statistics DB 1320 can include a non-volatile memory. The non-volatile memory means a storage medium that stores and maintains information even when not powered and can use the stored information again when powered. The non-volatile memory can include at least one of, for example, a flash memory, a hard disk, a solid state drive (SSD), a multimedia card micro memory, a card-type memory (e.g., an SD or XD memory, etc.), a read-only memory (ROM), a magnetic memory, a magnetic disk, or an optical disk.

[0105] Referring to FIG. 13, Figure 3 , it is shown that the session time DB 1310 and the session use history statistics DB 1320 are included in the memory 1300, but are not limited thereto. According to an embodiment of the disclosure, at least one of the session time DB 1310 or the session use history statistics DB 1320 can be configured in a database form independent of the memory 1300. According to another embodiment of the disclosure, at least one of the session time DB 1310 or the session use history statistics DB 1320 can be connected with the electronic device 1000 via the communication interface 1100 by wired or wireless communication as a component of an external device or an external server, not as a component of the electronic device 1000.

[0106] Figure 4 is a flowchart of an operation method of an electronic device according to an embodiment of the disclosure.

[0107] Referring to FIG. 13, Figure 4 At operation S410, the electronic device 1000 can output a response message to the user's input message. According to an embodiment of the disclosure, the electronic device 1000 can analyze the user's input message by using a natural language understanding model to recognize a domain, an intent, and an entity associated with the input message. According to an embodiment of the disclosure, the electronic device 1000 can parse text included in the input message in a morpheme, a word, or a phrase unit, and infer a meaning of a word extracted from the parsed text by using a natural language understanding model using a language feature (e.g., a grammatical element) of the morpheme, the word, or the phrase. The electronic device 1000 can identify the domain, the intent, and the entity by comparing the inferred word meaning with each of predetermined domains, intents, and entities provided by the natural language understanding model.

[0108] The electronic device 1000 can output a response message set corresponding to the domain, the intent, and the entity identified from the input message. According to an embodiment of the disclosure, the electronic device 1000 can generate a response message for solving a problem related to the domain, the intent, and the entity using a natural language generation model, and output the generated response message.

[0109] In operation S420, the electronic device 1000 can determine a default session time in which the session is maintained based on a difficulty level of the output response message. According to an embodiment of the disclosure, the electronic device 1000 can obtain an output label value of the difficulty level by applying the response message as input to a deep neural network model, and can determine a session time mapped to the obtained label value as the default session time. The deep neural network model can be an artificial intelligence model trained through supervised learning in which a plurality of training messages previously obtained are applied as input and a label value of the difficulty level is applied as an output true value.

[0110] In operation S430, the electronic device 1000 can determine an additional session time based on the conversation history information before the response message is output. According to an embodiment of the disclosure, in operation S410, the electronic device 1000 can determine whether to provide the additional session time based on at least one of the number of response messages previously output before the response message is output, or a session usage time used in a scenario performed through the response message previously output. According to an embodiment of the disclosure, a preset ratio of the sum of the session usage times for each scenario can be determined as the additional session time.

[0111] In operation S440, the electronic device 1000 can wait for a user to input an additional message during a session duration that is the sum of the default session time and the additional session time.

[0112] In operation S450, the electronic device 1000 can determine the end of the session based on whether the additional message is input. According to an embodiment of the disclosure, the electronic device 1000 can confirm whether the session duration has elapsed, and output a session end notification message based on the confirmed result.

[0113] The electronic device 1000 can output a preset session end notification message for the response message, but is not limited thereto. According to an embodiment of the disclosure, the electronic device 1000 can generate a session end notification message that asks whether a problem related to the intent and the entity identified from the response message is solved by using a natural language generation model, and output the generated response message.

[0114] Figure 5ais a diagram illustrating a method of training a deep neural network model for classifying a difficulty level of a response message, performed by an electronic device according to an embodiment of the disclosure.

[0115] Referring to Figure 5a , the electronic device 1000 can perform training by applying a plurality of training messages 510-1 to 510-n (hereinafter also referred to as a first training message 510-1 and a second training message 510-2) as input, and applying a difficulty label value 520-1 to 520-n of each of the plurality of training messages 510-1 to 510-n as an output true value. As messages previously obtained for training the deep neural network model 1214, the plurality of training messages 510-1 to 510-n can include examples of response messages of a chatbot with respect to a user input message. The plurality of training messages 510-1 to 510-n can include a context related to a request, a question, an answer, or the provision of information. For example, the plurality of training messages 510-1 to 510-n can be "The price of Galaxy S7 LTE 128G is 762,000 won." or "Did you check whether the power cord is inserted?"

[0116] The plurality of training messages 510-1 to 510-n can be input to the embedding module 1212 before being input to the deep neural network model 1214. The embedding module 1212 can parse the input text and extract at least one word therefrom, and can quantize the extracted at least one extracted word into a vector. For example, the embedding module 1212 can convert at least one word into an embedding vector by using a known machine learning model such as Word2vec, GloVe, or onehot encoding. However, the embedding model used by the embedding module 1212 is not limited to the above-described examples.

[0117] The processor 1200 of the electronic device 1000 (refer to Figure 3 ) can extract at least one word by parsing the plurality of training messages 510-1 to 510-n in units of words using the embedding module 1212, and convert the extracted word into an embedding vector. The processor 1200 can arrange the embedding vector of at least one word in a matrix form. In the embodiment of the disclosure shown in Figure 5a , the first training message 510-1 can be converted into a first embedding vector by the embedding module 1212, and the second training message 510-2 can be converted into a second embedding vector by the embedding module 1212.

[0118] The processor 1200 can input the embedding vector to the deep neural network model 1214. In this case, the embedding vector can be input to the deep neural network model 1214 as a feature vector.

[0119] The deep neural network model 1214 is an artificial intelligence model configured to perform training by using the embedding vectors provided from the embedding module 1212. According to an embodiment of the disclosure, the deep neural network model 1214 can be trained through supervised learning in which the embedding vectors are applied as input and the difficulty label values 520-1 to 520-n are applied as true values. Here, the "difficulty level" refers to the degree of difficulty in solving a problem through a request, a question, an answer, or information provided by the plurality of training messages 510-1 to 510-n. In Figure 5a In the illustrated embodiment of the disclosure, the difficulty level can be an integer value ranging from 0 to n, but is not limited thereto. The "label value of the difficulty level" refers to a numerical value preset with respect to the difficulty. For example, the first training message 510-1 "The price of the Galaxy S7 LTE model is 762,000 won." can be mapped to a difficulty level of 0, and thus the first label value 520-1 of the difficulty level of 0 can be trained to output a true value, and the second training message 510-2 "Do you check if the power cord is properly plugged in?" can be mapped to a difficulty level of 1, and thus the second label value 520-2 of the difficulty level of 1 can be trained to output a true value.

[0120] According to an embodiment of the disclosure, the deep neural network model 1214 can include a plurality of hidden layers, which are inner layers that perform operations. The deep neural network model 1214 can include, for example, at least one of a CNN, a recurrent neural network model (RNN), an RBM, a DBN, a BRDNN, or a deep Q-network, but is not limited to the above examples. However, the deep neural network model 1214 is not limited to the above examples, and can include all known neural network models based on deep learning.

[0121] When the deep neural network model 1214 is implemented as a CNN, a feature value can be extracted from the embedding vector applied as input by using a filter having a preset size and a preset number of channels, a plurality of layers including the extracted feature value can be obtained, and a feature vector map can be obtained by applying weights to the plurality of layers. In the process of obtaining the feature vector map, a rectified linear unit (Relu) model can be used, and in order to improve efficiency, the training model can be regularized through dropout, and an operation of performing pooling or max pooling can also be added in the process. Then, the feature values obtained through the pooling or max pooling are merged through a fully connected layer, and can be trained to output a label related to a compression ratio value through an activation function including softmax, sigmoid, and hyperbolic tangent. The values of the parameters included in the deep neural network model 1214 can be changed by training the deep neural network model 1214. For example, the weight and bias values of the layers included in the deep neural network model 1214 can be changed.

[0122] Figure 5b is a diagram illustrating a method of obtaining information about a difficulty level of a response message by using a deep neural network model, performed by an electronic device, according to an embodiment of the disclosure.

[0123] Referring to Figure 5b , the electronic device 1000 can output a response message 530 for a user's input message. The output response message 530 can be input to the embedding module 1212 by the electronic device 1000 before being input to the deep neural network model 1214. When the text included in the response message 530 is input to the embedding module 1212, the embedding module 1212 can extract at least one word by parsing the input text in units of words, and can output an embedding vector by quantizing the extracted at least one word. The embedding vector output by the embedding module 1212 can be input to the deep neural network model 1214.

[0124] As shown in Figure 5a , the deep neural network model 1214 is an artificial intelligence model trained through supervised learning, in which a plurality of training messages 510-1 to 510-n (refer to Figure 5a ) are applied as input, and a difficulty label value 520-1 to 520-n (refer to Figure 5a ) is applied as an output true value. When the embedding vector converted from the response message 530 is input to the deep neural network model 1214, a label value predicted as a difficulty level of the response message 530 can be output. In the embodiment of the disclosure shown in Figure 5b , the difficulty level of the response message 530 "The price of the Galaxy Z Fold 5G 256G model is 2,199,000 won." can be predicted as 0, and a label value of the difficulty level 0 can be output by the deep neural network model 1214.

[0125] Figure 5c is a diagram illustrating a method of determining a default conversation time based on a difficulty level, performed by an electronic device, according to an embodiment of the disclosure.

[0126] Referring to Figure 5c , the label value of the difficulty level of the response message 530 (refer to Figure 5b ) output by the deep neural network model 1214 (refer to Figure 5b ) can be input to the default conversation time determination module 1210. The default conversation time determination module 1210 is a module configured to obtain information about the difficulty level of the response message 530 and determine a default conversation time based on the difficulty level of the response message 530. According to an embodiment of the disclosure, the default conversation time determination module 1210 can search for a conversation time from the conversation time DB 1310 according to the difficulty.

[0127] Session Time DB 1310 is a database that stores session times according to difficulty levels. According to an embodiment of this disclosure, Session Time DB 1310 can store data for multiple difficulty marker values ​​1312-1 to 1312-n (hereinafter also referred to as the first marker value 1312-1 and the nth marker value 1312-n) and multiple session times 1314-1 to 1314-n (hereinafter also referred to as the first session time 1314-1 and the nth session time 1314-n). The multiple session times 1314-1 to 1314-n can be mapped to each of the multiple difficulty marker values ​​1312-1 to 1312-n. According to an embodiment of this disclosure, Session Time DB 1310 can store the data of the multiple difficulty marker values ​​1312-1 to 1312-n and the multiple session times 1314-1 to 1314-n as key-value types. Figure 5c In the embodiment of this disclosure shown, the first marker value 1312-1 with a difficulty level of 0 and the first session time 1314-1 of 30 seconds can be stored as a key-value type in the session time DB 1310. Similarly, the nth marker value 1312-n with a difficulty level of n and the nth session time 1314-n of 90 minutes can be stored as a key-value type in the session time DB 1310.

[0128] Electronic device 1000 can search for session times in session time database 1310 based on the difficulty level of the response message, and provide information about the session time to electronic device 1000 by using default session time determination module 1210. Figure 5c In the illustrated embodiment, the default session time determination module 1210 can receive input of a flag value for the difficulty level of the response message 530 (see reference). Figure 5b It can search for the first session time 1314-1, which is the session time corresponding to the marker value of difficulty level 0 in session time DB 1310. The first session time 1314-1 can be 30 seconds, and by default, the session time determination module 1210 can obtain information about the session time (30 seconds) from the session time DB 1310.

[0129] Figure 6 This is a flowchart illustrating an operation method of an electronic device according to an embodiment of the present disclosure.

[0130] refer to Figure 6 , Figure 6 Operations S610 and S620 in the illustrated operation are Figure 4 Details of operation S410 shown. Figure 6 The operations S630 to S660 shown are Figure 4Details of the illustrated operation S420.

[0131] At operation S610, the electronic device 1000 can determine a response message mapped as corresponding to the input message. The electronic device 1000 can identify a domain, an intent, and an entity of the input message by analyzing the input message received from the user. According to an embodiment of the disclosure, the electronic device 1000 can parse the text in morphemes, words, or phrases by using a natural language understanding model, and infer the meaning of the parsed text by using linguistic features (e.g., grammatical elements) of the morphemes, words, or phrases. The electronic device 1000 can identify the domain, the intent, and the entity by comparing the inferred word meaning with each of predetermined domains, intents, and entities provided by the natural language understanding model.

[0132] The electronic device 1000 can output a response message regarding the identified domain, intent, and entity. According to an embodiment of the disclosure, the response message can be pre-set to correspond to the domain, the intent, and the entity. However, the disclosure is not limited thereto, and the electronic device 1000 can generate a response message related to the domain, the intent, and the entity by using a natural language generation model.

[0133] At operation S620, the electronic device 1000 can output the determined response message.

[0134] At operation S630, the electronic device 1000 can identify whether a default conversation time for the response message is set.

[0135] At operation S640, when the default conversation time for the response message is not set, the electronic device 1000 can output a marked value of a difficulty level of the response message by a deep neural network model. Here, the "deep neural network model" is an artificial model trained through supervised learning in which a plurality of training messages are applied as input and a marked value of a difficulty level set with respect to each of the plurality of training messages is applied as an output true value. The deep neural network model is a model trained before the input message is received from the user in operation S610. The electronic device 1000 can obtain a marked value predicted as a difficulty level of the response message by applying the response message as input to the deep neural network model.

[0136] At operation S650, the electronic device 1000 can determine a conversation time corresponding to the obtained marked value of the difficulty level as a default conversation time. According to an embodiment of the disclosure, the electronic device 1000 can search for a conversation time DB 1310 (refer to FIG. 13) via the deep neural network model by using the obtained marked value of the difficulty level, and determine a conversation time corresponding to the marked value of the difficulty level as the default conversation time. Figure 5c), to obtain a session time mapped as corresponding to the tagged value. The electronic device 1000 can determine the session time obtained from the session time DB 1310 as a default session time.

[0137] The default session time can be defined according to information provided by a response message, a response, or a request. In operation S660, when a default session time of a response message is set, the electronic device 1000 can determine a preset session time as the default session time.

[0138] Figure 7 is a time table 700 illustrating a relationship between a default session time according to an input message input by a user and a session use time according to a response message provided by a chatbot according to an embodiment of the disclosure.

[0139] Referring to Figure 7 When the user inputs a first input message 711 at a first time point t1, the chatbot can output a first response message 721 for the first input message 711 at a second time point t2 (hereinafter, the first response message 721 and the second response message 722 can also be referred to as a previously output response message 721 and a previously output response message 722). A first default session time 731 can be provided for a predetermined time from the second time point t2 at which the first response message 721 is output. It has been described above with reference to Figure 5c and Figure 6 A detailed method of determining a default session time is described, and thus a redundant description is omitted. The user can input a second input message 712 at a third time point t3 before the first default session time 731 elapses, and in this case, a session time can be updated.

[0140] The first session use time 741 refers to a time used by the user to perform a specific operation according to a request, an answer, or information provided by the first response message 721 after the second time point t2 at which the first response message 721 is provided to the user, or a time used until the third time point t3 at which the second input message 712 is input for an additional inquiry or an additional request. For example, the first default session time 731 can be 5 minutes, and the first session use time 741 can be 4 minutes. When the session use time exceeds the default session time, the session can be terminated.

[0141] The chatbot can output a second response message 722 related to the second input message 712 at a fourth time point t4. In the same manner as the first default session time 731, a second default session time 732 can be provided for a preset time from the fourth time point t4 at which the second response message 722 is output. The user can input a third input message 713 at a fifth time point t5 before the second default session time 732 elapses, and in this case, a session time can be updated.

[0142] The second session usage time 742 refers to the time taken by the user to perform a specific operation based on the request, response, or information provided by the second response message 722 after the fourth time point t4 when the second response message 722 is provided to the user, or the time taken until the fifth time point t5 when the third input message 713 is entered in response to an additional inquiry or request. For example, the second default session time 732 may be 10 minutes, and the second session usage time 742 may be 6 minutes.

[0143] A "scenario" refers to a problem-solving action (or process) provided by a chatbot in response to a user's input, such as a request, inquiry, response, or information provision. According to embodiments of this disclosure, a scenario can be defined as the point in time from when a response message is provided to the user to when the user inputs additional information. Figure 7 In the embodiments of this disclosure shown, the first scenario may refer to the time interval between the second time point t2 of outputting the first response message 721 and the third time point t3 of inputting the second input message 712, and the second scenario may refer to the time interval between the fourth time point t4 of outputting the second response message 722 and the fifth time point t5 of inputting the third input message 713.

[0144] Electronic device 1000 can determine additional session time based on information relating to at least one of the following: the number of response messages previously output before the current output response message is output, or the session usage time used in a scenario executed via previously output response messages. According to embodiments of this disclosure, when the number of previously output response messages is one or more, electronic device 1000 can determine whether to provide additional session time. A preset ratio of the sum of session usage time for each scenario can be determined as the additional session time. (Refer to...) Figures 8 to 10 This disclosure describes a detailed embodiment of how an electronic device 1000 determines additional session time.

[0145] Figure 8 This is a flowchart illustrating a method performed by an electronic device, according to an embodiment of the present disclosure, for determining whether to provide additional session time.

[0146] refer to Figure 8 , Figure 8 The operations S810 to S830 shown are Figure 4 Details of operation S430 are shown. Operation S810 can be performed... Figure 4 The operation shown in S420 will be executed afterward. Figure 4 The operation S440 shown can be performed after operations S820 and S830.

[0147] The electronic device 1000 can identify whether there is a previously output response message before the current output response message at operation S810.

[0148] When it is confirmed that there is a previously output response message at operation S820, the electronic device 1000 can determine a certain ratio of the sum of the session usage time for each scenario as an additional session time according to the previously output response message. Referring to Figure 7 , when the current output response message is the third response message 723 (referring to Figure 7 ), since two response messages including the first response message 721 (referring to Figure 7 ) and the second response message 722 (referring to Figure 7 ) are output, the electronic device 1000 can determine whether to provide an additional session time. Also, since the first session usage time 741 (referring to Figure 7 ) and the second session usage time 742 (referring to Figure 7 ) are used for the first scenario and the second scenario, respectively, the electronic device 1000 can determine a certain ratio (e.g., 20%) of the sum of the first session usage time 741 and the second session usage time 742 as an additional session time. For example, when the first session usage time 741 is 4 minutes and the second session usage time 742 is 6 minutes, the electronic device 1000 can determine 2 minutes (i.e., 20% of the sum of the first session usage time 741 and the second session usage time 742, which is 10 minutes) as an additional session time of the third response message 723 (referring to Figure 7 ).

[0149] When it is confirmed that there is no previously output response message at operation S830 (i.e., when the current output response message is the first response message), the electronic device 1000 can determine not to provide an additional session time.

[0150] Figure 9 is a flowchart illustrating a method of determining whether to provide an additional session time, performed by an electronic device, according to an embodiment of the disclosure.

[0151] Referring to Figure 9 , Figure 9 , operations S910 through S940 illustrated in Figure 4 are details of operation S430. Operation S910 can be performed after operation S420 illustrated in Figure 4 is performed. Figure 4 Operation S440 illustrated in

[0152] The electronic device 1000 can calculate a session time use ratio for each scene, at operation S910. According to an embodiment of the disclosure, the electronic device 1000 can calculate the session time use ratio by performing an arithmetic operation of dividing the session use time by the default session time. For example, when the default session time is 10 minutes and the session use time is 6 minutes, the session time use ratio is 60%.

[0153] According to an embodiment of the disclosure, the electronic device 1000 can calculate the session time use ratio for each scene according to the previously output response message. Referring to FIG. 7, when the currently output response message is the third response message 723 (referring to FIG. 7), the electronic device 1000 can calculate the first session time use ratio for the first scene and the second session time use ratio for the second scene. Figure 7 , the electronic device 1000 can calculate the first session time use ratio for the first scene and the second session time use ratio for the second scene. Figure 7 ) is 5 minutes and the first session use time 741 (referring to FIG. 7) is 4 minutes, the first session time use ratio is 80%; and when the second default session time 732 (referring to FIG. 7) is 10 minutes and the second session use time 742 (referring to FIG. 7) is 6 minutes, the second session time use ratio is 60%. Figure 7 Figure 7 Figure 7 Figure 7

[0154] The electronic device 1000 can compare the session time use ratio with a preset threshold α, at operation S920. The electronic device 1000 can determine whether the session time use ratio exceeds the threshold α.

[0155] The electronic device 1000 can determine to provide an additional session time when the session time use ratio exceeds the preset threshold α, at operation S930. For example, when the preset threshold α is 50%, since the first session time use ratio is 80% and the second session time use ratio is 60%, the electronic device 1000 can determine that the session time use ratio has exceeded the preset threshold α, and determine to provide an additional session time.

[0156] The additional session time can be determined as a certain ratio of the sum of the session use times for each scene. For example, when the first session use time is 4 minutes and the second session use time is 6 minutes, the additional session time can be determined as 2 minutes, i.e., 20% of the sum of the first session use time and the second session use time.

[0157] The electronic device 1000 can determine not to provide an additional session time when the session time use ratio is equal to or less than the preset threshold α, at operation S940.

[0158] Figure 10 ​​​​This is a flowchart illustrating a method performed by an electronic device, according to an embodiment of the present disclosure, for determining whether to provide additional session time.

[0159] refer to Figure 10 , Figure 10 Operations S1010 to S1040 shown are Figure 4 Details of operation S430 are shown. Operation S1010 can be performed... Figure 4 The operation shown in S420 will be executed afterward. Figure 4 The operation S440 shown can be performed after operations S1030 and S1040.

[0160] In operation S1010, electronic device 1000 can calculate the average session usage time before outputting a response message. (Reference) Figure 7 The current output response message is the third response message 723 (see reference). Figure 7 When the electronic device 1000 calculates the average session usage time in the first scenario and the average session usage time in the second scenario, it can calculate the average session usage time in the first scenario. For example, when the first session usage time used in the first scenario is 741 (reference...), the electronic device 1000 can calculate the average session usage time in the first scenario and the average session usage time in the second scenario. Figure 7 The second session, used in the second scenario, lasted 4 minutes and 742 seconds (reference). Figure 7 If the time interval is 6 minutes, the average time can be calculated to be 5 minutes.

[0161] In operation S1020, the electronic device 1000 can compare the calculated average value with a preset threshold t. th A comparison is made to determine whether the average exceeds the threshold t. th .

[0162] In operation S1030, when the average value exceeds the preset threshold t th At that time, electronic device 1000 can determine to provide additional session time. For example, when a preset threshold t th When the time is 3 minutes, since the average value calculated in operation S1010 is 5 minutes, the electronic device 1000 can determine to provide additional session time.

[0163] Additional session time can be determined as a percentage of the sum of session usage time for each scenario. For example, when the first session usage time is 4 minutes and the second session usage time is 6 minutes, the additional session time can be determined to be 2 minutes, that is, 20% of the sum of the first and second session usage times.

[0164] In operation S1040, when the average value is equal to or less than the preset threshold t th At that time, electronic device 1000 can determine that it does not provide additional session time.

[0165] The more the previously output response messages (refer to Figure 7 ) before the current output response message 723, the greater the burden on the user to end the session. When time is taken due to the user's action (e.g., checking a fault, identifying a device state, etc.) according to the current response message 723, or when the user does not input an additional message due to a mistake, the default session time can pass, and thus the session can end. In this case, the user can need to repeatedly input the same input message to achieve the same response message 723, which is cumbersome and reduces the user's convenience.

[0166] According to the electronic device 1000 according to the embodiment shown in Figures 8 to 10 , the information on at least one of the number of the previously output response messages 721 and 722 (refer to Figure 7 ) before the current output response message 723 and the session usage time according to each scenario of the previously output response messages 721 and 722 can be used to provide the additional session time, thereby solving the problem of the session ending unexpectedly due to the default session time passing. In addition, by considering the session usage time used by the user to achieve the current response message 723, an optimized additional session time can be provided for each user, thereby improving the user's convenience.

[0167] Figure 11 is a flowchart illustrating a method of outputting a session end notification performed by an electronic device according to an embodiment of the disclosure.

[0168] Referring to Figure 11 , Figure 11 , operations S1110 to S1150 shown in Figure 4 are details of operation S450 shown in Figure 4 . Operation S1110 can be performed after operation S440 shown in

[0169] At operation S1110, the electronic device 1000 can identify whether a session duration time has passed. According to an embodiment of the disclosure, the session duration time can be determined by the sum of the default session time and the additional session time. The electronic device 1000 can identify whether an additional message of the user has been input during the session duration time from when the response message is output.

[0170] At operation S1120, the electronic device 1000 can output a session end notification 250. According to an embodiment of the disclosure, the electronic device 1000 can generate a message notifying the user that the session duration time has ended using a natural language generation model by using the intent and entity related to the response message, and can output the generated response message. The session end notification message can be, for example, "Did you solve the problem about the refrigerator? End the chat."

[0171] The electronic device 1000 can identify whether a user's response message to the end notification is input, in operation S1130. The user's response message can be a message indicating whether the question has been resolved with Y / N (such as "Yes" or "No"), but is not limited thereto. The response message can be a message related to an additional question or an additional request.

[0172] In operation S1140, when it is confirmed that the user's response message is input, the electronic device 1000 can update the session duration.

[0173] In operation S1150, when it is confirmed that the user's response message is not input, the electronic device 1000 can determine to end the session. According to an embodiment of the disclosure, even when the response message is input, when the response message is a message notifying that the question has been resolved (for example, for "Did you solve the question about the refrigerator? End the chat.", the response message is "Yes"), the electronic device 1000 can determine to end the session.

[0174] In Figure 11 In the disclosed embodiment of the disclosure, when the session duration has elapsed, the electronic device 1000 can not immediately end the session and can output a session end notification message, thereby notifying the user of whether the session will end and allowing the user to input an additional message if necessary. Accordingly, the electronic device 1000 of the disclosure can prevent the problem of the session ending unexpectedly and improve the convenience of the user.

[0175] Figure 12 FIG. 12 is a flowchart illustrating a method of adjusting a default session time performed by an electronic device according to an embodiment of the disclosure.

[0176] Referring to Figure 12 In operation S1210, the electronic device 1000 can store statistical data regarding session usage history information of a plurality of users. According to an embodiment of the disclosure, the session usage history information can be statistical data of a plurality of users regarding at least one of the number of session extension requests, the number of scenarios performed before a response message is output, or the session usage time. According to an embodiment of the disclosure, the electronic device 1000 can classify and store the session usage history information for each response message.

[0177] In operation S1220, the electronic device 1000 can adjust the default session time based on the session usage history information. The electronic device 1000 can extend or shorten the default session time based on the session usage history information of a plurality of users related to the currently output response message. Detailed embodiments of the disclosure in which the electronic device 1000 extends or shortens the default session time will be described with reference to Figure 13

[0178] ​Figure 13 is a flowchart illustrating a method of extending or shortening a default session time, performed by an electronic device, according to an embodiment of the disclosure.

[0179] Referring to Figure 13 In operation S1310, the electronic device 1000 can identify whether there is a request for extending a session in a previous session. According to an embodiment of the disclosure, the electronic device 1000 can search for statistics data about session use history information related to a currently output response message from session use history information of a plurality of users stored in a session use history statistics DB 1320 (refer to FIG. 13B). Figure 3 According to a search result of the session use history statistics DB 1320, the electronic device 1000 can obtain information about whether a plurality of users have requested a session extension in a previous session before the response message is output.

[0180] In operation S1320, when it is confirmed that there is a session extension request, the electronic device 1000 can increase the session extension count by 1.

[0181] In operation S1330, the electronic device 1000 can determine whether the session extension count exceeds α% of the total number of sessions.

[0182] In operation S1340, when it is confirmed that the session extension count exceeds α% of the total number of sessions, the electronic device 1000 can extend the default session time of the response message.

[0183] In operation S1350, when it is confirmed that the session extension count is α% or less of the total number of sessions, the electronic device 1000 can update the session time.

[0184] In operation S1360, when it is confirmed that there is no session extension request in a previous session, the electronic device 1000 can determine whether a maximum value of session use time for a response message is less than β% of a default session time. According to an embodiment of the disclosure, the electronic device 1000 can obtain information about session use time of a plurality of users for a currently output response message from the session use history statistics DB 1320. The electronic device 1000 can identify a maximum value of session use time from the information about session use time of a plurality of users obtained from the session use history statistics DB 1320, and determine whether the maximum value is less than β% (e.g., 50%) of the default session time.

[0185] In operation S1370, when the maximum value of the session use time is less than β%, the electronic device 1000 can shorten the default session time of the response message. The electronic device 1000 can shorten the default session time by, for example, 25%.

[0186] According to Figure 12 and Figure 13 The electronic device 1000 according to the embodiment of the disclosure shown in FIGS. 1 to 6 can store session use history information about other users of the response message, and extend or shorten the default session time by using the stored session use history information of a plurality of users, thereby having a technical effect of adaptively adjusting the default session time by reflecting the experiences of a plurality of users.

[0187] The program executed by the electronic device 1000 described herein can be implemented as hardware components, software components, and / or a combination of hardware components and software components. The program can be executed by any system capable of executing computer-readable instructions.

[0188] The software can include a computer program, code, instructions, or a combination thereof, and can configure a processing device to operate as needed or command the processing device independently or collectively.

[0189] The software can be implemented as a computer program including instructions stored in a computer-readable storage medium. Examples of the computer-readable recording medium can include magnetic storage media (e.g., ROM, RAM, floppy disks, and hard disks), and optical recording media (e.g., compact disc-ROM (CD-ROM) and digital versatile disc (DVD)). The computer-readable recording medium can also be distributed over networked computer systems so as to store and execute the computer-readable code in a distributed manner. The medium can be read by a computer, stored in a memory, and executed by a processor.

[0190] The computer-readable storage medium can be provided in the form of a non-transitory storage medium. The term "non-transitory" simply means that the storage medium is a tangible device, and does not include a signal, but the term does not distinguish between the case where data is semi-permanently stored in the storage medium and the case where data is temporarily stored in the storage medium.

[0191] The program according to the embodiment of the disclosure disclosed in the specification can be included in and provided in a computer program product. The computer program product can be traded between a seller and a buyer as a product.

[0192] The computer program product can include a software (S / W) program and a non-transitory computer-readable recording medium storing the S / W program. For example, the computer program product can include a program module that is downloaded from an application store of an electronic market (e.g., Google Play TMThis refers to software program products (e.g., downloadable applications) electronically distributed by manufacturers of online marketplaces, app stores, etc. For electronic distribution, at least a portion of the software program may be stored on a storage medium or temporarily generated. In this case, the storage medium may be the manufacturer's server, the marketplace's server, or a relay server temporarily storing the software program.

[0193] The computer program product may include the storage medium of a server or the storage medium of a device in a system comprising a server and a device. Alternatively, when a third device (e.g., a smartphone) is present that communicates with the server or device, the computer program product may include the storage medium of the third device. Alternatively, the computer program product may include software programs transferred from the server to the device or the third device, or from the third device to the device.

[0194] In this scenario, one of the server, device, and third device may execute the method according to the present disclosure by executing a computer program product. Alternatively, two or more of the server, device, and third device may execute a computer program product to execute the method according to the present disclosure in a distributed manner.

[0195] For example, a server can execute a computer program product stored on the server to control a device communicatively connected to the server to perform a method according to an embodiment of this disclosure.

[0196] In another example, a third device may execute a computer program product to control a device communicating with the third device to perform a method according to an embodiment of the present disclosure.

[0197] When a third device executes a computer program product, it may download the computer program product from a server and execute the downloaded computer program product. Alternatively, the third device may execute a computer program product provided in a free-loaded state and perform the method according to an embodiment of this disclosure.

[0198] Although this disclosure has been shown and described with reference to various embodiments thereof, those skilled in the art will understand that various changes in form and detail may be made without departing from the spirit and scope of this disclosure as defined by the appended claims and their equivalents.

Claims

1. A method for determining the session duration of a chatbot, the method comprising: Output a response message in response to user input; The default session duration for which the session is maintained is determined based on the difficulty level of the output response message. The additional session time is determined based on the chatbot's dialogue history information before the response message is output; Waiting for user input of additional messages during the session duration, where the session duration is the sum of the default session time and the additional session time; as well as The session ends based on whether the additional message was entered.

2. The method according to claim 1, wherein, Determining the default session time includes: By applying the response message as input to a deep neural network model, a label value for the difficulty level is obtained; and The session time that will be mapped to the obtained tag value will be determined as the default session time.

3. The method according to claim 2, in, The deep neural network model is trained through supervised learning. Among them, multiple training messages are used as input, and The difficulty level marker is used as the output truth value.

4. The method according to claim 1, wherein, The dialogue history information includes information about at least one of the following: the number of previously output response messages in the session before the output of the response message, or the session usage time in the scenario executed based on the previously output response message.

5. The method according to claim 4, wherein, The additional session time is determined by a preset ratio of the sum of the session usage time for each scenario.

6. The method according to claim 1, further comprising: Stores session usage history information for multiple users regarding the output response messages; as well as The default session time is adjusted based on the stored session usage history information of multiple users.

7. The method according to claim 6, wherein, The session usage history information includes statistics for the plurality of users, the statistics for the plurality of users relating to at least one of the following: the number of session extension requests at a time point before the output of the response message, the number of scenarios executed before the response message, or the session usage time for the response message.

8. An electronic device for determining the session duration of a chatbot, the electronic device comprising: The communication interface is configured to send and receive data with other devices. Memory; as well as A processor is configured to execute one or more instructions. The processor is further configured to execute one or more instructions to: Input messages from the user are obtained through the communication interface. Output a response message in response to the user's input message. Based on the difficulty level of the output response message, determine the default session duration for which the session is maintained. Based on the chatbot's conversation history information before outputting the response message, the additional session time is determined. Waiting for user input of additional messages during the session duration, where the session duration is the sum of the default session time and the additional session time, and The session ends based on whether the additional message was entered.

9. The electronic device according to claim 8, wherein, The processor is also configured to execute the one or more instructions to: By applying the response message as input to a deep neural network model, the output label value of the difficulty level is obtained; as well as The session time that will be mapped to the obtained tag value will be determined as the default session time.

10. The electronic device according to claim 9, in, The deep neural network model is trained through supervised learning. Among them, multiple training messages are used as input, and The difficulty level marker is used as the output truth value.

11. The electronic device according to claim 8, wherein, The dialogue history information includes information about at least one of the following: the number of previously output response messages in the session before the output of the response message, or the session usage time in the scenario executed based on the previously output response message.

12. The electronic device according to claim 11, wherein, The processor is also configured to execute one or more instructions to determine the additional session time by a preset ratio of the sum of the session usage times for each scenario.

13. The electronic device according to claim 8, wherein, The processor is also configured to execute the one or more instructions to: The database in the memory stores session usage history information for multiple users regarding the output response messages, and The default session time is adjusted based on the stored session usage history information of multiple users.

14. The electronic device according to claim 13, wherein, The session usage history information includes statistics for the plurality of users, the statistics for the plurality of users relating to at least one of the following: the number of session extension requests at a time point before the output of the response message, the number of scenarios executed before the response message, or the session usage time for the response message.

15. At least one non-transitory computer program product comprising a computer-readable storage medium, wherein, The computer-readable storage medium includes instructions executable by the device for: Output a response message in response to user input; The default session duration for which the session is maintained is determined based on the difficulty level of the output response message. The additional session time is determined based on the chatbot's conversation history information before the response message is output; Waiting for user input of additional messages during the session duration, where the session duration is the sum of the default session time and the additional session time; as well as The session ends based on whether the additional message was entered.

Images