Electronic device and operation method of electronic device

Abstract

This electronic device may include: a memory for storing at least one instruction; and a processor for executing the at least one instruction stored in the memory, wherein the processor executes the least one instruction to: transmit a user input and data related to the type classification of the user input to a server; obtain, from the server, classification information obtained through a large language model (LLM) on the basis of the user input and the data; and perform control to provide a final response corresponding to the user input on the basis of a first response corresponding to the user input based on the obtained classification information and a second response corresponding to the user input obtained from the server through the large language model.

Description

Electronic devices and methods of operating electronic devices

[0001] The present disclosure relates to an electronic device and a method of operating the electronic device. Specifically, the present disclosure relates to an electronic device that provides a response to a user input and a method of operating the electronic device.

[0002] An artificial intelligence system is a computer system that implements human-level intelligence. It is a system in which the machine learns and makes judgments on its own, and its recognition rate improves with use.

[0003] It consists of element technologies that mimic the cognitive and judgment functions of the human brain by utilizing machine learning technology and machine learning algorithms.

[0004] The element technologies may include, for example, at least one of a linguistic understanding technology that recognizes human language / characters, an inference / prediction technology that judges information and logically infers and predicts, and a knowledge representation technology that processes human experience information into knowledge data.

[0005] Recently, with the advancement of artificial intelligence system technology, a technology has been developed and used to extract keywords from user-provided queries and query a database using the extracted keywords to obtain responses to the queries.

[0006] At this time, to obtain the user's desired response, the accuracy of keyword extraction from the query must be high. If keyword extraction from the user's query is not performed properly, the desired response cannot be obtained from the database.

[0007] Additionally, even if the query does not contain clear keywords, the desired response cannot be obtained from the database.

[0008] Additionally, if the query seeks a response that is not included in the stored data stored in the database, an accurate response may not be provided.

[0009] One embodiment of the present disclosure provides an electronic device. The electronic device may include a memory storing at least one instruction. The electronic device may include at least one processor that executes at least one instruction stored in the memory. The at least one processor may transmit data related to a user input and a classification of the type of the user input to a server by executing the at least one instruction. The at least one processor may obtain classification information from the server, obtained through a Large Language Model (LLM), based on the user input and data, by executing the at least one instruction. The at least one processor may control, by executing the at least one instruction, to provide a final response corresponding to the user input based on a first response corresponding to the user input based on the obtained classification information and a second response corresponding to the user response obtained from the server through the Large Language Model.

[0010] One embodiment of the present disclosure may provide a method of operating an electronic device. The method of operating the electronic device may include a step of transmitting data related to a user input and a classification of the type of the user input to a server. The method of operating the electronic device may include a step of obtaining classification information from the server, obtained through a Large Language Model (LLM), based on the user input and the data. The method of operating the electronic device may include a step of controlling the electronic device to provide a final response corresponding to the user input based on a first response corresponding to the user input based on the obtained classification information and a second response corresponding to the user input obtained from the server through the Large Language Model.

[0011] As one embodiment of the present disclosure, a computer-readable recording medium having recorded thereon a program for performing at least one method of the method of operating the disclosed electronic device on a computer may be provided.

[0012] The technical problems to be achieved in this document are not limited to the technical problems mentioned above, and other technical problems not mentioned can be clearly understood by a person having ordinary skill in the technical field to which the present disclosure belongs from the description below.

[0013] The present disclosure may be understood in conjunction with the following detailed description and accompanying drawings, wherein reference numerals refer to structural elements.

[0014] FIG. 1 is a drawing for explaining the operation of an electronic device according to one embodiment of the present disclosure.

[0015] FIG. 2 is a block diagram illustrating the configuration of an electronic device according to one embodiment of the present disclosure.

[0016] FIG. 3 is a flowchart illustrating the operation of an electronic device according to one embodiment of the present disclosure.

[0017] FIG. 4A is a flowchart illustrating the operation of an electronic device providing a final response using a first response and a second response according to one embodiment of the present disclosure.

[0018] FIG. 4b is a flowchart illustrating the operation of an electronic device providing a final response using a first response and a second response according to one embodiment of the present disclosure.

[0019] FIG. 5 is a flowchart illustrating an operation of obtaining an identified final response using a server according to one embodiment of the present disclosure.

[0020] FIG. 6 is a flowchart illustrating an operation of performing second keyword mapping using a server according to one embodiment of the present disclosure.

[0021] FIG. 7 is a flowchart illustrating an operation of performing second keyword mapping using a server according to one embodiment of the present disclosure.

[0022] FIG. 8 is a diagram for explaining an operation for determining whether to provide feedback on a reference score according to one embodiment of the present disclosure.

[0023] FIG. 9 is a block diagram illustrating a second server providing a first response and a first server providing a second response according to one embodiment of the present disclosure.

[0024] FIG. 10 is a drawing for explaining the configuration of an electronic device according to one embodiment of the present disclosure.

[0025] The terms used in this disclosure will be briefly explained, and one embodiment of the present disclosure will be specifically described.

[0026] Throughout this disclosure, the expression "at least one of a or b" refers to a alone, b alone, both a and b, or variations thereof. The expression "at least one of a, b, or c" refers to a alone, b alone, c alone, both a and b, both a and c, both b and c, all a, b, and c, or variations thereof.

[0027] The terms used in this disclosure are selected from widely used, current terms, taking into account the functions of one embodiment of the disclosure. However, these terms may vary depending on the intentions of those skilled in the art, precedents, the emergence of new technologies, etc. Furthermore, in certain cases, terms may be arbitrarily selected by the applicant, and in such cases, their meanings will be described in detail in the description of the relevant embodiments of the disclosure. Therefore, the terms used in this disclosure should not be defined simply as names of terms, but rather based on the meanings of the terms and the overall content of the disclosure.

[0028] Singular expressions may include plural expressions unless the context clearly indicates otherwise. Terms used herein, including technical or scientific terms, have the same meaning as commonly understood by a person of ordinary skill in the art described herein.

[0029] Throughout this disclosure, when a part is said to "include" a component, this does not exclude other components, but rather implies the inclusion of other components, unless otherwise specifically stated. Furthermore, terms such as "part," "module," and the like described herein refer to a unit that processes at least one function or operation, which may be implemented in hardware or software, or a combination of hardware and software.

[0030] The expression “configured to” as used herein can be used interchangeably with, for example, “suitable for,” “having the capacity to,” “designed to,” “adapted to,” “made to,” or “capable of.” The term “configured to” does not necessarily mean something that is “specifically designed to” in terms of hardware. Instead, in some contexts, the expression “a system configured to” can mean that the system is “capable of” doing something together with other devices or components. For example, the phrase “a processor configured (or set) to perform A, B, and C” may mean a dedicated processor (e.g., an embedded processor) for performing those operations, or a generic-purpose processor (e.g., a CPU or application processor) that can perform those operations by executing one or more software programs stored in memory.

[0031] Additionally, when a component is referred to as being “connected” or “connected” to another component in the present disclosure, it should be understood that the component may be directly connected or connected to the other component, but may also be connected or connected via another component in between, unless otherwise specifically stated.

[0032] Below, with reference to the attached drawings, embodiments of the present disclosure are described in detail so that those skilled in the art can easily implement the present disclosure. However, one embodiment of the present disclosure may be implemented in various different forms and is not limited to the embodiments described herein. In addition, in the drawings, parts irrelevant to the description are omitted to clearly describe one embodiment of the present disclosure, and similar parts are designated with similar drawing reference numerals throughout the present disclosure.

[0033] Hereinafter, embodiments of the present disclosure will be described in detail with reference to the drawings.

[0034] FIG. 1 is a drawing for explaining the operation of an electronic device according to one embodiment of the present disclosure.

[0035] Referring to FIG. 1, in one embodiment of the present disclosure, an electronic device (100) may provide a final response (response, 130) corresponding to the user input (110) to a user who has provided user input (110) to the electronic device (100).

[0036] In one embodiment of the present disclosure, when the user input (110) is “Recommend a movie to watch with my girlfriend,” the electronic device (100) may provide a final response (130) related to the user input, “True Love,” which is a response corresponding to the user input (110).

[0037] At this time, the "user input (110)" may be a request that a user using the electronic device (100) inputs into the electronic device (100) by voice or by text, etc. The user input (110) may be data converted into text data by converting audio data that the user inputs into voice to the electronic device (100) through a microphone, etc.

[0038] In one embodiment of the present disclosure, an electronic device (100) may transmit user input (110) provided by a user to a server (200). The server (200) may include a Large Language Model (LLM). A "Large Language Model" may refer to an artificial intelligence model capable of understanding and generating the structure, grammar, and meaning of sentences by learning a large amount of language data.

[0039] In one embodiment of the present disclosure, the large language model may be a model trained based on linguistic understanding that recognizes and applies / processes human language / characters, and thus may be capable of performing operations such as natural language processing, machine translation, conversational systems, question-answering, and speech recognition / synthesis. Furthermore, the large language model may be a model trained based on inference prediction, which logically infers and predicts by judging information, and thus may be capable of performing operations such as knowledge / probability-based inference, optimization prediction, preference-based planning, and recommendation.

[0040] In one embodiment of the present disclosure, a large language model included in the server (200) may be provided with a prompt as input. In this case, a "prompt" may refer to an input value provided to a generative artificial intelligence model so that the generative artificial intelligence model performs a specific task.

[0041] In one embodiment of the present disclosure, the prompt may include natural language text.

[0042] In one embodiment of the present disclosure, a prompt may refer to an input value provided to a large language model, which is one of the generative artificial intelligence models, to perform a specific task. A prompt may refer to an input value for commanding, directing, questioning, or requesting the large language model to perform a specific action.

[0043] In one embodiment of the present disclosure, user input (110) may be transmitted to the server (200) as a prompt provided by the user. Hereinafter, the user input (110) may be provided as a prompt to the large language model.

[0044] In one embodiment of the present disclosure, the electronic device (100) may provide a user input (110) to a large language model included in a server (200) and request that the large language model generate a response (210) corresponding to the user input (110).

[0045] In one embodiment of the present disclosure, the electronic device (100) can obtain a response (210) generated for a user input (110) using a large language model included in the server (200).

[0046] Additionally, the large language model included in the server (200) may be provided with data as input, including not only user input (110), but also information for performing a specific action set in advance using the user input (110).

[0047] In one embodiment of the present disclosure, a prompt explaining in natural language what action the giant language model should take may be provided as data to the giant language model, thereby outputting a desired result from the giant language model.

[0048] In one embodiment of the present disclosure, the electronic device (100) may provide a large language model included in the server (200) with not only user input (110), but also data set to perform a specific action with respect to the user input (110).

[0049] In one embodiment of the present disclosure, the electronic device (100) may transmit to the server (200) as data a prompt set to classify the user input (110) into one of a plurality of preset types corresponding to the type of the user input (110).

[0050] In one embodiment of the present disclosure, the electronic device (100) may include a plurality of response generation modules corresponding to a plurality of types of user input. Each of the plurality of response generation modules may be a module used to generate a corresponding response from among a plurality of responses stored in a database included in the electronic device (100) according to the type of user input.

[0051] In one embodiment of the present disclosure, the electronic device (100) can obtain classification information that classifies the type of user input (110) using a large language model from a server (200). Based on the classification information obtained from the server (200), the electronic device (100) can provide a response (120) corresponding to the user input (110). Using the obtained classification information, the electronic device (100) can provide a response (120) corresponding to the user input (110) using a response generation module corresponding to the type of the user input (110) among a plurality of response generation modules.

[0052] In one embodiment of the present disclosure, the electronic device (100) may generate a final response (130) to be provided to the user using the response (210) obtained from the server (200) and the response (120) obtained from the electronic device (100). Hereinafter, the response (120) generated by the electronic device (100) may be referred to as the first response (120). The response (210) obtained from the server (200) may be referred to as the second response (210). The electronic device (100) may generate a final response (130) to be provided to the user using the first response (120) and the second response (210).

[0053] In one embodiment of the present disclosure, the electronic device (100) can obtain a response list based on the first response (120) and the second response (210). The “response list” may be generated by adding the first response (120) and the second response (210).

[0054] In one embodiment of the present disclosure, the electronic device (100) may provide the server (200) with a response list and identification data set to identify a final response (130) corresponding to a user input (110) from the response list. At this time, the identification data may be set to identify the final response (130) corresponding to the user input (110) from the response list and may be provided as a prompt to the macro language model. In one embodiment of the present disclosure, the electronic device (100) may obtain the final response (130) identified from the response list from the server (200) using the macro language model. The electronic device (100) may provide the final response (130) obtained from the server (200) to the user.

[0055] Through the electronic device (100) of the present invention, the intent of the user input (110) provided by the user can be accurately identified. Furthermore, the accuracy of the final response (130) provided to the user in response to the user input (110) can be improved and provided to the user. This can improve the user experience of the user using the electronic device (100).

[0056] In one embodiment of the present disclosure, the electronic device (100) may be implemented as an electronic device of various shapes, such as a smart phone, a laptop computer, a tablet PC, a television, a wearable device, a head mounted display device, a digital signage, a personal computer (PC), etc.

[0057] Furthermore, in one embodiment of the present disclosure, the electronic device (100) may be implemented as an electronic device of various shapes, such as a desktop, a set-top box, or a server device. In this case, the electronic device (100) may receive user input (110) from an external electronic device including a display or audio connected through an input / output interface, and may provide the generated final response (130) to the external electronic device or an external server. Furthermore, the electronic device (100) may receive user input (110) from an external server through a communication interface (190), and may provide the generated final response (130) to the external server.

[0058] FIG. 2 is a block diagram illustrating the configuration of an electronic device according to one embodiment of the present disclosure.

[0059] Referring to FIGS. 1 and 2, in one embodiment of the present disclosure, an electronic device (100) may include a display (140), a memory (150), at least one processor (160), a cache memory (170), an input / output interface (180), and a communication interface (190).

[0060] However, not all of the components illustrated in FIG. 2 are essential components. The electronic device (100) may be implemented with more components than those illustrated in FIG. 2, or may be implemented with fewer components.

[0061] The display (140), memory (150), at least one processor (160), cache memory (170), input / output interface (180), and communication interface (190) may each be electrically and / or physically connected to each other.

[0062] In one embodiment of the present disclosure, the display (140) may include any one of a liquid crystal display, a plasma display, an organic light emitting diode display, and an inorganic light emitting diode display. However, the present disclosure is not limited thereto, and the display (140) may include other types of displays capable of displaying images.

[0063] In one embodiment of the present disclosure, at least one processor (160) may display and provide the generated final response (130) to the user through the display (140). In addition, at least one processor (160) may provide the user with content corresponding to the obtained final response (130) through the display (140).

[0064] In one embodiment of the present disclosure, if the final response (130) is the title of a specific movie, drama, TV program, or book, at least one processor (160) may display the final response (130) through the display (140) and may also display the movie, drama, TV program, or book corresponding to the final response (130). In addition, if the final response (130) refers to a specific web page, game application, or weather application, at least one processor (160) may display the final response (130) through the display (140) and may also execute and display the specific web page, game application, or weather application corresponding to the final response (130).

[0065] However, the present disclosure is not limited thereto, and at least one processor (160) may display only content corresponding to the generated final response (130) on the display (140) and provide it to the user.

[0066] In one embodiment of the present disclosure, memory (150) may store instructions, data structures, and program codes that can be read by at least one processor (160). Operations performed by at least one processor (160) may be implemented by executing instructions or codes of a program stored in memory (150). In one embodiment of the present disclosure, there may be one or more memories (150).

[0067] In one embodiment of the present disclosure, the memory (150) may include at least one of a flash memory type, a hard disk type, a multimedia card micro type, a card type memory (e.g., SD or XD memory, etc.), a RAM (Random Access Memory), a SRAM (Static Random Access Memory), a ROM (Read-Only Memory), an EEPROM (Electrically Erasable Programmable Read-Only Memory), a PROM (Programmable Read-Only Memory), a Mask ROM, a Flash ROM, etc.), a hard disk drive (HDD), or a solid state drive (SSD).

[0068] In one embodiment of the present disclosure, the memory (150) may store instructions or program codes for performing functions or operations of the electronic device (100). The instructions, algorithms, data structures, program codes, and application programs stored in the memory (150) may be implemented in a programming or scripting language such as, for example, C, C++, Java, or an assembler.

[0069] In one embodiment of the present disclosure, a memory (150) may store a cache control module (151), a response generation module (152), a response list generation module (156), a response identification module (157), and a response provision module (158). The response generation module (152) may include a first generation module (153), a second generation module (154), and a third generation module (155).

[0070] However, not all modules illustrated in FIG. 2 are essential modules. The memory (150) may store more or fewer modules than the modules illustrated in FIG. 2.

[0071] In one embodiment of the present disclosure, a 'module' included in the memory (150) may mean a unit that processes a function or operation performed by at least one processor (160). The 'module' included in the memory (150) may be implemented as software such as instructions, an algorithm, a data structure, or a program code.

[0072] In one embodiment of the present disclosure, the cache control module (151) may be configured with instructions or program codes relating to an operation or function for determining whether cache data stored in the cache memory (170) is valid.

[0073] In one embodiment of the present disclosure, “cache data” may be data stored in a cache memory (170), which may be data that has been pre-requested from at least one processor (160) and provided to at least one processor (160) from a server (200). The cache data may be a second response (210) that has been pre-obtained from the server (200) and stored in the cache memory (170) in response to a user input (110) provided by at least one processor (160).

[0074] In one embodiment of the present disclosure, the cache control module (151) may be configured with instructions or program codes relating to an operation or function of checking whether a second response (210) corresponding to the user input (110) is stored in the cache memory (170) as cache data before providing the user input (110) to the server (200) to request a second response (210).

[0075] In one embodiment of the present disclosure, the cache control module (151) may be configured with instructions or program codes relating to an operation or function of searching the header of at least one cache data stored in the cache memory (170) to determine whether cache data corresponding to the same prompt as the user input (110) to be requested is stored in the cache memory (170).

[0076] In one embodiment of the present disclosure, the cache control module (151) may be configured with instructions or program codes relating to an operation or function of checking whether cache data stored in the cache memory (170) is valid before providing a user input (110) to the server (200) to request a second response (210).

[0077] In one embodiment of the present disclosure, the cache control module (151) may be configured with instructions or program codes related to an operation or function of searching the header of cache data stored in the cache memory (170) and checking whether the validity period of the cache data has expired, etc. to check whether the cache data is valid.

[0078] In one embodiment of the present disclosure, the cache control module (151) may be configured with commands or program codes related to an operation or function of reading cache data stored in a cache memory (170).

[0079] In one embodiment of the present disclosure, at least one processor (160) can check whether a second response (210) corresponding to a user input (110) is included as cache data in a cache memory (170) by executing instructions or program code of a cache control module (151).

[0080] In one embodiment of the present disclosure, at least one processor (160) may read the second response (210) from the cache memory (170) by executing instructions or program codes of the cache control module (151) upon determining that the cache data includes the second response (210) corresponding to the user input (110). At least one processor (160) may obtain the cache data stored in the cache memory (170) as the second response (210) upon determining that the cache data includes the second response (210) corresponding to the user input (110).

[0081] In one embodiment of the present disclosure, the response generation module (152) may be configured with instructions or program codes relating to an operation or function of obtaining, as a first response (120), the stored data corresponding to the user input (110) from among a plurality of stored data stored in a database included in the electronic device (100).

[0082] In one embodiment of the present disclosure, a database included in an electronic device (100) may include a non-relational database. The database may include a plurality of stored data stored in a key-value manner. In this case, a key may correspond to a keyword, and a value may correspond to a response corresponding to the keyword. The response generation module (152) may be configured with commands or program codes related to an operation or function of retrieving a response corresponding to a keyword included in a user input (110) from the database.

[0083] However, the present disclosure is not limited thereto, and the database included in the electronic device (100) may include a relational, hierarchical, network, or object-oriented type, etc.

[0084] In one embodiment of the present disclosure, the memory (150) may include a natural language processing module including a natural language processing (NLP) model. At least one processor (160) may perform embedding on a user input (110) using the natural language processing module. At least one processor (160) may perform preprocessing, such as tokenizing the user input (110), using the natural language processing module, and perform embedding, which converts the preprocessed user input into a vector representation.

[0085] In one embodiment of the present disclosure, a plurality of stored data stored in a database may be data that has undergone embedding in a vector representation. The response generation module (152) may be configured with commands or program codes related to an operation or function of bringing a value corresponding to a key having a high similarity to a vectorized user input (110) among the plurality of stored data stored in the database included in the electronic device (100) as a first response (120).

[0086] At least one processor (160) may generate, as a first response (120), stored data corresponding to a user input (110) from among a plurality of stored data stored in a database included in the electronic device (100) by executing instructions or program codes of the response generation module (152). In this case, generating the first response (120) may mean an operation of retrieving and acquiring a response corresponding to a keyword included in the user input (110) from the database.

[0087] In one embodiment of the present disclosure, the type of user input (110) may include a first type in which the user input (110) includes a first keyword that is preset. The type of user input (110) may include a second type in which the user input (110) includes a second keyword that is preset. The type of user input (110) may include a third type in which the user input (110) does not include the first keyword or the second keyword.

[0088] In one embodiment of the present disclosure, the database included in the electronic device (100) may include a first database storing responses corresponding to a first keyword, a second database storing responses corresponding to a second keyword, and a third database storing responses corresponding to keywords other than the first and second keywords. However, the present disclosure is not limited thereto, and it goes without saying that the responses corresponding to the first keyword, the responses corresponding to the second keyword, and the responses corresponding to other keywords may be stored in a single database.

[0089] In one embodiment of the present disclosure, the response generation module (152) may include a first generation module (153), a second generation module (154), and a third generation module (155).

[0090] The first generation module (153) may be configured with commands or program codes relating to an operation or function of retrieving a response corresponding to a user input (110) from a first database in order to obtain a first response (120) corresponding to a first type of user input (110).

[0091] The second generation module (154) may be configured with commands or program codes for an operation or function of retrieving a response corresponding to a user input (110) from a second database in order to obtain a first response (120) corresponding to a second type of user input (110).

[0092] The third generation module (155) may be configured with commands or program codes for an operation or function of retrieving a response corresponding to a user input (110) from a third database in order to obtain a first response (120) corresponding to a third type of user input (110).

[0093] At least one processor (160) can obtain a first response (120) by executing commands or program codes of a generation module corresponding to the type of user input (110) among the first to third generation modules (154, 154, 155) using classification information obtained from the server (200) according to the type of user input (110).

[0094] Hereinafter, the first to third types and the first to third generation modules (153, 154, 155) will be described later in FIGS. 6 and 7.

[0095] In one embodiment of the present disclosure, the response list generation module (156) may be configured with commands or program codes related to an operation or function of obtaining a response list based on a first response (120) obtained through the response generation module (152) and a second response (210) obtained from a server (200).

[0096] At this time, the "response list" may be generated by adding the first response (120) and the second response (210). Contents included in both the first response (120) and the second response (210) may be generated so as not to overlap in the response list.

[0097] In one embodiment of the present disclosure, the first response (120) and the second response (210) may each include at least one response corresponding to the user input (110). The response list may include at least one response corresponding to the user input (110).

[0098] At least one processor (160) can obtain a response list by executing instructions or program codes of a response list generation module (156) using a first response (120) generated in an electronic device (100) and a second response (210) obtained from a server (200).

[0099] In one embodiment of the present disclosure, the response identification module (157) may be configured with commands or program codes relating to an operation or function for identifying a final response (130) corresponding to a user input (110) among the generated response list.

[0100] In one embodiment of the present disclosure, the response identification module (157) can identify the response with the highest correlation to the user input (110) among the multiple responses included in the generated response list as the final response (130) corresponding to the user input (110).

[0101] In one embodiment of the present disclosure, the response identification module (157) may be configured with instructions or program codes relating to an operation or function of performing embedding into a vector representation each of a plurality of responses included in a response list, and identifying a response that is adjacent to the vector representation of the user input (110) as a final response (130) corresponding to the user input (110).

[0102] However, the present disclosure is not limited thereto, and the response identification module (157) may be configured with commands or program codes relating to operations or functions that utilize other preset methods to identify a response highly related to the user input (110) among multiple responses as the final response (130).

[0103] In addition, the operation performed through the response identification module (157) may be performed using a large language model included in the server (200). At least one processor (160) may transmit to the server (200) the generated response list and identification data set to identify the final response (130) corresponding to the user input (110) from the response list. At least one processor (160) may obtain the final response (130) identified in the response list from the server (200) using the large language model. At this time, the identification data may be provided to the large language model as a prompt instructing the large language model to identify the final response (130) corresponding to the user input (110) from the response list.

[0104] In one embodiment of the present disclosure, the response providing module (158) may be configured with instructions or program codes relating to an action or function that provides a final response (130) identified through the response identification module (157) to the user.

[0105] In one embodiment of the present disclosure, the response providing module (158) may be configured with instructions or program codes relating to an operation or function of controlling the display (140) to display the acquired final response (130). The response providing module (158) may be configured with instructions or program codes relating to an operation or function of controlling the display (140) to display content corresponding to the acquired final response (130).

[0106] At least one processor (160) can control the display (140) to display the acquired final response (130) or content corresponding to the final response (130) by executing instructions or program codes of the response providing module (158).

[0107] In addition, when an audio device such as a speaker is included in the electronic device (100), the response provision module (158) may be configured with commands or program codes regarding an operation or function that controls the audio device so that the generated final response (130) is provided as sound. The response provision module (158) may be configured with commands or program codes regarding an operation or function that converts the generated final response (130) into sound data and provides it to the user through the audio device.

[0108] At least one processor (160) can control a speaker to provide the acquired final response (130) as sound by executing instructions or program codes of the response providing module (158).

[0109] However, the present disclosure is not limited thereto. The response providing module (158) may be configured with commands or program codes relating to an operation or function that controls an input / output interface (180) or a communication interface (190) to provide the acquired final response (130) to an external electronic device including a display or speaker or to an external server.

[0110] At least one processor (160) can control an input / output interface (180) or a communication interface (190) to provide the acquired final response (130) to an external electronic device including a display or speaker or to an external server by executing instructions or program codes of a response providing module (158).

[0111] In one embodiment of the present disclosure, at least one processor (160) may be configured as at least one of a Central Processing Unit, a Microprocessor, a Graphic Processing Unit, an Application Processor (AP), an Application Specific Integrated Circuits (ASICs), a Digital Signal Processors (DSPs), a Digital Signal Processing Devices (DSPDs), a Programmable Logic Devices (PLDs), a Field Programmable Gate Arrays (FPGAs), an Application Processor, a Neural Processing Unit, or an AI-only processor designed with a hardware structure specialized for processing an AI model, but is not limited thereto.

[0112] In one embodiment of the present disclosure, at least one processor (160) may be configured as a processing circuitry such as a System on Chip (SoC) or an Integrated Circuit (IC).

[0113] In one embodiment of the present disclosure, at least one processor (160) can execute various types of modules stored in the memory (150). In one embodiment of the present disclosure, at least one processor (160) can execute at least one module among a cache control module (151), a response generation module (152), a response list generation module (156), a response identification module (157), or a response provision module (158) stored in the memory (150).

[0114] In one embodiment of the present disclosure, at least one processor (160) can execute at least one instruction that constitutes various types of modules stored in the memory (150). The plurality of modules (151, 152, 153, 154, 155, 156, 157, 158) included in the memory (150) may be configurations implemented by at least one processor (160) included in the electronic device (100) executing a program or instruction stored in the memory (150). At least one processor (160) can individually or collectively execute at least one instruction in the memory (150).

[0115] In one embodiment of the present disclosure, at least one processor (160) may include a plurality of processors.

[0116] In one embodiment of the present disclosure, at least one module among the cache control module (151), the response generation module (152), the response list generation module (156), the response identification module (157), or the response provision module (158) stored in the memory (150) may be executed by any one of the plurality of processors. The remaining modules among the cache control module (151), the response generation module (152), the response list generation module (156), the response identification module (157), or the response provision module (158) stored in the memory (150) may be executed by another processor among the plurality of processors.

[0117] In one embodiment of the present disclosure, at least one of the first generation module (153), the second generation module (154), or the third generation module (155) included in the response generation module (152) may be executed by any one of the plurality of processors. The remaining modules of the first generation module (153), the second generation module (154), or the third generation module (155) included in the response generation module (152) may be executed by another processor among the plurality of processors.

[0118] In one embodiment of the present disclosure, a second response (210) obtained from a server (200) via a communication interface (190) may be stored as cache data in the cache memory (170).

[0119] In one embodiment of the present disclosure, data processed by at least one processor (160) by executing instructions or program code of the cache control module (151) may be cache data read and acquired from the cache memory (170).

[0120] In one embodiment of the present disclosure, the cache memory (170) may be a non-volatile memory. The cache memory (170) may include at least one of a flash memory type, a hard disk type, a multimedia card micro type, a card type memory (e.g., an SD or XD memory, etc.), a hard disk drive (HDD), or a solid state drive (SSD).

[0121] In one embodiment of the present disclosure, the cache memory (170) and the memory (150) are illustrated as separate configurations in FIG. 2, but the present disclosure is not limited thereto. The cache memory (170) may also be a configuration included in the memory (150).

[0122] In one embodiment of the present disclosure, the cache memory (170) may be a memory for storing cache data for a certain period of time. The cache memory (170) may temporarily store cache data. The cache memory (170) may be a memory for storing a second response (210) provided from a server (200) as cache data in response to a request from at least one processor (160).

[0123] In one embodiment of the present disclosure, the cache memory (170) may be a memory for storing some of the data stored in the memory (150) as cache data in order to quickly transfer the data to at least one processor (160). The cache memory (170) may be designed to have a structure having a higher layer than the memory (150) and may provide cache data to at least one processor (160).

[0124] In one embodiment of the present disclosure, the input / output interface (180) can receive at least one of image data, audio data, or text data from an external electronic device, etc., under the control of at least one processor (160).

[0125] In one embodiment of the present disclosure, the input / output interface (180) can perform input / output operations with an external electronic device using at least one of input / output methods including an HDMI port (High-Definition Multimedia Interface port), a DVI (Digital Visual Interface), a component jack, a PC port, or a USB port (Universal Serial Bus port). However, the present disclosure is not limited to the above-described input / output methods.

[0126] In one embodiment of the present disclosure, the communication interface (190) can perform data communication with an external server or external electronic device under the control of at least one processor (160).

[0127] In one embodiment of the present disclosure, the communication interface (190) may perform data communication with an external server or an external electronic device using at least one of data communication methods including, for example, wired LAN, wireless LAN, Wi-Fi, Bluetooth, zigbee, Wi-Fi Direct (WFD), infrared Data Association (IrDA), Bluetooth Low Energy (BLE), near field communication (NFC), Wireless Broadband Internet (Wibro), World Interoperability for Microwave Access (WiMAX), Shared Wireless Access Protocol (SWAP), Wireless Gigabit Alliance (WiGig), and RF communication.

[0128] In one embodiment of the present disclosure, the electronic device (100) may further include a user interface capable of obtaining user input. The user interface may receive input from a user using the electronic device (100) under the control of at least one processor (160).

[0129] In one embodiment of the present disclosure, the user interface may include a touch unit, a push button, a voice recognition unit, etc. In one embodiment, the electronic device (100) may obtain a user input (110) provided by a user through the user interface. The electronic device (100) may obtain a user input (110) from a user who touches, presses, provides a voice, or performs an action such as a hand gesture on the user interface.

[0130] In one embodiment of the present disclosure, when an electronic device (100) includes a voice recognition unit (e.g., a microphone), the electronic device (100) can acquire a user input (110) by digitizing an analog voice signal of a user acquired through the voice recognition unit. At least one processor (160) can acquire a user input (110) by digitizing an analog voice signal of a user acquired through the voice recognition unit.

[0131] The electronic device (100) may obtain a user input (110) by converting a user's voice obtained through a voice recognition unit into text. At least one processor (160) may transmit the digitized user input (110) to an external server performing an STT (Speech to Text) operation through a communication interface (190), and may obtain the user input (110) converted into text from the external server through the communication interface (190). However, the present disclosure is not limited thereto, and at least one processor (160) may obtain the user input (110) converted into text from the digitized user input (110) through a module performing an STT operation.

[0132] However, the present disclosure is not limited thereto, and it is obvious that at least one processor (160) may obtain user input (110) based on user input provided using a touch screen or keyboard, etc.

[0133] The present disclosure is not limited thereto, and the electronic device (100) may be connected to an external user interface via an input / output interface (180). In this case, the user interface connected via the input / output interface (180) may include a keyboard, a mouse, a remote control, a microphone, etc. The electronic device (100) may also obtain user input (110) from an external user interface via the input / output interface (180).

[0134] In one embodiment of the present disclosure, the server (200) may include a communication interface (220), a memory (230), and at least one processor (240). The server (200) may be a high-performance computing device, higher than the electronic device (100), capable of processing complex operations and tasks using large amounts of data, such as training, inference, management, and distribution of large-scale language models.

[0135] The server (200) can perform data communication with the electronic device (100) through a communication interface (220) under the control of at least one processor (240).

[0136] The communication interface (220) may include a communication circuit that can perform data communication between the server (3000) and another electronic device (e.g., the electronic device (100)) using at least one of data communication methods including, for example, wired LAN, wireless LAN, Wi-Fi, Bluetooth, ZigBee, Wi-Fi Direct (WFD), infrared Data Association (IrDA), Bluetooth Low Energy (BLE), Near Field Communication (NFC), Wireless Broadband Internet (Wibro), World Interoperability for Microwave Access (WiMAX), Shared Wireless Access Protocol (SWAP), Wireless Gigabit Alliances (WiGig), and RF communication.

[0137] The communication interface (220) can transmit and receive a user input (110) and a second response (210) generated in response to the user input (110) to and from the electronic device (100) under the control of at least one processor (240).

[0138] In one embodiment of the present disclosure, the server (200) may receive user input (110) or embedded user input from the electronic device (100) via the communication interface (220). The server (200) may receive data, calculation data, classification data, and identification data, which will be described later, from the electronic device (100) via the communication interface (220).

[0139] In one embodiment of the present disclosure, the server (200) may provide the electronic device (100) with a second response (210) generated using a large language model included in the memory (230) in response to a user input (110) via a communication interface (220). The server (200) may provide the electronic device (100) with classification information that classifies the type of the user input (110) using the large language model included in the memory (230) via the communication interface (220).

[0140] In one embodiment of the present disclosure, memory (230) may store instructions, data structures, and program codes that can be read by the processor (240). Operations performed by the processor (240) may be implemented by executing instructions or codes of the program stored in memory (230).

[0141] The memory (230) may include a flash memory type, a hard disk type, a multimedia card micro type, a card type memory (e.g., SD or XD memory, etc.), and may include a non-volatile memory including at least one of a ROM (Read-Only Memory), an EEPROM (Electrically Erasable Programmable Read-Only Memory), a PROM (Programmable Read-Only Memory), a magnetic memory, a magnetic disk, and an optical disk, and a volatile memory such as a RAM (Random Access Memory) or an SRAM (Static Random Access Memory).

[0142] In one embodiment of the present disclosure, the memory (230) may store one or more instructions and / or programs that cause the server (200) to operate to classify the type of user input (110) and generate a second response (210) corresponding to the user input (110). Furthermore, the memory (230) may store one or more instructions and / or programs that cause the server (200) to select a final response (130) corresponding to the user input (110) from a list of responses.

[0143] In one embodiment of the present disclosure, the memory (230) may store instructions and / or programs for implementing functions of a large language model.

[0144] In one embodiment of the present disclosure, at least one processor (240) can control the overall operations of the server (200) by executing one or more instructions of a program stored in the memory (230).

[0145] In one embodiment of the present disclosure, at least one processor (240) can control overall operations for obtaining a second response (210) corresponding to a user input (110) by executing one or more instructions of a program stored in a memory (230). At least one processor (240) can control overall operations for classifying a type of a user input (110) by executing one or more instructions of a program stored in a memory (230). At least one processor (240) can control overall operations for identifying a final response (130) corresponding to a user input (110) from a response list by executing one or more instructions of a program stored in a memory (230).

[0146] In one embodiment of the present disclosure, at least one processor (240) may be configured as at least one of a Central Processing Unit, a microprocessor, a Graphic Processing Unit, Application Specific Integrated Circuits (ASICs), Digital Signal Processors (DSPs), Digital Signal Processing Devices (DSPDs), Programmable Logic Devices (PLDs), Field Programmable Gate Arrays (FPGAs), an Application Processor, a Neural Processing Unit, or an AI-only processor designed with a hardware structure specialized for processing an AI model, but is not limited thereto.

[0147] FIG. 3 is a flowchart illustrating the operation of an electronic device according to one embodiment of the present disclosure.

[0148] Referring to FIGS. 1, 2 and 3, in one embodiment of the present disclosure, a method of operating an electronic device (100) may include a step (S100) of obtaining a user input (110) requesting a final response (130).

[0149] In one embodiment of the present disclosure, in the step (S100) of obtaining a user input (110), at least one processor (160) may obtain a user input (110) from a user using an electronic device (100) through an input / output interface (180) or a communication interface (190).

[0150] In one embodiment of the present disclosure, a method of operating an electronic device (100) may include a step (S200) of transmitting data related to a user input (110) and a type classification of the user input (110) to a server (200).

[0151] In one embodiment of the present disclosure, data related to type classification of user input (110) may be data set to classify the user input (110) into one type corresponding to the user input (110) among a plurality of preset types.

[0152] In one embodiment of the present disclosure, in the step (S200) of transmitting user input (110) and data to the server (200), at least one processor (160) may transmit the user input (110) and data to the server (200).

[0153] In one embodiment of the present disclosure, in the step (S200) of transmitting data related to user input (110) and type classification of the user input (110) to the server (200), the data may be transmitted to the server (200) as a prompt.

[0154] Hereinafter, the data will be described in FIGS. 4a to 7.

[0155] In one embodiment of the present disclosure, the operating method of the electronic device (100) may include a step (S300) of acquiring classification information obtained through a large language model based on user input and data from a server (200). At this time, the classification information may include information classifying the type of user input (110).

[0156] In one embodiment of the present disclosure, in the step of obtaining classification information (S300), at least one processor (160) may obtain classification information generated through a large language model from a server (200). At least one processor (160) may classify the type of user input (110) using the obtained classification information.

[0157] Hereinafter, the above-mentioned operations will be described in detail with reference to FIGS. 4a to 7.

[0158] In one embodiment of the present disclosure, the operating method of the electronic device (100) may include a step (S400) of obtaining a first response (120) corresponding to a user input (110) based on the obtained classification information.

[0159] In one embodiment of the present disclosure, in the step (S400) of obtaining a first response (120) corresponding to a user input (110), at least one processor (160) may obtain a first response (120) for the user input (110) using the obtained classification information. At least one processor (160) may generate a first response (120) corresponding to the type of the classified user input (110) using a response generation module (152).

[0160] In one embodiment of the present disclosure, the operating method of the electronic device (100) may include a step (S500) of obtaining a second response (210) corresponding to a user input (110) obtained through a large language model from a server (200).

[0161] In one embodiment of the present disclosure, the method of operating the electronic device (100) may further include a step of providing a user input (110) to the server (200) prior to the step (S500) of obtaining a second response (210) from the server (200). However, the present disclosure is not limited thereto, and the second response (210) may be generated from the server (200) using the user input (110) provided in the step (S200) of transmitting the user input (110) and data to the server (200).

[0162] In one embodiment of the present disclosure, in the step (S500) of obtaining a second response (210) obtained to correspond to a user input (110), at least one processor (160) may provide the user input (110) to the server (200) so that the second response (210) obtained using a large language model may be obtained from the server (200).

[0163] At least one processor (160) can obtain a second response (210) from the server (200) via a communication interface (190).

[0164] In one embodiment of the present disclosure, the method of operating the electronic device (100) may include a step (S600) of providing a final response (130) corresponding to the user input (110) based on the first response (120) and the second response (210).

[0165] In one embodiment of the present disclosure, in the step (S600) of providing a final response (130) corresponding to a user input (110), at least one processor (160) may obtain the final response (130) corresponding to the user input (110) from the first response (120) and the second response (210) using a response list generation module (156) and a response identification module (157).

[0166] In one embodiment of the present disclosure, the method of operating the electronic device (100) may further include a step of providing the acquired final response (130) to the user. At least one processor (160) may control the response providing module (158) to provide the acquired final response (130) to the user who provided the user input (110).

[0167] Although FIG. 3 illustrates multiple operations being performed separately, the present disclosure is not limited thereto. In the operation of the electronic device (100), some of the multiple operations illustrated in FIG. 3 may be omitted, or new operations may be added. Furthermore, it goes without saying that two or more of the multiple operations illustrated in FIG. 3 may be performed together in a single step.

[0168] FIG. 4A is a flowchart illustrating the operation of an electronic device providing a final response using a first response and a second response according to one embodiment of the present disclosure. FIG. 4B is a flowchart illustrating the operation of an electronic device providing a final response using a first response and a second response according to one embodiment of the present disclosure.

[0169] Hereinafter, steps identical to those described in Fig. 3 are given the same drawing reference numerals, and redundant descriptions are omitted.

[0170] Referring to FIGS. 1, 2, 3 and 4A, in one embodiment of the present disclosure, in operation S100, the electronic device (100) may obtain a user input (110).

[0171] In one embodiment of the present disclosure, the user input (110) may be a request for a response including content such as a specific movie, such as "Recommend a movie to watch with my girlfriend." However, the present disclosure is not limited thereto, and the user input (110) may also be a request including a routine question such as "How's the weather today?"

[0172] In operation S200, the electronic device (100) may transmit user input (110) and data to the server (200) (S210). At this time, the data may be a prompt configured to classify the user input (110) into a corresponding one of a plurality of preset types.

[0173] In one embodiment of the present disclosure, a plurality of preset types may be set according to a plurality of responses stored in a database included in the electronic device (100). In one embodiment of the present disclosure, the plurality of types may include a first reference type and a second reference type. The first reference type may be a type that includes a plurality of responses stored in the database. The second reference type may be a type that does not include a plurality of responses stored in the database.

[0174] In one embodiment of the present disclosure, if a plurality of stored data items stored in a database are stored in the form of content (e.g., movies, animations, dramas, etc.) corresponding to keywords, the first reference type may be a type corresponding to a request related to the content. The second reference type may be a type other than content, such as a type corresponding to a request related to weather, dates, politics, or economics.

[0175] In one embodiment of the present disclosure, the plurality of types may include a third reference type and a fourth reference type.

[0176] In one embodiment of the present disclosure, when a plurality of stored data stored in a database are stored in the form of keyword-content (movie, animation, drama, etc.), the keyword may indicate a genre of the content, such as “romance” or “comedy,” or may include words indicating a target audience for viewing the content, such as “girlfriend,” “family,” or “child.”

[0177] In one embodiment of the present disclosure, the third criterion type may be a type including multiple keywords corresponding to multiple contents stored in the database. The third criterion type may be a type including keywords such as "romance" and "girlfriend" corresponding to contents stored in the database.

[0178] In this case, a search request for content like "Search for romantic movies to watch with my girlfriend" has a specific target (e.g., romance, girlfriend, etc.) and includes keywords to describe the content. Therefore, it may fall under the third criterion type.

[0179] In one embodiment of the present disclosure, the fourth criterion type may be a type that likely does not include multiple keywords corresponding to multiple contents stored in the database. The fourth criterion type may include keywords such as "weekend" that are not keywords corresponding to contents stored in the database. In this case, a content recommendation request, such as "Recommend a movie to watch on the weekend," may not include keywords describing the content, as the request target is not specified (e.g., "weekend"). Therefore, the fourth criterion type may be applicable.

[0180] However, since the above example is an embodiment in which multiple stored data included in the database are stored in the form of keyword-content (movies, animations, dramas, etc.), and keywords are stored as words representing the genre or subject of the content, the present disclosure is not limited thereto. The first criterion type, the second criterion type, the third criterion type, and the fourth criterion type may be set differently depending on the form, type, type, etc. of the multiple stored data included in the database.

[0181] That is, the first reference type, the second reference type, the third reference type, and the fourth reference type are preset so that the operations of the electronic device (100) according to the present disclosure are performed separately, and of course, they can be set differently depending on the scenario in which the operations of the electronic device (100) are to be performed separately.

[0182] In one embodiment of the present disclosure, the data may include classification data. The classification data may be data configured to classify the user input (110) into a first type because the user input (110) includes a first keyword that is preset. The classification data may be data configured to classify the user input (110) into a second type because the user input (110) includes a second keyword that is preset. The classification data may be data configured to classify the user input (110) into a third type because the user input (110) does not include the first keyword or the second keyword. The classification data may be provided as a prompt to a large language model included in the server (200).

[0183] At this time, the first keyword, the second keyword, the third keyword, the first type, the second type, and the third type will be described later in FIGS. 6 and 7.

[0184] In operation S200, the electronic device (100) may transmit calculation data set to calculate a score of a user input (110) to the server (200) (S220). In this case, the "score of the user input (110)" may refer to a score that can be used to classify the type of the user input (110) by determining whether the user input (110) includes at least one keyword suitable for generating a final response (130). The calculation data may be provided as a prompt to a large language model included in the server (200).

[0185] In operation S300, the electronic device (100) can obtain classification information classifying the type of user input (110) from the server (200).

[0186] In one embodiment of the present disclosure, the classification information may include information regarding whether the user input (110) is classified into a first criterion type or a second criterion type. The classification information may include information regarding whether the user input (110) is classified into a third criterion type or a fourth criterion type. The classification information may include information regarding whether the user input (110) is classified into a first type, a second type, or a third type.

[0187] Additionally, the classification information may include a score of the user input (110) that serves as a basis for the large language model included in the server (200) to classify the type of the user input (110).

[0188] In one embodiment of the present disclosure, a large language model can classify the type of a user input (110) based on the score of the user input (110). If the score of the user input (110) is equal to or greater than a preset threshold score, the user input (110) can be classified into a first criterion type. If the score of the user input (110) is less than the preset threshold score, the user input (110) can be classified into a second criterion type. In this case, the "threshold score" may be a score that serves as a criterion for the large language model to classify the type of the user input (110) into the first criterion type and the second criterion type. A relationship between the threshold score and the calculated score of the user input (110) can be established through calculation data and transmitted to the server (200).

[0189] Referring to FIGS. 4A and 4B , in operation S310, the operation of the electronic device (100) may vary depending on the classification information. As the user input (110) is classified into the first criterion type, the electronic device (100) may determine whether the calculated score of the user input (110) is greater than or equal to the criterion score (S320). At this time, the "criterion score" may be a score set as a criterion for asking the user whether to provide feedback. The criterion score may be set to a score that is a certain value greater than the threshold score.

[0190] In operation S320, if the score of the user input (110) is equal to or greater than the reference score, the electronic device (100) may not perform an operation of asking the user whether to provide feedback on the threshold score. In operation S320, if the score of the user input (110) is less than the reference score, the electronic device (100) may perform an operation of asking the user whether to provide feedback on the threshold score (S330).

[0191] In operation S330, the electronic device (100) can determine whether a user input providing feedback on a threshold score is obtained from the input / output interface (180) or the communication interface (190).

[0192] In one embodiment of the present disclosure, when a positive user input is provided regarding the score of the calculated user input (110) and the classification thereof into the first criterion type in operation S330, the electronic device (100) may not provide feedback regarding the threshold score. When a negative user input is provided regarding the score of the calculated user input (110) and the classification thereof into the first criterion type in operation S330, the electronic device (100) may provide feedback regarding the threshold score.

[0193] In one embodiment of the present disclosure, if a user input (110) indicates that it is appropriate to classify the user input (110) as a second criterion type rather than a first criterion type, or if a user input indicates that it is appropriate to classify the user input (110) as a first criterion type rather than a second criterion type, the electronic device (100) may determine that a negative user input has been provided. Accordingly, the settings of the calculation data transmitted to the server (200) may be changed to change the value of the threshold score.

[0194] Through operations S320 and S330, the accuracy of the operation of classifying the type of user input (110) using the large language model included in the server (200) can be increased and user optimization can be performed.

[0195] As the user input (110) is classified into the first criterion type, the electronic device (100) can determine whether the cache data obtained in advance from the server (200) includes a second response (210) to the user input (110) (S340).

[0196] In operation S340, the electronic device (100) can check whether an input identical to the user input (110) has been previously transmitted to the server (200) and a second response (210) corresponding to the input has been previously obtained from the server (200) and included in the cache data.

[0197] The electronic device (100) can read the second response (210) from the cache memory (170) storing the cache data, since the second response (210) to the user input (110) is included in the cache data previously acquired from the server (200). Accordingly, the time required to acquire the second response (210) can be reduced compared to the case of transmitting the user input (110) to the server (200) and acquiring the second response (210) generated from the server (200).

[0198] The electronic device (100) may transmit the user input (110) to the server (200) (S510) as the cache data does not include a second response (210) to the user input (110).

[0199] In operation S520, the electronic device (100) can obtain a second response (210) corresponding to the user input (110) generated using a large language model from the server (200).

[0200] In operation S530, the electronic device (100) may store the user input (110) and the acquired second response (210) in the cache memory (170). However, the electronic device (100) may not perform operation S530.

[0201] In operation S400, the electronic device (100) can generate a first response (120) corresponding to the user input (110) using the response generation module (152). Operation S400 will be described below with reference to FIGS. 6 and 7.

[0202] In operation S610, the electronic device (100) can generate a response list based on the first response (120) generated in operation S400 and the second response (210) obtained in operation S520. The electronic device (100) can generate a response list by adding the first response (120) and the second response (210) and removing duplicate responses.

[0203] In operation S620, the electronic device (100) may transmit a response list and identification data to the server (200). At this time, the identification data may be set to identify the final response (130) corresponding to the user input (110) among the response list. The identification data may be set differently depending on the type of user input (110). The identification data may be provided as a prompt to a large language model included in the server (200).

[0204] In one embodiment of the present disclosure, a user input (110) classified as a third criterion type may be a request of a specific target type. Therefore, in this case, the identification data may be set to assign equal weights to the first response (120) and the second response (210) to select a response corresponding to the user input (110) to enhance response accuracy.

[0205] In one embodiment of the present disclosure, a user input (110) classified as a fourth criterion type may be a request for which a target request has not been determined. Accordingly, in this case, the identification data may be set to select a response corresponding to the user input (110) by assigning a high weight to a first response (120) generated from a database included in the electronic device (100) that reflects highly reliable data and the user's usage experience for user optimization, and a low weight to a second response (210).

[0206] In operation S310, as the user input (110) is classified as a second reference type, the electronic device (100) may transmit the user input (110) to the server (200) (S700). In the case of the second reference type, since the request is not related to content stored in a database included in the electronic device (100), the electronic device (100) may not generate a first response (120).

[0207] In operation S800, the electronic device (100) can obtain a second response (210) generated using a large language model from the server (200).

[0208] In operation S900, the electronic device (100) may generate a final response (130) to be ultimately provided to the user, depending on the type of the user input (110). At this time, if the type of the user input (110) is a first reference type, the final response (130) generated in operation S900 may be a response selected in operation S630. If the type of the user input (110) is a second reference type, the final response (130) generated in operation S900 may be a response generated in operation S800.

[0209] FIG. 5 is a flowchart illustrating the operation of obtaining an identified final response using a server according to one embodiment of the present disclosure. Hereinafter, steps identical to those described in FIGS. 4a and 4b are assigned the same reference numerals, and redundant descriptions are omitted.

[0210] Referring to FIG. 1, FIG. 4a, FIG. 4b and FIG. 5, FIG. 5 illustrates an operation in which an electronic device (100) classifies the type of user input (110) through a server (200) and obtains a second response (210) generated in response to the user input (110).

[0211] In one embodiment of the present disclosure, an electronic device (100) may obtain a user input (110) in operation S100. The electronic device (100) may transmit the user input (110) and data to a server (200) in operation S200. The data may include classification data.

[0212] The electronic device (100) may also transmit calculation data to the server (200) in operation S200.

[0213] In operation S301, the server (200) provides the user input (110) and data transmitted to the large language model as inputs, thereby classifying the type of the user input (110). The server (200) can generate classification information including information classifying the type of the user input (110) using the large language model. The server (200) can calculate a score of the user input (110) used in classifying the type of the user input (110) using the user input (110) and calculation data.

[0214] In operation S302, the server (200) can provide classification information to the electronic device (100).

[0215] In operation S310, the electronic device (100) can classify the type of user input (110) using the provided classification information. However, the present disclosure is not limited thereto, and operations S302 and S310 may be performed in a single step.

[0216] In operation S330, the electronic device (100) may provide feedback to the server (200) regarding the score of the user input (110) used to classify the type of the user input (110). If negative feedback is provided regarding the calculated user input (110) to the server (200), the server (200) may recalculate the score of the user input (110) and classify the type of the user input (110) based on the score.

[0217] In operation S340, the electronic device (100) can check whether the second response corresponding to the user input (110) is included in the cache data stored in the cache memory (170).

[0218] In operation S510, the electronic device (100) may transmit the user input (110) to the server (200) as the second response corresponding to the user input (110) is not included in the cache data.

[0219] In operation S521, the server (200) may provide the user input (110) provided to the giant language model as a prompt to generate a second response (210).

[0220] In operation S522, the server (200) can provide the generated second response (210) to the electronic device (100).

[0221] In operation S610, the electronic device (100) can generate a response list using a first response (120) generated in the electronic device (100) and a second response (210) provided from the server (200) in response to a user input (110).

[0222] In operation S620, the electronic device (100) can transmit the generated response list and identification data to the server (200).

[0223] In operation S631, the server (200) provides a response list and identification data as a prompt to the giant language model, so that the final response (130) corresponding to the user input (110) can be identified from the response list.

[0224] In operation S632, the server (200) can provide the identified final response (130) to the electronic device (100). The electronic device (100) can provide the provided final response (130) to the user as a response corresponding to the user input (110).

[0225] FIG. 6 is a diagram illustrating an operation for determining whether to provide feedback on a reference score according to one embodiment of the present disclosure. Hereinafter, steps identical to those described in FIG. 3 are assigned the same reference numerals, and any redundant descriptions are omitted.

[0226] Referring to FIGS. 1, 2, 3, 4a, 4b and 6, in one embodiment of the present disclosure, FIG. 6 illustrates a plurality of operations performed in a step (S400) of generating a first response (120) to a user input (110) using classification information.

[0227] In operation S420, the electronic device (100) may obtain a user input (110). At this time, the user input (110) may be classified into a first reference type.

[0228] In operation S421, the electronic device (100) may obtain classification information. At this time, the classification information may include information that can determine whether the user input (110) is of the third reference type or the fourth reference type. In addition, the classification information may include information that can determine whether the user input (110) is of the first type, the second type, or the third type.

[0229] In operation S430, the electronic device (100) can analyze the type of the user input (110). The electronic device (100) can re-analyze the type of the user input (110) classified as a first reference type, which is a type that includes multiple responses included in the database.

[0230] The preset plurality of types may include a third reference type and a fourth reference type. The electronic device (100) may classify a type of user input (110) classified as a first reference type into a third reference type or a fourth reference type.

[0231] However, the present disclosure is not limited thereto, and the type of user input (110) may be classified into any one of the first to fourth criterion types as a single action.

[0232] In operation S440, the electronic device (100) can check whether a user input (110) classified as a third criterion type includes a first keyword. At this time, the first keyword may mean a keyword that is preset to generate a first response (120) corresponding to the user input (110) using a plurality of responses stored in a database, and if the keyword is included in the user input (110), a response corresponding to the keyword is generated as the first response (120).

[0233] In one embodiment of the present disclosure, the first keyword may be at least one keyword that is directly associated with a corresponding response, such as “title”, “subject”, or “movie actor name”.

[0234] In operation S450, the electronic device (100) may obtain a first response (120) using a first generation module for a user input (110) classified as a first type, including a first keyword. At this time, the first generation module may be a module that generates the first response (120) using the first keyword and a database that stores a response corresponding to the first keyword.

[0235] In operation S460, the electronic device (100) may obtain a second response using a second generation module for a user input (110) classified as a second type, including a second keyword. At this time, the second keyword may mean a keyword that is preset to be obtained as a first response (120) corresponding to the user input (110) when obtaining a first response (120) corresponding to the user input (110) using a plurality of responses stored in a database, when the second keyword is included in a user input (110) that does not include the first keyword, or when there is a keyword with a high similarity to the second keyword.

[0236] In one embodiment of the present disclosure, the second keyword may be at least one keyword that indirectly evokes a corresponding response, such as a genre such as “melodrama,” “action,” or “comedy.”

[0237] In operation S460, the electronic device (100) may transmit the user input (110) and the second keyword to the server (200). The electronic device (100) may transmit mapping data set to map the user input (110), the second keyword, and the keyword with the highest similarity to the keyword included in the user input (110) among the second keywords to the server (200). At this time, the mapping data may be provided as a prompt to a large language model included in the server (200).

[0238] The server (200) can use a large language model to map the keyword with the highest similarity to the keyword included in the user input (110) among the second keywords and provide the same to the electronic device (100).

[0239] In one embodiment of the present disclosure, if the user input (110) is "Search for a movie to watch with my girlfriend," among the second keywords including genres, the keyword "melodrama" may have the highest similarity with the multiple keywords (e.g., "girlfriend," "movie," etc.) included in the user input (110). In this case, "melodrama" may be mapped to the user input (110).

[0240] In operation S470, the electronic device (100) may obtain a first response (120) using a second generation module for a user input (110) classified as a second type, including a second keyword. At this time, the second generation module may be a module that obtains the first response (120) using a database that stores the second keyword and a response corresponding to the second keyword.

[0241] In operation S480, the electronic device (100) can check whether the user input (110) includes a keyword corresponding to the first keyword or the second keyword.

[0242] In operation S480, as it is determined that the user input (110) includes a first keyword, the electronic device (100) can obtain a first response (120) for the user input (110) using the first generation module in operation S491.

[0243] In operation S480, when it is determined that the user input (110) includes a keyword corresponding to the second keyword, the electronic device (100) can perform operations S460 and S470.

[0244] In operation S480, if it is determined that the user input (110) does not include keywords corresponding to the first and second keywords, the electronic device (100) may perform operation S492. The user input (110) determined to not include keywords corresponding to the first and second keywords may be classified into the third type.

[0245] In operation S492, the electronic device (100) may obtain a first response (120) corresponding to the user input (110) using a third generation module. At this time, the third generation module may be a module that generates the first response (120) using the viewing history of the user using the electronic device (100) or the user's preferences.

[0246] In one embodiment of the present disclosure, if the user input (110) is “Recommend something to watch on the weekend”, which is classified as the fourth criterion type, the user input (110) may be classified as the third type because it does not include keywords corresponding to the first and second keywords.

[0247] In this case, the electronic device (100) can use the third generation module to generate content with the highest viewing history or content with the highest preference as the first response (120) by considering the viewing history or preferences of the user who provided the user input (110).

[0248] FIG. 7 is a flowchart illustrating the operation of performing second keyword mapping using a server according to one embodiment of the present disclosure. Hereinafter, steps identical to those described in FIG. 6 are assigned the same reference numerals, and redundant descriptions are omitted.

[0249] Referring to FIGS. 1, 6 and 7, FIG. 7 illustrates an operation of mapping keywords of user input (110) to second keywords using a server (200), where user input (110) is classified into a second type.

[0250] In one embodiment of the present disclosure, the electronic device (100) can classify the type of user input (110) in operation S430. At this time, the user input (110) may be classified into a first reference type. However, the present disclosure is not limited thereto, and in operation S430, the user input (110) may be classified into one of a third reference type or a fourth reference type.

[0251] In one embodiment of the present disclosure, the electronic device (100) can check whether a first keyword is included in the user input (110) in operation S440.

[0252] When the electronic device (100) determines that the user input (110) does not include the first keyword and includes a keyword corresponding to the second keyword, the electronic device (100) can transmit the user input (110), the second keyword, and mapping data to the server (200) in operation S461.

[0253] At this time, the second keyword may include multiple genres such as "melodrama," "action," "drama," "comedy," etc. However, in one embodiment, the present disclosure is not limited thereto, and the second keyword may be composed of words that can be set in advance.

[0254] The server (200) can provide a user input (110), a second keyword, and mapping data as inputs to a large language model, and map at least one keyword among the second keywords that has a high similarity to a keyword included in the user input (110).

[0255] In operation S463, the server (200) may provide a second keyword mapped to the user input (110) to the electronic device (100).

[0256] In operation S470, the electronic device (100) provides a second keyword mapped to the user input (110) to the second generation module, thereby obtaining a first response (120) corresponding to the user input (110).

[0257] FIG. 8 is a diagram for explaining an operation for determining whether to provide feedback on a reference score according to one embodiment of the present disclosure.

[0258] Referring to FIGS. 2 and 8, in one embodiment of the present disclosure, FIG. 8 illustrates a case where a score of a user input (110) is less than a reference score, and the electronic device (100) asks the user whether to provide feedback on the threshold score.

[0259] In one embodiment of the present disclosure, the electronic device (100) can ask the user about the intent of the user input (110) and check whether the type of the user input (110) is classified by the server (200).

[0260] In one embodiment of the present disclosure, the inquiry (800) provided by the electronic device (100) to the user may be a text-based image, such as, "Are you sure you asked about a movie to watch with your girlfriend?" However, the present disclosure is not limited thereto, and the electronic device (100) may also provide the inquiry (800) to the user in an audio format, such as voice.

[0261] The electronic device (100) can receive user feedback regarding an inquiry (800) through a microphone, remote control, speaker, etc. capable of receiving the user's voice. The present disclosure is not limited thereto, and the electronic device (100) can receive user feedback through motion recognition using a touch panel, terminal, camera, etc.

[0262] The electronic device (100) may use the provided user input to provide feedback to the server (200) regarding the classification of the type of the user input (110). Specifically, the electronic device (100) may reflect the feedback in data provided to classify the user input (110) into a certain type and transmit it to the server (200). Accordingly, the threshold score used by the server (200) to classify the type of the user input (110) may be changed.

[0263] The electronic device (100) of the present disclosure can thereby increase the accuracy of the type classification of the user input (110) and increase the accuracy of the generated final response (130) in generating the final response (130) corresponding to the user input (110).

[0264] FIG. 9 is a block diagram illustrating a second server providing a first response and a first server providing a second response according to one embodiment of the present disclosure. Hereinafter, components identical to those described in FIG. 2 are assigned the same reference numerals, and redundant descriptions are omitted.

[0265] Referring to FIGS. 2 and 9, in one embodiment of the present disclosure, an electronic device (900) may include a display (140), a memory (150), at least one processor (160), a cache memory (170), an input / output interface (180), and a communication interface (190). The memory (150) may include a cache control module (151), a response list generation module (156), and a response provision module (158).

[0266] The electronic device (900) can transmit user input (110) and data, etc. to the first server (200). The electronic device (900) can obtain a second response (210) generated from the first server (200).

[0267] In one embodiment of the present disclosure, the first server (200) may include a communication interface (220), a memory (230), and at least one processor (240). In this case, the first server (200) may be a server that performs the same operation as the server (200) described in FIG. 2.

[0268] In one embodiment of the present disclosure, the generation of the first response (120) through the response generation module (913) performed in the aforementioned electronic device (900) may be performed by the second server (910).

[0269] In one embodiment of the present disclosure, the second server (910) may include a communication interface (911), a memory (912), and at least one processor (914). The second server (910) may perform data communication with the electronic device (900) through the communication interface (911).

[0270] In one embodiment of the present disclosure, the communication interface (911) included in the second server (910) may include a communication circuit that can perform data communication between the second server (910) and another electronic device (e.g., the electronic device (900) or the first server (200)) using at least one of data communication methods including, for example, wired LAN, wireless LAN, Wi-Fi, Bluetooth, ZigBee, Wi-Fi Direct (WFD), infrared Data Association (IrDA), Bluetooth Low Energy (BLE), Near Field Communication (NFC), Wireless Broadband Internet (Wibro), World Interoperability for Microwave Access (WiMAX), Shared Wireless Access Protocol (SWAP), Wireless Gigabit Alliances (WiGig), and RF communication.

[0271] The communication interface (911) can transmit and receive data for generating a first response (120) to and from the electronic device (900) under the control of at least one processor (914). In one embodiment of the present disclosure, the second server (910) can receive user input (110) and classification information from the electronic device (900) through the communication interface (911), and transmit the first response (120) generated in response to the user input (110) to the electronic device (900).

[0272] Memory (912) may store instructions, data structures, and program codes that can be read by at least one processor (914). Operations performed by at least one processor (914) may be implemented by executing instructions or codes of a program stored in memory (912).

[0273] The memory (912) may include a flash memory type, a hard disk type, a multimedia card micro type, a card type memory (e.g., SD or XD memory, etc.), and may include a non-volatile memory including at least one of a ROM (Read-Only Memory), an EEPROM (Electrically Erasable Programmable Read-Only Memory), a PROM (Programmable Read-Only Memory), a magnetic memory, a magnetic disk, and an optical disk, and a volatile memory such as a RAM (Random Access Memory) or a SRAM (Static Random Access Memory).

[0274] The memory (912) may store one or more instructions and / or programs that operate to cause the server (910) to generate a first response (120) using user input (110) and classification information. For example, the memory (912) may store instructions and / or programs for implementing functions of the response generation module (913). The response generation module (913) may include a database in which responses corresponding to keywords of the first type, the second type, and the third type are stored.

[0275] Since the description related to the operations of the response generation module (913) has already been described in the description of the previous drawings, a repeated description is omitted.

[0276] At least one processor (914) included in the second server (910) can control the overall operations of the second server (910). For example, at least one processor (914) can execute one or more instructions of a program stored in the memory (912).

[0277] At least one processor (914) may be configured as, but is not limited to, at least one of, for example, a Central Processing Unit, a microprocessor, a Graphic Processing Unit, Application Specific Integrated Circuits (ASICs), Digital Signal Processors (DSPs), Digital Signal Processing Devices (DSPDs), Programmable Logic Devices (PLDs), Field Programmable Gate Arrays (FPGAs), an Application Processor, a Neural Processing Unit, or an AI-specific processor designed with a hardware structure specialized for processing an AI model.

[0278] FIG. 10 is a drawing for explaining the configuration of an electronic device according to one embodiment of the present disclosure. Hereinafter, configurations identical to those described in FIG. 2 are assigned the same reference numerals, and redundant descriptions are omitted.

[0279] Referring to FIGS. 2 and 10, in one embodiment of the present disclosure, the operations performed by the electronic device (100) and the server (200) described above can be performed solely by the electronic device (100).

[0280] An electronic device (100) may include a display (140), a memory (1010), at least one processor (160), a cache memory (170), an input / output interface (180), and a communication interface (190). The memory (1010) may include an input type classification module (1011), a first response generation module (1012), a second response generation module (1013), a cache control module (151), a response list generation module (156), a response identification module (157), and a response provision module (158).

[0281] In one embodiment of the present disclosure, the operations performed in the aforementioned server (200) may be independently performed in the electronic device (100) by the input type classification module (1011), the second response generation module (1013), and the response identification module (157). At this time, since the electronic device (100) has relatively low computing performance compared to the server (200), the input type classification module (1011), the second response generation module (1013), and the response identification module (157) used by the electronic device (100) may be generative language models that have been lightweighted to be customized for the computing performance of the electronic device (1000).

[0282] To solve the above-described technical problem, in one embodiment of the present disclosure, an electronic device is provided. The electronic device may include a memory storing at least one instruction. The electronic device may include at least one processor that executes at least one instruction stored in the memory. The at least one processor may transmit data related to a user input and a classification of the type of the user input to a server by executing the at least one instruction. The at least one processor may obtain classification information from the server through a Large Language Model (LLM) based on the user input and data by executing the at least one instruction. The at least one processor may control the execution of the at least one instruction to provide a final response corresponding to the user input based on a first response corresponding to the user input based on the received classification information and a second response corresponding to the user input obtained from the server through the Large Language Model.

[0283] In one embodiment of the present disclosure, the data may be data configured to classify a user input into one of a plurality of preset types corresponding to the user input. At least one processor may execute at least one command to obtain a first response using a response generation module corresponding to the type of user input among a plurality of response generation modules corresponding to each of the plurality of types based on the classification information.

[0284] In one embodiment of the present disclosure, at least one processor may provide a second response as a final response based on classification information by executing at least one command, based on which a type of user input does not correspond to a plurality of response generation modules.

[0285] In one embodiment of the present disclosure, the plurality of types may include a first type including a first preset keyword, a second type including a second preset keyword, and a third type not including the first keyword or the second keyword. At least one processor may obtain a first response by using a first generation module corresponding to the first type among the plurality of response generation modules when the user input is classified into the first type based on classification information by executing at least one command. At least one processor may obtain a first response by using a second generation module corresponding to the second type among the plurality of response generation modules when the user input is classified into the second type based on classification information by executing at least one command. At least one processor may obtain a first response corresponding to the third type by executing at least one command when the user input is classified into the third type based on classification information.

[0286] In one embodiment of the present disclosure, the data may include classification data. At least one processor may, by executing at least one command, classify the user input into a first type if the user input includes a first keyword. At least one processor may, by executing at least one command, classify the user input into a second type if the user input does not include the first keyword and includes a keyword corresponding to the second keyword. At least one processor may, by executing at least one command, transmit classification data set to classify the user input into a third type if the user input does not include a keyword corresponding to the first keyword or the second keyword to the server. At least one processor may, by executing at least one command, obtain classification information from the server, which classifies the user input acquired through the large language model into one of the first type, the second type, or the third type.

[0287] In one embodiment of the present disclosure, at least one processor may, by executing at least one command, determine whether a user input includes at least one keyword suitable for providing a final response, and transmit calculation data corresponding to the calculation of a score of the user input to a server. At least one processor may, by executing at least one command, obtain a score obtained through a large-scale language model based on the user input and the calculation data from the server. At least one processor may, by executing at least one command, transmit feedback regarding the score to the server.

[0288] In one embodiment of the present disclosure, an electronic device may include a cache memory that stores cache data previously acquired from a server. At least one processor may, by executing at least one command, acquire a second response from the cache memory if the cache data includes a second response to a user input. At least one processor may, by executing at least one command, acquire a second response acquired through a large-scale language model from the server if the cache data does not include the second response.

[0289] In one embodiment of the present disclosure, at least one processor may obtain a response list based on the first response and the second response by executing at least one command. At least one processor may select a final response corresponding to the user input from the response list by executing at least one command and provide the final response.

[0290] In one embodiment of the present disclosure, at least one processor may transmit to the server a response list and identification data configured to identify a final response corresponding to a user input from the response list by executing at least one command. At least one processor may obtain from the server a final response identified from the response list using a large-scale language model based on the response list and the identification data by executing at least one command.

[0291] In one embodiment of the present disclosure, the user input may be a prompt requesting a search or recommendation for specific content. The final response may include specific content corresponding to the user input.

[0292] To solve the above-described technical problem, one embodiment of the present disclosure provides a method of operating an electronic device. The method of operating the electronic device may include a step of transmitting data related to a user input and a classification of the type of the user input to a server. The method of operating the electronic device may include a step of obtaining classification information from the server, obtained through a Large Language Model (LLM), based on the user input and the data. The method of operating the electronic device may include a step of controlling the electronic device to provide a final response corresponding to the user input based on a first response corresponding to the user input based on the received classification information and a second response corresponding to the user input obtained from the server through the Large Language Model.

[0293] In one embodiment of the present disclosure, the data may be data configured to classify a user input into one of a plurality of preset types corresponding to the user input. The step of providing a final response may include a step of obtaining a first response using a response generation module corresponding to the type of user input among a plurality of response generation modules corresponding to each of the plurality of types, based on the classification information.

[0294] In one embodiment of the present disclosure, the step of providing a final response may include the step of providing a second response as the final response based on the classification information, when the type of the user input does not correspond to a plurality of response generation modules.

[0295] In one embodiment of the present disclosure, the plurality of types may include a first type including a preset first keyword, a second type including a preset second keyword, and a third type that does not include the first keyword and the second keyword. In the step of obtaining the first response, when the user input is classified into the first type according to the classification information, the first response may be obtained using a first generation module corresponding to the first type among the plurality of response generation modules. In the step of obtaining the first response, when the user input is classified into the second type according to the classification information, the first response may be obtained using a second generation module corresponding to the second type among the plurality of response generation modules. In the step of obtaining the first response, when the user prompt is classified into the third type according to the classification information, the first response may be obtained using a third generation module corresponding to the third type among the plurality of response generation modules.

[0296] In one embodiment of the present disclosure, the system data may include classification data. In the step of providing the system data to the server, classification data may be provided to the server, wherein the user input is classified into a first type if the user input includes a first keyword, the user input is classified into a second type if the user input does not include the first keyword and includes a keyword corresponding to the second keyword, and the user input is classified into a third type if the user input does not include a keyword corresponding to the first or second keyword. In the step of obtaining classification information, classification information may be obtained from the server, wherein the user input obtained through the large language model is classified into one of the first type, the second type, or the third type.

[0297] In one embodiment of the present disclosure, a method of operating an electronic device may include a step of determining whether a user input includes at least one keyword suitable for providing a final response, and transmitting calculation data corresponding to calculating a score of the user input to a server. The method of operating the electronic device may include a step of obtaining a score obtained through a large language model based on the user input and the calculation data. The method of operating the electronic device may include a step of transmitting feedback regarding the score to the server.

[0298] In one embodiment of the present disclosure, a method of operating an electronic device may include a step of determining whether cache data previously acquired from a server includes a second response to a user input. If the cache data includes the second response to the user input, the method of operating the electronic device may include a step of acquiring the second response from a cache memory that stores the cache data. If the cache data does not include the second response, the method of operating the electronic device may include a step of acquiring the second response acquired through a large language model from the server.

[0299] In one embodiment of the present disclosure, the step of providing a final response may include obtaining a list of responses based on the first response and the second response. The step of generating a final response may include identifying a final response corresponding to the user input from the list of responses and providing the final response.

[0300] In one embodiment of the present disclosure, the step of providing a final response may include transmitting to the server a response list and identification data configured to identify a final response corresponding to the user input from the response list. The step of generating a final response may include obtaining from the server a final response identified from the response list using a large-scale language model based on the response list and the identification data.

[0301] In order to solve the above-described technical problem, a computer-readable recording medium having recorded thereon a program for performing at least one method of an embodiment of an operating method of an electronic device disclosed in the present disclosure on a computer can be provided.

[0302] The program executed by the electronic device described in this disclosure may be implemented as hardware components, software components, and / or a combination of hardware components and software components. The program may be executed by any system capable of executing computer-readable instructions.

[0303] Software may include a computer program, code, instructions, or a combination of one or more of these, which may configure a processing device to do a desired thing or may independently or collectively command a processing device to do a desired thing.

[0304] Software may be implemented as a computer program containing instructions stored on a computer-readable storage medium. Examples of computer-readable storage media include magnetic storage media (e.g., read-only memory (ROM), random-access memory (RAM), floppy disks, hard disks, etc.) and optical readable media (e.g., CD-ROMs, DVDs (Digital Versatile Discs)). The computer-readable storage media may be distributed across network-connected computer systems, so that computer-readable code may be stored and executed in a distributed manner. The storage media may be readable by a computer, stored in a memory, and executed by a processor.

[0305] Computer-readable storage media may be provided in the form of non-transitory storage media. Here, the term "non-transitory storage media" simply means a tangible device that does not contain signals (e.g., electromagnetic waves). This term does not distinguish between cases where data is permanently stored in the storage media and cases where data is temporarily stored. For example, a "non-transitory storage medium" may include a buffer in which data is temporarily stored.

[0306] Additionally, programs according to the embodiments disclosed herein may be provided as part of a computer program product. The computer program product may be traded as a commodity between sellers and buyers.

[0307] A computer program product may include a software program and a computer-readable storage medium storing the software program. For example, a computer program product may include a product in the form of a software program (e.g., a downloadable application) distributed electronically by an electronic device manufacturer or through an electronic marketplace (e.g., the Samsung Galaxy Store). For electronic distribution, at least a portion of the software program may be stored on a storage medium or temporarily created. In this case, the storage medium may be a server of the electronic device manufacturer, a server of the electronic marketplace, or a storage medium of an intermediary server that temporarily stores the software program.

[0308] Although the embodiments described above have been described with limited examples and drawings, those skilled in the art will appreciate that various modifications and variations can be made based on the above description. For example, appropriate results can still be achieved even if the described techniques are performed in a different order than described, and / or components such as the described computer system or modules are combined or combined in a different manner than described, or are replaced or substituted with other components or equivalents.

Claims

1. In an electronic device (100), A memory (150) storing at least one instruction; and At least one processor (160) comprising a processing circuit, The electronic device (100) executes the at least one processor (160) individually or collectively the at least one instruction stored in the memory (150), Transmitting data related to user input and classification of the type of said user input to the server (200), Obtain classification information obtained through a Large Language Model (LLM) based on the user input and the data from the server (200), An electronic device (100) that controls to provide a final response corresponding to the user input based on a first response corresponding to the user input based on the acquired classification information and a second response corresponding to the user input acquired from the server (200) through the large language model.

2. In paragraph 1, The above data is data set to classify the user input into one of a plurality of preset types corresponding to the user input, The above electronic device (100) is, An electronic device (100) that obtains the first response by using a response generation module corresponding to the type of the user input among a plurality of response generation modules corresponding to each of the plurality of types based on the classification information.

3. In paragraph 2, The above electronic device (100) is, An electronic device (100) that provides the second response as the final response based on the classification information, as the type of the user input does not correspond to the plurality of response generation modules.

4. In either of the second or third paragraphs, The above plurality of types include a first type including a preset first keyword, a second type including a preset second keyword, and a third type not including the first keyword and the second keyword, The above electronic device (100) is, Based on the classification information, when the user input is classified into the first type, the first response is obtained using the first generation module corresponding to the first type among the plurality of response generation modules, Based on the classification information, when the user input is classified into the second type, the first response is obtained using a second generation module corresponding to the second type among the plurality of response generation modules, An electronic device (100) that obtains the first response by using a third generation module corresponding to the third type among the plurality of response generation modules based on the classification information, when the user input is classified into the third type.

5. In paragraph 4, The above data includes classification data, The above electronic device (100) is, The classification data set to classify the user input into the first type when the first keyword is included in the user input, classify the user input into the second type when the first keyword is not included in the user input and a keyword corresponding to the second keyword is included, and classify the user input into the third type when the keyword corresponding to the first keyword and the second keyword is not included in the user input is transmitted to the server (200). An electronic device (100) that obtains classification information, obtained through the large language model, from the server (200) that classifies the user input into one of the first type, the second type, or the third type.

6. In any one of the clauses 1 to 5, The above electronic device (100) is, Determine whether the user input includes at least one keyword suitable for providing the final response and transmit calculation data corresponding to the score calculation of the user input to the server (200); Obtaining the score obtained through the large language model based on the user input and the calculation data from the server (200), An electronic device (100) that transmits feedback on the score to the server (200).

7. In any one of the clauses 1 to 6, The electronic device (100) further includes a cache memory that stores cache data obtained in advance from the server (200), The above electronic device (100) is, As the second response to the user input is included in the cache data, the second response is obtained from the cache memory, An electronic device (100) that obtains the second response obtained through the large language model from the server (200) as the second response is not included in the cache data.

8. In any one of the clauses 1 to 7, The above electronic device (100) is, Obtain a response list based on the first response and the second response, An electronic device (100) that identifies a final response corresponding to the user input from the obtained response list and provides the final response.

9. In paragraph 8, The above electronic device (100) is, Transmitting the above response list and the identification data set to identify the final response corresponding to the user input among the above response list to the server (200), An electronic device (100) that obtains the final response identified from the response list from the server (200) using the large language model based on the response list and the identification data.

10. In any one of the clauses 1 to 9, The above user input is a prompt requesting a search or recommendation of specific content, The final response is an electronic device (100) that includes the specific content corresponding to the user input.

11. In the operating method of an electronic device (100), A step (S200) of transmitting data related to user input and type classification of the user input to a server; A step (S300) of obtaining classification information obtained through a large language model (LLM) based on the user input and the data from the server; and An operating method of an electronic device (100), comprising a step (S600) of controlling to provide a final response corresponding to the user input based on a first response corresponding to the user input based on the acquired classification information and a second response corresponding to the user input acquired from the server through the large language model.

12. In paragraph 11, The above data is data set to classify the user input into one of a plurality of preset types corresponding to the user input, The step (S600) of providing the final response above is: An operating method of an electronic device (100) further comprising a step of obtaining the first response by using a response generation module corresponding to the type of the user input among a plurality of response generation modules corresponding to each of the plurality of types, according to the classification information.

13. In paragraph 12, The step (S600) of providing the final response above is: A method of operating an electronic device (100) further comprising the step of providing the second response as the final response according to the classification information, when the type of the user input does not correspond to the plurality of response generation modules.

14. In any one of paragraphs 12 or 13, The above plurality of types include a first type including a preset first keyword, a second type including a preset second keyword, and a third type not including the first keyword and the second keyword, In the step of obtaining the above first response, According to the above classification information, when the user input is classified into the first type, the first response is obtained using the first generation module corresponding to the first type among the plurality of response generation modules, According to the above classification information, when the user input is classified into the second type, the first response is obtained using a second generation module corresponding to the second type among the plurality of response generation modules, An operating method of an electronic device (100) for obtaining the first response by using a third generation module corresponding to the third type among the plurality of response generation modules according to the classification information, when the user input is classified into the third type.

15. A computer-readable recording medium having recorded thereon a program for performing the method described in any one of Articles 11 to 14 on a computer.

Images