WO2025084597A1 - Device and method for generating personalized model, and storage medium - Google Patents

Abstract

This electronic device may comprise: a memory for storing instructions; a communication circuit; and at least one processor. When executed individually or collectively by the at least one processor, the instructions can cause the electronic device to: obtain a reference model from a server for federated learning; generate a first model for the service from the reference model by using user input data for using the service; transmit, to the server, first information indicating the service; in response to the first information, obtain, from the server, second information indicating external electronic devices using the service; obtain a second model for the service of the external electronic device from the external electronic device with which a connection is established on the basis of the second information; and, on the basis of the difference between the first model and the second model, which is less than a reference value, generate a third model for the service by using the first model and the second model.

Description

Device, method, and storage medium for generating personalized models

The descriptions below relate to a device, method, and storage medium for generating a personalized model.

The electronic device can collect data based on a user's input. The electronic device can generate a model using the data based on the user's input. For example, the model can include an artificial intelligence model (AI model). For example, the electronic device can generate (or learn) the model based on federated learning. For example, the federated learning can refer to a technique in which an external electronic device connected to the electronic device and the electronic device cooperate to generate the model.

The above information may be provided as related art for the purpose of assisting in understanding the present disclosure. No claim or determination is made as to whether any of the above is applicable as prior art related to the present disclosure.

An electronic device as described above may include a memory storing instructions. The electronic device may include a communication circuit. The electronic device may include a processor. The instructions, when executed by the processor, may cause the electronic device to obtain a reference model from a server for federated learning via the communication circuit. The instructions, when executed by the processor, may cause the electronic device to generate a first model for a service from the reference model using user input data for use of the service. The instructions, when executed by the processor, may cause the electronic device to transmit first information instructing the service to the server. The instructions, when executed by the processor, may cause the electronic device to obtain, from the server, second information instructing an external electronic device to use the service, in response to the first information. The instructions, when executed by the processor, may cause the electronic device to obtain, from the external electronic device with which a connection has been established based on the second information, a second model for the service of the external electronic device. The instructions, when executed by the processor, may cause the electronic device to generate a third model for the service using the first model and the second model, based on a difference between the first model and the second model being less than a reference value.

An electronic device as described above may include a memory storing instructions. The electronic device may include a communication circuit. The electronic device may include a processor. The instructions, when executed by the processor, may cause the electronic device to transmit a reference model generated for federated learning to each of a first external electronic device, a second external electronic device, and a third external electronic device. The instructions, when executed by the processor, may cause the electronic device to obtain a first model for a service generated from the reference model from the first external electronic device, a second model for the service generated from the reference model from the second external electronic device, and a third model for the service generated from the reference model from the third external electronic device, via the communication circuit. The instructions, when executed by the processor, may cause the electronic device to recognize a first difference between the first model and the second model and a second difference between the first model and the third model. The instructions, when executed by the processor, may cause the electronic device to generate a fourth model for the service from the first model using the second model based on the first difference being less than a threshold value and the second difference being greater than or equal to the threshold value.

In a method performed by an electronic device as described above, the method may include an operation of obtaining a reference model from a server for federated learning. The method may include an operation of generating a first model for the service from the reference model using input data of a user for using the service. The method may include an operation of transmitting first information indicating the service to the server. The method may include an operation of obtaining, in response to the first information, second information indicating an external electronic device using the service from the server. The method may include an operation of obtaining, from the external electronic device with which a connection is established based on the second information, a second model for the service of the external electronic device. The method may include an operation of generating a third model for the service using the first model and the second model based on a difference between the first model and the second model that is less than a reference value.

A non-transitory computer-readable storage medium as described above may store one or more programs including instructions that, when executed by a processor of an electronic device including a communication circuit, cause the electronic device to obtain a reference model from a server for federated learning via the communication circuit. The non-transitory computer-readable storage medium may store one or more programs including instructions that, when executed by the processor, cause the electronic device to generate a first model for the service from the reference model using user input data for use of the service. The non-transitory computer-readable storage medium may store one or more programs including instructions that, when executed by the processor, cause the electronic device to transmit first information indicative of the service to the server. The non-transitory computer-readable storage medium may store one or more programs including instructions that, when executed by the processor, cause the processor to, in response to the first information, obtain second information from the server, which indicates an external electronic device that uses the service. The non-transitory computer-readable storage medium may store one or more programs including instructions that, when executed by the processor, cause the processor to obtain a second model for the service of the external electronic device from the external electronic device with which a connection has been established based on the second information. The non-transitory computer-readable storage medium may store one or more programs including instructions that, when executed by the processor, cause the processor to generate a third model for the service using the first model and the second model based on a difference between the first model and the second model that is less than a reference value.

A method performed by an electronic device as described above may include an operation in which the electronic device transmits a reference model generated for federated learning to each of a first external electronic device, a second external electronic device, and a third external electronic device. The method may include an operation in which a first model for a service generated from the reference model is acquired from the first external electronic device, a second model for the service generated from the reference model is acquired from the second external electronic device, and a third model for the service generated from the reference model is acquired from the third external electronic device, via the communication circuit. The method may include an operation in which a first difference between the first model and the second model and a second difference between the first model and the third model are recognized. The method may include an operation in which a fourth model for the service is generated from the first model using the second model, based on the first difference being less than a reference value and the second difference being greater than or equal to the reference value.

A non-transitory computer-readable storage medium as described above may store one or more programs including instructions that, when executed by a processor of an electronic device including a communication circuit, cause the electronic device to transmit a reference model generated for federated learning to each of a first external electronic device, a second external electronic device, and a third external electronic device. The non-transitory computer-readable storage medium may store one or more programs including instructions that, when executed by the processor, cause the electronic device to obtain a first model for a service generated from the reference model from the first external electronic device, a second model for the service generated from the reference model from the second external electronic device, and a third model for the service generated from the reference model from the third external electronic device, via the communication circuit. The non-transitory computer-readable storage medium may store one or more programs including instructions that, when executed by the processor, cause the processor to recognize a first difference between the first model and the second model and a second difference between the first model and the third model. The non-transitory computer-readable storage medium may store one or more programs including instructions that, when executed by the processor, cause the processor to generate a fourth model for the service from the first model using the second model based on the first difference being less than a threshold value and the second difference being greater than or equal to the threshold value.

FIG. 1 is a block diagram of an electronic device within a network environment according to various embodiments.

FIG. 2 illustrates an example of a method for an electronic device to generate a personalized model based on an external electronic device, according to one embodiment.

FIG. 3 illustrates an example of a network for generating personalized models based on peer to peer (P2P), according to one embodiment.

FIG. 4 illustrates an example of a signal flow for a method of generating a personalized model based on P2P, according to one embodiment.

FIGS. 5A and 5B illustrate an example of a method for learning a model of an electronic device using a model obtained from an external electronic device, according to one embodiment.

FIG. 6 illustrates an example of an operational flow for a method of generating a personalized model based on P2P, according to one embodiment.

FIG. 7 illustrates an example of a network for generating a personalized model based on a server, according to one embodiment.

FIG. 8 illustrates an example signal flow for a method for generating a personalized model based on a server, according to one embodiment.

FIG. 9 illustrates an example of an operational flow for a method of generating a personalized model based on a server, according to one embodiment.

The terms used in this disclosure are only used to describe specific embodiments and may not be intended to limit the scope of other embodiments. The singular expression may include the plural expression unless the context clearly indicates otherwise. The terms used herein, including technical or scientific terms, may have the same meaning as commonly understood by a person having ordinary skill in the art described in this disclosure. Among the terms used in this disclosure, terms defined in general dictionaries may be interpreted as having the same or similar meaning in the context of the related technology, and shall not be interpreted in an ideal or excessively formal meaning unless explicitly defined in this disclosure. In some cases, even if a term is defined in this disclosure, it cannot be interpreted to exclude embodiments of the present disclosure.

In the various embodiments of the present disclosure described below, a hardware-based approach is described as an example. However, since the various embodiments of the present disclosure include techniques using both hardware and software, the various embodiments of the present disclosure do not exclude a software-based approach.

In addition, in the present disclosure, in order to determine whether a specific condition is satisfied or fulfilled, the expression "more than" or "less than" may be used, but this is only a description for expressing an example and does not exclude the description of more than or less than. A condition described as "more than" may be replaced with "more than," a condition described as "less than" may be replaced with "less than," and a condition described as "more than and less than" may be replaced with "more than and less than." In addition, hereinafter, "A" to "B" mean at least one of the elements from A to (including A) and from (including B).

FIG. 1 is a block diagram of an electronic device within a network environment according to various embodiments.

Referring to FIG. 1, in a network environment (100), an electronic device (101) may communicate with an electronic device (102) via a first network (198) (e.g., a short-range wireless communication network), or may communicate with at least one of an electronic device (104) or a server (108) via a second network (199) (e.g., a long-range wireless communication network). According to one embodiment, the electronic device (101) may communicate with the electronic device (104) via the server (108). According to one embodiment, the electronic device (101) may include a processor (120), a memory (130), an input module (150), an audio output module (155), a display module (160), an audio module (170), a sensor module (176), an interface (177), a connection terminal (178), a haptic module (179), a camera module (180), a power management module (188), a battery (189), a communication module (190), a subscriber identification module (196), or an antenna module (197). In some embodiments, the electronic device (101) may omit at least one of these components (e.g., the connection terminal (178)), or may have one or more other components added. In some embodiments, some of these components (e.g., the sensor module (176), the camera module (180), or the antenna module (197)) may be integrated into one component (e.g., the display module (160)).

The processor (120) may control at least one other component (e.g., a hardware or software component) of the electronic device (101) connected to the processor (120) by executing, for example, software (e.g., a program (140)), and may perform various data processing or calculations. According to one embodiment, as at least a part of the data processing or calculations, the processor (120) may store a command or data received from another component (e.g., a sensor module (176) or a communication module (190)) in the volatile memory (132), process the command or data stored in the volatile memory (132), and store result data in the nonvolatile memory (134). According to one embodiment, the processor (120) may include a main processor (121) (e.g., a central processing unit or an application processor) or an auxiliary processor (123) (e.g., a graphic processing unit, a neural processing unit (NPU), an image signal processor, a sensor hub processor, or a communication processor) that can operate independently or together therewith. For example, if the electronic device (101) includes a main processor (121) and a secondary processor (123), the secondary processor (123) may be configured to use lower power than the main processor (121) or to be specialized for a given function. The secondary processor (123) may be implemented separately from the main processor (121) or as a part thereof.

The auxiliary processor (123) may control at least a portion of functions or states associated with at least one of the components of the electronic device (101) (e.g., the display module (160), the sensor module (176), or the communication module (190)), for example, on behalf of the main processor (121) while the main processor (121) is in an inactive (e.g., sleep) state, or together with the main processor (121) while the main processor (121) is in an active (e.g., application execution) state. In one embodiment, the auxiliary processor (123) (e.g., an image signal processor or a communication processor) may be implemented as a part of another functionally related component (e.g., a camera module (180) or a communication module (190)). In one embodiment, the auxiliary processor (123) (e.g., a neural network processing device) may include a hardware structure specialized for processing artificial intelligence models. The artificial intelligence models may be generated through machine learning. Such learning may be performed, for example, in the electronic device (101) itself on which the artificial intelligence model is executed, or may be performed through a separate server (e.g., server (108)). The learning algorithm may include, for example, supervised learning, unsupervised learning, semi-supervised learning, or reinforcement learning, but is not limited to the examples described above. The artificial intelligence model may include a plurality of artificial neural network layers. The artificial neural network may be one of a deep neural network (DNN), a convolutional neural network (CNN), a recurrent neural network (RNN), a restricted Boltzmann machine (RBM), a deep belief network (DBN), a bidirectional recurrent deep neural network (BRDNN), deep Q-networks, or a combination of two or more of the above, but is not limited to the examples described above. In addition to the hardware structure, the artificial intelligence model may additionally or alternatively include a software structure.

The memory (130) can store various data used by at least one component (e.g., processor (120) or sensor module (176)) of the electronic device (101). The data can include, for example, software (e.g., program (140)) and input data or output data for commands related thereto. The memory (130) can include volatile memory (132) or nonvolatile memory (134).

The program (140) may be stored as software in the memory (130) and may include, for example, an operating system (142), middleware (144), or an application (146).

The input module (150) can receive commands or data to be used in a component of the electronic device (101) (e.g., a processor (120)) from an external source (e.g., a user) of the electronic device (101). The input module (150) can include, for example, a microphone, a mouse, a keyboard, a key (e.g., a button), or a digital pen (e.g., a stylus pen).

The audio output module (155) can output an audio signal to the outside of the electronic device (101). The audio output module (155) can include, for example, a speaker or a receiver. The speaker can be used for general purposes such as multimedia playback or recording playback. The receiver can be used to receive an incoming call. According to one embodiment, the receiver can be implemented separately from the speaker or as a part thereof.

The display module (160) can visually provide information to an external party (e.g., a user) of the electronic device (101). The display module (160) can include, for example, a display, a holographic device, or a projector and a control circuit for controlling the device. According to one embodiment, the display module (160) can include a touch sensor configured to detect a touch, or a pressure sensor configured to measure the intensity of a force generated by the touch.

The audio module (170) can convert sound into an electrical signal, or vice versa, convert an electrical signal into sound. According to one embodiment, the audio module (170) can obtain sound through an input module (150), or output sound through an audio output module (155), or an external electronic device (e.g., an electronic device (102)) (e.g., a speaker or a headphone) directly or wirelessly connected to the electronic device (101).

The sensor module (176) can detect an operating state (e.g., power or temperature) of the electronic device (101) or an external environmental state (e.g., user state) and generate an electric signal or data value corresponding to the detected state. According to one embodiment, the sensor module (176) can include, for example, a gesture sensor, a gyro sensor, a barometric pressure sensor, a magnetic sensor, an acceleration sensor, a grip sensor, a proximity sensor, a color sensor, an IR (infrared) sensor, a biometric sensor, a temperature sensor, a humidity sensor, or an illuminance sensor.

The interface (177) may support one or more designated protocols that may be used to directly or wirelessly connect the electronic device (101) with an external electronic device (e.g., the electronic device (102)). According to one embodiment, the interface (177) may include, for example, a high definition multimedia interface (HDMI), a universal serial bus (USB) interface, an SD card interface, or an audio interface.

The connection terminal (178) may include a connector through which the electronic device (101) may be physically connected to an external electronic device (e.g., the electronic device (102)). According to one embodiment, the connection terminal (178) may include, for example, an HDMI connector, a USB connector, an SD card connector, or an audio connector (e.g., a headphone connector).

The haptic module (179) can convert an electrical signal into a mechanical stimulus (e.g., vibration or movement) or an electrical stimulus that a user can perceive through a tactile or kinesthetic sense. According to one embodiment, the haptic module (179) can include, for example, a motor, a piezoelectric element, or an electrical stimulation device.

The camera module (180) can capture still images and moving images. According to one embodiment, the camera module (180) can include one or more lenses, image sensors, image signal processors, or flashes.

The power management module (188) can manage power supplied to the electronic device (101). According to one embodiment, the power management module (188) can be implemented as, for example, at least a part of a power management integrated circuit (PMIC).

The battery (189) can power at least one component of the electronic device (101). In one embodiment, the battery (189) can include, for example, a non-rechargeable primary battery, a rechargeable secondary battery, or a fuel cell.

The communication module (190) may support establishment of a direct (e.g., wired) communication channel or a wireless communication channel between the electronic device (101) and an external electronic device (e.g., the electronic device (102), the electronic device (104), or the server (108)), and performance of communication through the established communication channel. The communication module (190) may operate independently from the processor (120) (e.g., the application processor) and may include one or more communication processors that support direct (e.g., wired) communication or wireless communication. According to one embodiment, the communication module (190) may include a wireless communication module (192) (e.g., a cellular communication module, a short-range wireless communication module, or a GNSS (global navigation satellite system) communication module) or a wired communication module (194) (e.g., a local area network (LAN) communication module or a power line communication module). Among these communication modules, a corresponding communication module may communicate with an external electronic device (104) via a first network (198) (e.g., a short-range communication network such as Bluetooth, wireless fidelity (WiFi) direct, or infrared data association (IrDA)) or a second network (199) (e.g., a long-range communication network such as a legacy cellular network, a 5G network, a next-generation communication network, the Internet, or a computer network (e.g., a LAN or WAN)). These various types of communication modules may be integrated into a single component (e.g., a single chip) or implemented as multiple separate components (e.g., multiple chips). The wireless communication module (192) may use subscriber information (e.g., international mobile subscriber identity (IMSI)) stored in the subscriber identification module (196) to identify or authenticate the electronic device (101) within a communication network such as the first network (198) or the second network (199).

The wireless communication module (192) can support a 5G network and next-generation communication technology after a 4G network, for example, NR access technology (new radio access technology). The NR access technology can support high-speed transmission of high-capacity data (eMBB (enhanced mobile broadband)), terminal power minimization and connection of multiple terminals (mMTC (massive machine type communications)), or high reliability and low latency (URLLC (ultra-reliable and low-latency communications)). The wireless communication module (192) can support, for example, a high-frequency band (e.g., mmWave band) to achieve a high data transmission rate. The wireless communication module (192) may support various technologies for securing performance in a high-frequency band, such as beamforming, massive multiple-input and multiple-output (MIMO), full dimensional MIMO (FD-MIMO), array antenna, analog beam-forming, or large scale antenna. The wireless communication module (192) may support various requirements specified in an electronic device (101), an external electronic device (e.g., electronic device (104)), or a network system (e.g., second network (199)). According to one embodiment, the wireless communication module (192) may support a peak data rate (e.g., 20 Gbps or more) for eMBB realization, a loss coverage (e.g., 164 dB or less) for mMTC realization, or a U-plane latency (e.g., 0.5 ms or less for downlink (DL) and uplink (UL) each, or 1 ms or less for round trip) for URLLC realization.

The antenna module (197) can transmit or receive signals or power to or from the outside (e.g., an external electronic device). According to one embodiment, the antenna module (197) can include an antenna including a radiator formed of a conductor or a conductive pattern formed on a substrate (e.g., a PCB). According to one embodiment, the antenna module (197) can include a plurality of antennas (e.g., an array antenna). In this case, at least one antenna suitable for a communication method used in a communication network, such as the first network (198) or the second network (199), can be selected from the plurality of antennas by, for example, the communication module (190). A signal or power can be transmitted or received between the communication module (190) and the external electronic device through the selected at least one antenna. According to some embodiments, in addition to the radiator, another component (e.g., a radio frequency integrated circuit (RFIC)) can be additionally formed as a part of the antenna module (197).

According to various embodiments, the antenna module (197) may form a mmWave antenna module. According to one embodiment, the mmWave antenna module may include a printed circuit board, an RFIC positioned on or adjacent a first side (e.g., a bottom side) of the printed circuit board and capable of supporting a designated high-frequency band (e.g., a mmWave band), and a plurality of antennas (e.g., an array antenna) positioned on or adjacent a second side (e.g., a top side or a side) of the printed circuit board and capable of transmitting or receiving signals in the designated high-frequency band.

At least some of the above components may be interconnected and exchange signals (e.g., commands or data) with each other via a communication method between peripheral devices (e.g., a bus, a general purpose input and output (GPIO), a serial peripheral interface (SPI), or a mobile industry processor interface (MIPI)).

In one embodiment, commands or data may be transmitted or received between the electronic device (101) and an external electronic device (104) via a server (108) connected to a second network (199). Each of the external electronic devices (102 or 104) may be the same or a different type of device as the electronic device (101). In one embodiment, all or part of the operations executed in the electronic device (101) may be executed in one or more of the external electronic devices (102, 104, or 108). For example, when the electronic device (101) is to perform a certain function or service automatically or in response to a request from a user or another device, the electronic device (101) may, instead of or in addition to executing the function or service itself, request one or more external electronic devices to perform at least a part of the function or service. One or more external electronic devices that receive the request may execute at least a part of the requested function or service, or an additional function or service related to the request, and transmit the result of the execution to the electronic device (101). The electronic device (101) may process the result as is or additionally and provide it as at least a part of a response to the request. For this purpose, for example, cloud computing, distributed computing, mobile edge computing (MEC), or client-server computing technology may be used. The electronic device (101) may provide an ultra-low latency service by using, for example, distributed computing or mobile edge computing. In another embodiment, the external electronic device (104) may include an IoT (Internet of Things) device. The server (108) may be an intelligent server using machine learning and/or a neural network. According to one embodiment, the external electronic device (104) or the server (108) may be included in the second network (199). The electronic device (101) can be applied to intelligent services (e.g., smart home, smart city, smart car, or healthcare) based on 5G communication technology and IoT-related technology.

An artificial intelligence model (or a deep learning model) may collect user data to improve accuracy and usability. For example, the artificial intelligence model may be trained (or generated, updated) based on the user's data. In this case, if the user's data is related to the user's personal information, the collection of the data for training the artificial intelligence model may be limited. Hereinafter, the artificial intelligence model may be referred to as a model.

The learning of the artificial intelligence model may include federated learning (FL). For example, in the FL, a server for the FL may generate (or learn) a model. The model initially generated by the server may be referred to as a base model or a reference model. The generated reference model may be transmitted from the server to electronic devices connected to the server. Each of the electronic devices may use a service based on the received reference model. For example, the service may be referred to as an AI service using the reference model. For example, the service may be provided based on the execution of a specific software application. Each of the electronic devices may collect data of the user according to the use of the service, and perform learning (or updating) of the model based on the collected data. The model learned based on the data may be a personalized model with respect to the user. After the learning is completed, each of the electronic devices can transmit parameters for the updated model to the server. For example, the parameters can include vector values that configure (or define) the updated model. For example, the parameters can represent information that configures the model. The server can generate a new model (or a reference model) based on the models received from each of the electronic devices. As described above, by using the FL, the model can be learned using the actual usage information of the user while protecting the personal information of the user. Accordingly, the model can be advanced.

In the above example, an example is described in which a new reference model is generated in the server based on models of the electronic devices transmitted to the server, and one of the electronic devices, other than the server, may update the model of the electronic device based on a model received from at least one electronic device different from the electronic device among the electronic devices. The FL performed based on the communication between the electronic devices as described above may be referred to as a P2P (peer to peer) FL.

In the case of the P2P FL, the electronic device can learn (or generate, update) the model of the electronic device based on a model (or parameters or vector values of the model) obtained from at least one electronic device different from the electronic device. For example, the electronic device can use the model of the at least one electronic device that has a relatively low relevance to a user of the electronic device (or a service used by the user, or usage information of the user) for learning. In addition, considering that the information used for the learning is a parameter of the model, preprocessing such as filtering or distinction for the learning cannot be used. Accordingly, the electronic device can perform indiscriminate learning based on the model (or parameters of the model) of the at least one electronic device.

For example, if the user of the electronic device is right-handed, it is assumed that the electronic device learns a model by using the usage information (e.g., keyboard input) of the right-handed user when using the service of the electronic device. The service may be a service (e.g., keyboard input service or keyboard input function) in which the electronic device obtains input from the user through a keyboard. In this case, if there is a left-handed user among the users of the at least one electronic device, the electronic device may learn the model of the electronic device based on the usage information of the left-handed user. In other words, the electronic device may learn the model based on information that is not suitable for the user of the electronic device. In this case, the learned model may be a distorted model with respect to the user of the electronic device.

Hereinafter, an electronic device, a method, and a storage medium according to embodiments of the present disclosure can group models generated based on usage data of other users that are relevant to a user when using FL. The electronic device, the method, and the storage medium according to embodiments of the present disclosure can generate (or learn, update) a personalized model for the user by generating (or learning, updating) a model using the grouped model. Accordingly, the electronic device, the method, and the storage medium according to embodiments of the present disclosure can provide a service with improved usability for the user by using a more personalized model for the user. In addition, the electronic device, the method, and the storage medium according to embodiments of the present disclosure can effectively protect the personal information of the user when generating a more personalized model for the user.

FIG. 2 illustrates an example of a method for an electronic device to generate a personalized model based on an external electronic device, according to one embodiment.

The electronic device (101) of FIG. 2 may represent an example of the electronic device (101) of FIG. 1. For example, in FIG. 2, an example of an electronic device (101) that is a smartphone is described, but the embodiments of the present disclosure are not limited thereto. For example, the electronic device (101) may be implemented by software or hardware and software, such as a server.

FIG. 2 illustrates examples (201, 202, 203) of a method for learning a model of an electronic device (101) and providing a service through the learned model based on other models related to the service while the electronic device (101) provides a service based on the model. In the examples (201, 202, 203) of FIG. 2, the service may include a voice recognition service. However, the embodiments of the present disclosure are not limited thereto.

In example (201), according to one embodiment, the electronic device (101) can provide a service. For example, the electronic device (101) can provide the service based on a reference model. For example, the reference model can be obtained from a server for FL. For example, the service can include a voice recognition service.

According to one embodiment, the electronic device (101) can obtain a voice (210) spoken by a user (205). For example, the electronic device (101) can obtain the voice (210) using a microphone (e.g., the input module (150) of FIG. 1). For example, the voice (210) can be “AAA”. After obtaining the voice (210), the electronic device (101) can perform voice recognition on the voice (210) based on the reference model. For example, the voice recognition can represent a function of recognizing a text representing the voice (210) spoken by the user (205). The electronic device (101) can recognize a text (220) corresponding to the voice (210) by performing the voice recognition. For example, the electronic device (101) may display a visual object representing the recognized text (220) through a display (e.g., the display module (160) of FIG. 1). For example, the text (220) according to the speech recognition performed based on the reference model may be “BBB.” For example, “BBB” may be different from “AAA,” which is a speech (210) spoken by the user (205).

According to one embodiment, the electronic device (101) may obtain an input (225) from a user (205) who recognizes a text (220) different from a voice (210). For example, the input (225) may represent an input of the user (205) for modifying the text (220) different from the voice (210) to be identical to the voice (210). For example, the input (225) may represent input data of the user (205) obtained while the electronic device (101) uses the service (or the voice recognition service) provided to the user (205). The input data may represent data used for learning the reference model of the electronic device (101). For example, the input (225) may be used for learning the reference model. For example, the input (225) may include text (230) (e.g., “AAA”) input by a body part (e.g., a finger) of the user (205) or an external device (e.g., a stylus pen), a body part of the user (205) or a body part in contact with the external device for inputting the text (230), or information about a method by which the user (205) inputs the text (230). For example, the method may include handwriting, a location of the input (225), and an input speed. The electronic device (101) may recognize the text (230) based on the input (225) as a result of voice recognition of the voice (210). In other words, the electronic device (101) may recognize the result of the voice recognition for the voice (210) as text (230) rather than text (220).

In example (201), an example is illustrated in which an input (225) for correcting an incorrectly recognized text (220) into text (230) is used for learning (or training, adaptive training, generation, update) of the reference model, but the embodiments of the present disclosure are not limited thereto. Unlike example (201), when the electronic device (101) acquires a voice (210) and accurately recognizes it as text (230) based on the reference model, the accurately recognized text (230) may also be used for learning the reference model. According to one embodiment, the electronic device (101) may use a designated algorithm (or judgment logic) for selecting data (e.g., an input (225) for changing text (220) into text (230), or text (230)) used for learning the reference model. For example, the designated algorithm may be an algorithm for selecting (or screening, identifying) the data. For example, only inputs (225) for changing text (220) into text (230) by the above-mentioned algorithm may be selected as data to be used for learning the reference model.

Referring to example (202), the electronic device (101) can obtain models (240) for the voice recognition service from three external electronic devices (not shown). For example, the electronic device (101) can obtain an input for executing an update for a model learned (or updated) based on an input (225) from the reference model. For example, the input can include an input for the display of the electronic device (101).

According to one embodiment, the electronic device (101) may, in response to obtaining the input for executing the update, obtain models (240) from the external electronic devices. For example, the electronic device (101) may obtain model #A (241) for the voice recognition service from a first external electronic device. For example, the electronic device (101) may obtain model #B (242) for the voice recognition service from a second external electronic device. For example, the electronic device (101) may obtain model #C (243) for the voice recognition service from a third external electronic device. In example (202), an example of obtaining one model for the voice recognition service from each of three external electronic devices is illustrated, but the embodiments of the present disclosure are not limited thereto. For example, the electronic device (101) may obtain models from four or more or two or fewer external electronic devices. Alternatively, for example, the electronic device (101) may obtain a plurality of models from an external electronic device. At this time, the plurality of models may be models for a service requested by the electronic device (101) (e.g., the voice recognition service of FIG. 2).

According to one embodiment, the electronic device (101) can recognize at least one model similar to the model of the electronic device (101) among the models (240) acquired from the external electronic devices. For example, the model of the electronic device (101) may be a model learned based on input data including the input (225) from the reference model. For example, the electronic device (101) can recognize the difference between each of the models (240) and the model of the electronic device (101) and recognize the at least one model by comparing the difference with a reference value. For example, the difference may be referred to as a distance. For example, the electronic device (101) can recognize the at least one model in which the difference is less than the reference value. Although the example describes a case in which the at least one model in which the difference is less than the reference value is recognized, the embodiments of the present disclosure are not limited thereto. For example, the electronic device (101) can recognize at least one model in which a similarity value between each of the models (240) and the model of the electronic device (101) exceeds another reference value. In other words, the smaller the difference (or the distance) or the larger the similarity value, the more similar each of the models (240) and the model of the electronic device (101) can be.

According to one embodiment, the electronic device (101) can learn (or generate, update) the model of the electronic device (101) using the at least one model. For example, the electronic device (101) can generate (or learn, update) a new model from the model of the electronic device (101) for the voice recognition service using the at least one model. For example, the new model may be a model for the voice recognition service. According to one embodiment, the electronic device (101) can generate the new model based on a representative value between parameters of the model and parameters of the at least one model. For example, the new model can be generated based on an average value between first vector values of the model and second vector values of one of the at least one model. Specific details related thereto are described below in FIG. 3.

Referring to example (203), the electronic device (101) can provide the voice recognition service by using the new model for the voice recognition service. For example, the electronic device (101) can obtain a voice (250) spoken by a user (205). For example, the voice (250) can include “CCC”. For example, the electronic device (101) can recognize a text (260) from the voice (250) by using the new model. For example, the text (260) can be “CCC”. Referring to example (203), the electronic device (101) can accurately recognize a voice (250) that has not been learned previously. For example, the electronic device (101) can recognize a text (260) as a voice recognition result from the voice (250) based on the new model learned by using the at least one external electronic device.

As described above, the electronic device, method, and storage medium according to embodiments of the present disclosure can learn (or generate, update) a model for a service by evaluating the similarity between a service used by a user and a model of an external electronic device for the service when using FL. In the following, FIG. 3 describes an example of a technique for an electronic device (101) to learn a model based on a model acquired from an external electronic device (or a technique for learning a personalized model based on P2P).

FIG. 3 illustrates an example of a network for generating a personalized model based on peer to peer (P2P), according to one embodiment.

Referring to FIG. 3, an example of a network for generating a personalized model using a model acquired from external electronic devices (310, 320, 330) is illustrated. For example, FIG. 3 illustrates an example of the network including one electronic device (101), one server (350), and three external electronic devices (310, 320, 330), but embodiments of the present disclosure are not limited thereto. For example, the number of components constituting the network may be changed.

Referring to FIG. 3, an exemplary situation is illustrated in which an electronic device (101), external electronic devices (310, 320, 330), and a server (350) are connected to each other based on a wired network and/or a wireless network. In FIG. 3, an example is illustrated in which the electronic device (101) and the external electronic device (330) and the external electronic device (310) and the external electronic device (320) are not connected to each other, but this is merely for convenience of explanation, and the embodiments of the present disclosure are not limited thereto. FIG. 3 assumes that the electronic device (101) and the external electronic device (330) and the external electronic device (310) and the external electronic device (320) are connected to each other. For example, the wired network may include a network such as the Internet, a local area network (LAN), a wide area network (WAN), or a combination thereof. For example, the wireless network may include a network such as long term evolution (LTE), fifth generation (5G) new radio (NR), wireless fidelity (WiFi), Zigbee, near field communication (NFC), Bluetooth, Bluetooth low-energy (BLE), or a combination thereof. Although the electronic device (101), external electronic devices (310, 320, 330), and server (350) are illustrated as being directly connected, the electronic device (101), external electronic devices (310, 320, 330), and server (350) may be indirectly connected via one or more routers and/or access points (APs).

Referring to FIG. 3, according to an embodiment, an electronic device (101) may include at least one of a processor (301), a communication circuit (303), and a memory (305). The processor (301), the communication circuit (303), and the memory (305) may be electrically and/or operably coupled with each other by a communication bus. Hereinafter, the hardware components being operably coupled may mean that a direct connection or an indirect connection is established, wired or wireless, between the hardware components such that a second hardware component is controlled by a first hardware component among the hardware components. Although illustrated based on different blocks, the embodiment is not limited thereto, and some of the hardware components illustrated in FIG. 3 (e.g., at least a portion of the processor (301), the communication circuit (303), and the memory (305)) may be included in a single integrated circuit such as a system on a chip (SoC). The type and/or number of hardware components included in the electronic device (101) are not limited to those illustrated in FIG. 3. For example, the electronic device (101) may include only some of the hardware components illustrated in FIG. 3.

According to one embodiment, the processor (301) of the electronic device (101) may include a hardware component for processing data based on one or more instructions. The hardware component for processing data may include, for example, an arithmetic and logic unit (ALU), a floating point unit (FPU), and a field programmable gate array (FPGA). As an example, the hardware component for processing data may include a central processing unit (CPU), a graphics processing unit (GPU), a digital signal processing (DSP), and/or a neural processing unit (NPU). The number of processors (301) may be one or more. For example, the processor (301) may have a multi-core processor structure such as a dual core, a quad core, or a hexa core. The processor (301) of FIG. 3 may include the processor (120) of FIG. 1.

For example, the processor (301) of the electronic device (101) may include various processing circuits and/or multiple processors. For example, the term "processor" as used herein, including in the claims, may include various processing circuits including at least one processor, one or more of which may be configured to perform the various functions described below individually and/or collectively in a distributed manner. As used herein, when "processor," "at least one processor," and "one or more processors" are described as being configured to perform various functions, these terms encompass, for example and without limitation, situations where one processor performs some of the recited functions and other processor(s) perform other parts of the recited functions, and also situations where one processor may perform all of the recited functions. Additionally, the at least one processor may include a combination of processors that perform the various functions enumerated/disclosed, for example, in a distributed manner. At least one processor is capable of executing program instructions to accomplish or perform various functions.

According to one embodiment, the communication circuit (303) of the electronic device (101) may include hardware for supporting transmission and/or reception of electrical signals between the electronic device (101), the server (350), and external electronic devices (310, 320, 330). The communication circuit (303) may include, for example, at least one of a modem, an antenna, and an optical/electronic (O/E) converter. The communication circuit (303) may support transmission and/or reception of electrical signals based on various types of communication means, such as Ethernet, Bluetooth, BLE (Bluetooth low energy), ZigBee, LTE (long term evolution), and 5G NR (new radio). The communication circuit (303) of FIG. 3 may include the communication module (190) and/or the antenna module (197) of FIG. 1.

According to one embodiment, the memory (305) of the electronic device (101) may include a hardware component for storing data and/or instructions input to and/or output from the processor (301). The memory (305) may include, for example, a volatile memory such as a random-access memory (RAM), and/or a non-volatile memory such as a read-only memory (ROM). The volatile memory may include, for example, at least one of a dynamic RAM (DRAM), a static RAM (SRAM), a Cache RAM, and a pseudo SRAM (PSRAM). The non-volatile memory may include, for example, at least one of a programmable ROM (PROM), an erasable PROM (EPROM), an electrically erasable PROM (EEPROM), a flash memory, a hard disk, a compact disc, and an embedded multimedia card (eMMC). The memory (305) of Fig. 3 may include the memory (130) of Fig. 1.

Although not shown in FIG. 3, according to one embodiment, the electronic device (101) may include a display. For example, the display of the electronic device (101) may output visualized information to a user. The number of displays included in the electronic device (101) may be one or more. For example, the display may be controlled by a processor (301) and/or a graphic processing unit (GPU) (not shown) to output visualized information to a user. The display may include a flat panel display (FPD) and/or electronic paper. The FPD may include a liquid crystal display (LCD), a plasma display panel (PDP), a digital mirror device (DMD), one or more light emitting diodes (LEDs), and/or micro LEDs. The LEDs may include organic LEDs (OLEDs). The display may include the display module (160) of FIG. 1.

Although not shown in FIG. 3, according to one embodiment, the electronic device (101) may include a camera. The camera may include one or more optical sensors (e.g., a CCD (charged coupled device) sensor, a CMOS (complementary metal oxide semiconductor) sensor) that generate electrical signals representing the color and/or brightness of light. The plurality of optical sensors included in the camera may be arranged in the form of a two-dimensional array. The camera may acquire electrical signals of each of the plurality of optical sensors substantially simultaneously, and may generate an image corresponding to light reaching the optical sensors of the two-dimensional grid and including a plurality of pixels arranged two-dimensionally. In addition, although not shown in FIG. 3, the electronic device (101) may include a sensor. For example, the sensor may include at least one sensor. For example, the sensor may include at least a portion of the sensor module (176) of FIG. 1. For example, the sensor may include an IMU (or an IMU sensor). For example, the sensor may include a gyro sensor, a gravity sensor, and/or an acceleration sensor. For example, the electronic device (101) may obtain input data (or data used for learning a model) according to the use of a service by a user of the electronic device (101) based on the camera or the sensor.

In addition, although not shown in FIG. 3, the electronic device (101) according to one embodiment may include an output means for outputting information in a form other than a visualized form. For example, the electronic device (101) may include a speaker for outputting an acoustic signal. For example, the electronic device (101) may include a motor for providing haptic feedback based on vibration.

Referring to FIG. 3, according to one embodiment, in the memory (305) of the electronic device (101), one or more instructions (or commands) representing operations and/or actions to be performed on data by the processor (301) of the electronic device (101) may be stored. A set of one or more instructions may be referred to as a program, firmware, an operating system, a process, a routine, a sub-routine, and/or an application. Hereinafter, when an application is installed in the electronic device (e.g., the electronic device (101)), it may mean that one or more instructions provided in the form of an application are stored in the memory (305), and that the one or more applications are stored in a format executable by the processor of the electronic device (e.g., a file having an extension specified by the operating system of the electronic device (101)). According to one embodiment, the electronic device (101) may execute one or more instructions stored in the memory (305) to perform the operation of FIG. 4 or FIG. 6.

Referring to FIG. 3, programs installed in the electronic device (101) may be classified into one of different layers including an application layer, a framework layer, and/or a hardware abstraction layer (HAL) based on a target. For example, programs (e.g., drivers) designed to target hardware (e.g., communication circuit (303), the display, the sensor) of the electronic device (101) may be classified within the hardware abstraction layer. For example, programs (e.g., similarity determination module (305-1), model training module (305-2), and/or service provision module (305-3)) designed to target at least one of the hardware abstraction layer and/or the application layer may be classified within the framework layer. Programs classified into the framework layer may provide an executable API (application programming interface) based on other programs.

For example, within the application layer, a program designed to target a user controlling an electronic device (101) may be classified. For example, a program classified into the application layer may include an application that provides a service to the user. However, the embodiment of the present disclosure is not limited thereto. For example, a program classified into the application layer may call an API to cause execution of a function supported by programs classified into the framework layer.

Referring to FIG. 3, an electronic device (101) according to an embodiment may discriminate the similarity between a model of the electronic device (101) and a model acquired from an external electronic device based on the execution of a similarity determination module (305-1). For example, the electronic device (101) may acquire a reference model from a server (350). For example, the electronic device (101) may generate (or learn, update) a first model by using input data acquired by a user of the electronic device (101) from the reference model. For example, the first model may be a model for the service learned by using the input data acquired while using the service from the reference model. For example, the electronic device (101) may provide first information indicating the service to the server (350). For example, the server (350) may provide second information indicating an external electronic device that uses the service to the electronic device (101). For convenience of explanation, it is assumed that the external electronic devices are the external electronic device (310) and the external electronic device (320). For example, the electronic device (101) can establish a connection with the external electronic device (310) based on the second information indicating the external electronic device (310) and the external electronic device (320), and obtain a second model of the external electronic device (310) for the service from the external electronic device (310). For example, the second model can be a model for the service learned by the external electronic device (310) from the reference model. In addition, for example, the electronic device (101) can establish a connection with the external electronic device (320) based on the second information indicating the external electronic device (310) and the external electronic device (320), and obtain a third model of the external electronic device (320) for the service from the external electronic device (320). For example, the third model may be a model for the service learned by an external electronic device (320) from the reference model.

According to one embodiment, the electronic device (101) can determine the similarity between a model of the electronic device (101) and a model of an external electronic device. For example, the electronic device (101) can determine the similarity by directly comparing the model of the electronic device (101) and the model of the external electronic device, or by inputting reference data to each of the model of the electronic device (101) and the model of the external electronic device and comparing the outputs thereof.

For example, the electronic device (101) can directly compare vector values constituting a model of the electronic device (101) with vector values constituting a model of an external electronic device. The comparison between the models as described above can be referred to by the following mathematical formula.

The above D may represent a difference indicating a similarity between models, the W ₁ may represent vector values of a model of an electronic device (101), the W ₂ may represent vector values of a model of an external electronic device, and the f may represent a function for measuring the similarity. For example, the f may include a cosine distance function, a Euclidean distance function, a Manhattan distance function, a correlation function, or an inner product function.

Alternatively, for example, the electronic device (101) may input reference data to each of a model of the electronic device (101) and a model of an external electronic device, and compare the output results to determine the similarity between the model of the electronic device (101) and the model of the external electronic device. For example, a method of determining the similarity using the reference data may be referred to as a TC (test case) method. The comparison using the TC method as described above may be referred to by the following mathematical equation.

The above D may represent a difference indicating a similarity between models, the α ₁ may represent a value of an interference function (M ₁ (t)) by a model of the electronic device (101) (i.e., α ₁ =M ₁ (t) and t is the reference data value), the α ₂ may represent a value of an interference function (M ₂ (t)) of a model of an external electronic device (i.e., α ₂ =M ₂ (t) and t is the reference data value), and the f may represent a function for measuring the similarity. For example, the f may include a cosine distance function, a Euclidean distance function, a Manhattan distance function, a correlation function, or an inner product function.

The method for determining the similarity through comparison between the above models may refer to the following examples. For example, the electronic device (101) may calculate distance values representing the difference between vector values constituting the model of the electronic device (101) and vector values constituting the model of the external electronic device. For example, the electronic device (101) may calculate first distance values representing the difference between first vector values constituting the first model and second vector values constituting the second model. For example, the electronic device (101) may obtain the first vector values. For example, the electronic device (101) may obtain the second vector values. For example, the first vector values may be [0.1223, 0.4864, 1.4133, 8.1531, ...]. For example, the second vector values may be [0.8997, 2.1253, 4.0011, 0.1113, ...]. The electronic device (101) may calculate first distance values representing a difference between the first vector values and the second vector values. For example, the first distance values may be [0.7774, 1.6389, 2.5878, 8.0418, ...]. For example, the electronic device (101) may calculate the first distance values as a difference for determining the similarity between the first model and the second model. In addition, for example, the electronic device (101) may calculate second distance values representing a difference between the first vector values constituting the first model and third vector values constituting the third model. For example, the electronic device (101) may obtain the third vector values. The third vector values may be [0.1664, 0.2844, 2,0053, 8.0003, ...]. The electronic device (101) may calculate the second distance values representing the difference between the first vector values and the third vector values. For example, the second distance values may be [0.0441, 0.202, 0.592, 0.1528, ...]. For example, the electronic device (101) may calculate the second distance values as the difference for determining the similarity between the first model and the third model.

Alternatively, a method for determining similarity through comparison using the TC may refer to the following example. It is assumed that the electronic device (101) provides a voice recognition service. For example, the electronic device (101) may input reference data (e.g., “Hello.”) to each of a model of the electronic device (101) and a model of an external electronic device. For example, the electronic device (101) may calculate reference vector values based on the model of the electronic device (101) from the reference data and reference vector values based on the model of the external electronic device (101) from the reference data. For example, the electronic device (101) may calculate first reference vector values based on the first model, second reference vector values based on the second model, and third reference vector values based on the third model from the reference data. For example, the first reference vector values can be [0.1336, 0.5808, 1.6322, 8.3521, ...]. For example, the second reference vector values can be [0.9999, 2.2311, 4.1101, 0.2131, ...]. For example, the third reference vector values can be [0.2541, 0.3991, 2.1911, 8.1019, ...]. For example, the electronic device (101) can calculate the first distance values representing the difference between the first reference vector values and the second reference vector values. For example, the first distance values can be [0.8663, 1.6503, 2.4779, 8.139, ...]. For example, the electronic device (101) can calculate the first distance values as differences for determining the similarity between the first model and the second model. In addition, the electronic device (101) can calculate the second distance values representing differences between the first reference vector values and the third reference vector values. For example, the second distance values can be [0.1205, 0.1817, 0.5589, 0.2502, ...]. For example, the electronic device (101) can calculate the second distance values as differences for determining the similarity between the first model and the third model.

For example, the electronic device (101) can calculate the difference based on the distance values. For example, the electronic device (101) can calculate the difference using a representative value (e.g., an average value) of the distance values. For example, the electronic device (101) can calculate the difference between the first model and the second model as 5.1296. Also, for example, the electronic device (101) can calculate the difference between the first model and the third model as 0.2136. In the above example, an example of calculating the difference between models is described, but the embodiments of the present disclosure are not limited thereto. For example, the electronic device (101) can also calculate the difference based on a comparison using the TC. For example, the electronic device (101) can determine whether the difference is less than the reference value (e.g., 1). For example, the electronic device (101) may determine that the difference between the first model and the second model is greater than or equal to the reference value, and may determine that the difference between the first model and the third model is less than the reference value. Accordingly, the electronic device (101) may utilize the third model for learning the first model.

In the above example, an example is described in which the electronic device (101) determines the similarity based on a model acquired from an external electronic device, but the embodiments of the present disclosure are not limited thereto. According to one embodiment, the electronic device (101) may transmit a signal requesting the model of the external electronic device together with the model of the electronic device (101) (e.g., the first model). In response, the external electronic device may perform a similarity determination between the model of the external electronic device (e.g., the second model) and the model of the electronic device (101) (e.g., the first model). At this time, if the external electronic device determines that the model of the external electronic device (e.g., the second model) and the model of the electronic device (101) (e.g., the first model) are similar, the external electronic device may provide the model of the external electronic device (e.g., the second model) to the electronic device (101). Thereafter, the electronic device (101) can create (or learn, train, update) a new model from a model of the electronic device (101) (e.g., the first model) using a model of the external electronic device (e.g., the second model).

Referring to FIG. 3, an electronic device (101) according to an embodiment may train a model of the electronic device (101) based on the execution of the model training module (305-2). For example, the electronic device (101) may generate (or train, adaptively train, learn, update) a new model using input data by a user from the reference model acquired from the server (350). Alternatively, the electronic device (101) may generate (or train, adaptively train, learn, update) a new model using a model acquired from an external electronic device (e.g., the external electronic device (310)) from a model generated from the reference model. For example, the electronic device (101) may generate a fourth model from the first model using the third model. For example, the fourth model may be a model for the service provided based on the first model. For example, the fourth model can be generated based on an average between the first vector values of the first model and the second vector values of the second model. For example, it is assumed that the first vector values are [0.1223, 0.4864, 1.4133, 8.1531, ...], and the third vector values are [0.1664, 0.2844, 2,0053, 8.0003, ...]. For example, the fourth vector values constituting the fourth model can be [0.1444, 0.3854, 1.7093, 8.0767, ...]. In other words, each of the fourth vector values can represent an average of each of the first vector values and each of the third vector values. In the above example, an example is described of using one second model for the first model, but the embodiment of the present disclosure is not limited thereto. For example, a plurality of models may be merged for the first model. In addition, in the above example, an example is described of creating (or obtaining, updating) a new model (e.g., the fourth model) by merging models based on an average, but the embodiment of the present disclosure is not limited thereto.

Referring to FIG. 3, the electronic device (101) can provide a service using a model by using the service providing module (305-3). For example, the electronic device (101) can provide the service by using the first model learned based on the input data from the reference model. Or, for example, the electronic device (101) can provide the service by using the fourth model generated based on the third model from the first model. For example, the service providing module (305-3) can be referred to as an inferencer.

In the example of FIG. 3, each of the external electronic devices (310, 320, 330) may include substantially the same components of the electronic device (101) (e.g., the processor (301), the communication circuit (303), and the memory (305)). For example, the external electronic device (310) may include a processor, a communication circuit, and a memory. For example, the external electronic device (320) may include a processor, a communication circuit, and a memory. For example, the external electronic device (330) may include a processor, a communication circuit, and a memory. The description of each of the processor, the communication circuit, and the memory may be substantially the same as the description of the processor (301), the communication circuit (303), and the memory (305).

Referring to FIG. 3, the server (350) may include a processor (351), a communication circuit (353), and a memory (355). The description of each of the processor (351), the communication circuit (353), and the memory (355) may be substantially identical to the description of the processor (301), the communication circuit (303), and the memory (305). For example, the server (350) may be associated with a specific service. For example, the specific service may include a domain.

According to one embodiment, the memory (355) of the server (350) may include a reference model generation module (355-1). For example, the server (350) may generate (or train) and store a reference model for FL based on the reference model generation module (355-1). For example, the server (350) may include an application programming interface (API) to be used when the electronic device (101) or external electronic devices (310, 320, 330) downloads the reference model. For example, the API may include a rest API.

FIG. 4 illustrates an example of a signal flow for a method of generating a personalized model based on P2P, according to one embodiment.

The electronic device (101) of FIG. 4 may be an example of the electronic device (101) of FIG. 3. The external electronic device (310) of FIG. 4 may be an example of the external electronic device (310) of FIG. 3. The server (350) of FIG. 4 may be an example of the server (350) of FIG. 3. In the example of FIG. 4, an example of a method for generating a personalized model based on the P2P through one external electronic device (310) is illustrated, but the embodiments of the present disclosure are not limited thereto. For example, the embodiments of the present disclosure may include a method for generating a personalized model based on the P2P through a plurality of external electronic devices.

Referring to FIG. 4, in operation (400), the server (350) can generate a reference model. For example, the server (350) can generate the reference model for FL. For example, the server (350) can store the generated reference model. For example, the reference model can be a model for a service provided by the electronic device (101).

In operation (405), the server (350) can transmit the reference model. For example, the server (350) can establish a connection with a plurality of electronic devices for the FL. For example, the server (350) can establish a connection with the electronic device (101). For example, the server (350) can establish a connection with an external electronic device (310). For example, the server (350) can transmit the reference model to each of the electronic device (101) and the external electronic device (101) with which the connection is established.

In operation (410), the electronic device (101) may obtain (or collect) input data. For example, the electronic device (101) may obtain the user's input data for the service. For example, the input data may include an input performed by the user to the electronic device (101). For example, the input data may include the user's utterance, the user's touch input, or the user's usage history. However, the embodiments of the present disclosure are not limited thereto.

According to one embodiment, the input data can be acquired while the user uses the service. For example, the electronic device (101) can provide the service using the reference model. For example, the electronic device (101) can generate output data from the input data using the reference model. The output data can be used to provide the service. While the service is provided using the reference model, the electronic device (101) can acquire the input data.

In operation (415), the electronic device (101) may generate a first model. For example, the electronic device (101) may generate (or train, adaptively train, learn, update) the first model using the input data from the reference model. For example, the first model may be a model for the service.

In operation (420), the electronic device (101) may transmit first information. For example, the electronic device (101) may transmit the first information indicating the service to the server (350). For example, the first information indicating the service may include at least one of a name or identification information of the service. For example, if the service is provided based on a specific software application, it may include at least one of a name or identification information of the specific software application. In the above example, an example in which the electronic device (101) transmits the first information indicating the service to the server (350) is described, but the embodiments of the present disclosure are not limited thereto. For example, the electronic device (101) may transmit location information of the electronic device (101) to the server (350) together with the first information. For example, the location information of the electronic device (101) may include GPS (global positioning system) coordinate information of the electronic device (101) or location information based on a communication technique (e.g., WiFi, BLE).

In FIG. 4, operation (420) is illustrated as being performed after operation (415), but the embodiments of the present disclosure are not limited thereto. For example, operation (420) may be performed simultaneously with operation (415), or may be performed before operation (415).

In operation (425), the server (350) can transmit second information. For example, the server (350) can transmit the second information indicating the external electronic device (310) using the service to the electronic device (101). For example, the server (350) can recognize the electronic device using the service based on the first information obtained from the electronic device (101) through operation (420). For example, the server (350) can recognize the external electronic device (310) based on the first information. For example, the server (350) can transmit the second information indicating the recognized external electronic device (310) to the electronic device (101).

In the above example, an example is described in which the server (350) generates the second information indicating the external electronic device (310) recognized based on the first information, and transmits the second information to the electronic device (101), but the embodiment of the present disclosure is not limited thereto. According to one embodiment, the server (350) may generate the second information by further using the location information received from the electronic device (101). For example, the server (350) may recognize a plurality of external electronic devices using the service, and determine an external electronic device (310) that can be connected (or has a high possibility of being connected) to the electronic device (101) by using the location information of the electronic device (101) among the plurality of external electronic devices. For example, the external electronic device (310) may be included in at least one external electronic device that is within a specified distance from the location of the electronic device (101) among the plurality of external electronic devices. For example, the server (350) can recognize the external electronic device (310) based on the location information obtained from each of the plurality of external electronic devices and the location information of the electronic device (101). Accordingly, the server (350) can generate the second information indicating the external electronic device (310) and transmit the second information to the electronic device (101).

According to one embodiment, the second information may include at least one of a MAC (media access control) address of the external electronic device (310) or information for connection between the external electronic device (310) and the electronic device (101).

In operation (430), the electronic device (101) and the external electronic device (310) may establish a connection. For example, the electronic device (101) may attempt to establish a connection with the external electronic device (310) based on the second information.

In operation (435), the electronic device (101) may request a model for the service. For example, the electronic device (101) may transmit a signal requesting the model for the service to the external electronic device (310). In operation (440), the external electronic device (310) may transmit a second model to the electronic device (101). For example, the second model may be a model for the service generated (or learned, trained, adapted, trained, updated) by the external electronic device (310) from the reference model.

In operation (445), the electronic device (101) can determine the similarity between the models. For example, the electronic device (101) can determine the similarity between the first model and the second model. For example, the electronic device (101) can determine whether a difference between the first model and the second model is less than a reference value. For example, determining the similarity may include a method of comparing the first model and the second model themselves, or a method of comparing vector values output from reference data input for the first model and the second model. For specific details on methods of determining the similarity, reference may be made to an operation based on the similarity determination module (305-1) of FIG. 3.

Referring to operations (435), (440), and (445), an example of a method for determining the similarity by using the second model acquired from the external electronic device (310) by the electronic device (101) is illustrated in the example of FIG. 4, but the embodiments of the present disclosure are not limited thereto. For example, when the electronic device (101) transmits the signal requesting the model for the service, the electronic device (101) may transmit the first model of the electronic device (101) to the external electronic device (310) together with the signal. The external electronic device (310) may determine the similarity between the first model and the second model. Upon determining that the difference between the second model and the first model is less than the reference value, the external electronic device (310) may transmit the second model to the electronic device (101). In this case, the electronic device (101) can perform the operation (450) without determining the similarity between the received second model and the first model.

In operation (450), the electronic device (101) may generate a third model. For example, the electronic device (101) may generate the third model by using the second model from the first model. For example, the electronic device (101) may generate the third model by using the first model and the second model. For example, the third model may be a model for the service. For example, the third model may be composed of third vector values having an average value of first vector values constituting the first model and second vector values constituting the second model. For example, each of the third vector values may have an average value of each of the first vector values and each of the second vector values.

Although not illustrated in FIG. 4, the electronic device (101) can provide the service using the generated third model. For example, the electronic device (101) can obtain the user's input data while providing the service using the third model. Thereafter, the electronic device (101) can learn (or train, adaptively train, update) the third model based on the input data obtained while providing the service using the third model.

In addition, although not shown in FIG. 4, the electronic device (101) may transmit the generated third model to the server (350). For example, the electronic device (101) may transmit the third model to be used for federated learning to the server (350). For example, the server (350) may receive the third model from the electronic device (101) and the model learned from the external electronic device (310). For example, the model received from the external electronic device (310) may be a model learned from the second model of the external electronic device (310). For example, the server (350) may generate an aggregated model according to the federated learning by using the third model and the model received from the external electronic device (310).

FIGS. 5A and 5B illustrate an example of a method for learning a model of an electronic device using a model obtained from an external electronic device, according to one embodiment.

The electronic device (101) of FIGS. 5A and 5B may be an example of the electronic device (101) of FIG. 3. The external electronic device (310) of FIGS. 5A and 5B may be an example of the external electronic device (310) of FIG. 3. The server (350) of FIGS. 5A and 5B may be an example of the server (350) of FIG. 3. Although FIGS. 5A and 5B illustrate examples (500, 550) of methods for utilizing a model acquired from one external electronic device (310), the embodiments of the present disclosure are not limited thereto. For example, models acquired from a plurality of external electronic devices may also be utilized.

FIG. 5A illustrates an example (500) of a method for an electronic device (101) to learn (or generate, update) a model of an electronic device (101) using a model obtained from an external electronic device (310). The electronic device (101), the server (350), and the external electronic device (310) may be connected to perform FL.

Referring to example (500), the server (350) can generate model #A (510). For example, model #A (510) can be generated by the server (350) and referenced as a reference model for a service. For example, the server (350) can transmit the generated model #A (510) to each of the electronic device (101) and the external electronic device (310).

For example, the electronic device (101) can provide the service using the acquired model #A (511). For example, the electronic device (101) can obtain input data of a user of the electronic device (101) while providing the service using model #A (511). For example, the electronic device (101) can generate (or learn, update) model #B (521) from model #A (511) using the input data of the user of the electronic device (101).

In addition, for example, the external electronic device (310) can provide the service using the acquired model #A (512). For example, the external electronic device (310) can obtain input data of a user of the external electronic device (310) while providing the service using model #A (512). For example, the external electronic device (310) can generate (or learn, update) model #C (522) from model #A (512) using the input data of the user of the external electronic device (310).

For example, the electronic device (101) can transmit first information indicating the service to the server (350) and receive second information indicating an external electronic device (310) that uses the service from the server (350). The electronic device (101) can be connected to the external electronic device (310) based on the second information. The electronic device (101) can request a model for the service from the external electronic device (310). For example, the electronic device (101) can obtain model #C (522) from the external electronic device (310). For example, the electronic device (101) can determine the similarity between model #B (521) and model #C (522). For example, the electronic device (101) may determine that the model #C (522) is similar to the model #B (521) if the difference calculated based on the distance values between the vector values of the model #C (522) and the vector values of the model #B (521) is less than a reference value. For example, the electronic device (101) may generate (or learn, update) the model #BC (523) by using the model #C (522) from the model #B (521).

With respect to the example (500) of FIG. 5a, if the user of the electronic device (101) is right-handed, the location of the input (e.g., keyboard input) used by the user of the electronic device (101) for the electronic device (101) may be relatively distributed on the right side of the electronic device (101). At this time, when the user of the electronic device (101) executes an update of a service providing the keyboard, a user similar to the user of the electronic device (101) (e.g., the user of the external electronic device (310)) may be recognized, and an update may be performed using a model of the external electronic device (310). At this time, the user of the external electronic device (310) may be right-handed. If the user of the external electronic device (310) is left-handed, the electronic device (101) may recognize that the difference calculated based on the distance values between the vector values of model #C (522) and the vector values of model #B (521) is greater than or equal to the reference value.

In addition, with respect to the example (500) of FIG. 5a, it is assumed that the user of the electronic device (101) uses a voice recognition service. For example, if the user of the electronic device (101) uses a standard language, the model #B (521) of the electronic device (101) can be learned (or generated, updated) based on the input data that is the standard language. The electronic device (101) can recognize a user similar to the user of the electronic device (101) (e.g., the user of the external electronic device (310)) and perform an update using the model of the external electronic device (310). The user of the external electronic device (310) can use the standard language. Thereafter, if the user of the electronic device (101) uses a dialect, the model of the electronic device (101) can be learned based on the input data that is the dialect. The electronic device (101) can recognize a user different from the user of the electronic device (101) and update using the model of the other external electronic device. At this time, the other user can use the dialect.

FIG. 5b illustrates an example (500) of a method in which an electronic device (101) learns (or generates, updates) a model for a service of the electronic device (101) using a model for the service obtained from an external electronic device (310), and learns (or generates, updates) a model of another service related to the service. The electronic device (101), the server (350), and the external electronic device (310) may be connected to perform FL.

Referring to example (550), the server (350) can generate model #A (561) and model #B (562). For example, model #A (561) can be referenced as a reference model for a first service generated by the server (350). For example, model #B (562) can be referenced as a reference model for a second service generated by the server (350). The second service can be related to the first service. For example, when the first service is a book recommendation service, the second service can be a movie recommendation service. For example, the first service and the second service are services that recommend content, and can be classified as related services. For example, a voice recognition service can be a service that is not related to the first service and the second service. For example, the server (350) can transmit the generated model #A (561) and model #B (562) to the electronic device (101) and the external electronic device (310), respectively.

For example, the electronic device (101) can provide the first service using the acquired model #A (571). For example, the electronic device (101) can obtain first input data of a user of the electronic device (101) while providing the first service using model #A (571). For example, the electronic device (101) can generate (or learn, update) model #A1 (581) from model #A (571) using the first input data of the user of the electronic device (101). In addition, the electronic device (101) can provide the second service using the acquired model #B (572). For example, the electronic device (101) can obtain second input data of a user of the electronic device (101) while providing the second service using model #B (572). For example, the electronic device (101) can generate (or learn, update) model #B1 (582) from model #B (572) using the second input data of the user of the electronic device (101).

In addition, for example, the external electronic device (310) can provide the first service using the acquired model #A (573). For example, the external electronic device (310) can obtain third input data of the user of the external electronic device (310) while providing the first service using model #A (573). For example, the external electronic device (310) can generate (or learn, update) model #A2 (583) from model #A (573) using the third input data of the user of the external electronic device (310). In addition, the external electronic device (310) can provide the second service using the acquired model #B (574). For example, the external electronic device (310) can obtain fourth input data of the user of the external electronic device (310) while providing the second service using model #B (574). For example, the external electronic device (310) can generate (or learn, update) model #B2 (584) from model #B (574) using the fourth input data of the user of the external electronic device (310).

For example, the electronic device (101) can transmit first information indicating the first service and the second service to the server (350), and receive second information indicating an external electronic device (310) that uses the first service and the second service from the server (350). The electronic device (101) can be connected to the external electronic device (310) based on the second information. The electronic device (101) can request a model for the first service and the second service from the external electronic device (310). For example, the electronic device (101) can obtain model #A2 (583) for the first service and model #B2 (584) for the second service from the external electronic device (310). For example, the electronic device (101) can determine the similarity between model #A1 (581) and model #A2 (583). For example, the electronic device (101) may determine that the model #A2 (583) is similar to the model #A1 (581) if the difference calculated based on the distance values between the vector values of the model #A1 (581) and the vector values of the model #A2 (583) is less than a reference value. For example, the electronic device (101) may generate (or learn, update) the model #A1A2 (591) from the model #A1 (581) by using the model #A2 (583). According to one embodiment, the electronic device (101) may determine the similarity between the model #A1 (581) and the model #A2 (583) and, while generating the model #A1A2 (591), determine the similarity to the model for the second service related to the first service. For example, the electronic device (101) can determine the similarity between model #B1 (582) and model #B2 (584). For example, the electronic device (101) can determine that model #B2 (584) is similar to model #B1 (582) when a difference calculated based on distance values between vector values of model #B1 (582) and vector values of model #B2 (584) is less than a reference value. For example, the electronic device (101) can generate (or learn, train, adaptively train, update) model #B1B2 (592) by using model #B2 (584) from model #B1 (582).

In the above example, in response to the electronic device (101) determining the similarity of the models for the second service, an example is described in which the electronic device generates (or learns, trains, adapts, trains, updates) the model #B1B2 (592), but the embodiments of the present disclosure are not limited thereto. For example, if the electronic device (101) determines that the models related to the first service (e.g., model #A1 (581) and model #A2 (583)) are similar, the electronic device (101) may generate (or learn, train, adaptively train, update) the model #B1B2 (592) using the models for the second service (e.g., model #B1 (582) and model #B2 (584)) related to the first service. In other words, the electronic device (101) can create (or learn, train, adaptively train, update) a new model for the second service using the models for the second service without determining the similarity between the models for the second service when the models related to the first service are similar.

FIG. 6 illustrates an example of an operational flow for a method of generating a personalized model based on P2P, according to one embodiment.

At least some of the above methods of FIG. 6 may be performed by the electronic device (101) of FIG. 3. For example, at least some of the above methods may be controlled by the processor (301) of the electronic device (101). In the following embodiments, the respective operations may be performed sequentially, but are not necessarily performed sequentially. For example, the order of the respective operations may be changed, and at least two operations may be performed in parallel.

According to one embodiment, in operation (610), the electronic device (101) may obtain a reference model. For example, the electronic device (101) may obtain the reference model from a server (e.g., server (350) of FIG. 3). For example, the reference model may be generated by the server (350). For example, the reference model may be a model for a service.

In one embodiment, in operation (620), the electronic device (101) may generate a first model from the reference model. For example, the electronic device (101) may generate (or train, learn, update) the first model using input data from the reference model. For example, the first model may be a model for the service.

According to one embodiment, the electronic device (101) may obtain (or collect) the input data. For example, the electronic device (101) may obtain the user's input data for the service. For example, the input data may include an input performed by the user to the electronic device (101). For example, the input data may include the user's utterance, the user's touch input, or the user's usage history. However, the embodiments of the present disclosure are not limited thereto.

According to one embodiment, the input data can be acquired while the user uses the service. For example, the electronic device (101) can provide the service using the reference model. For example, the electronic device (101) can generate output data from the input data using the reference model. The output data can be used to provide the service. While the service is provided using the reference model, the electronic device (101) can acquire the input data.

According to one embodiment, in operation (630), the electronic device (101) may transmit first information indicating the service. For example, the electronic device (101) may transmit the first information indicating the service to the server (350). For example, the first information indicating the service may include a name or identification information of the service. For example, if the service is a service provided based on a specific software application, it may include a name or identification information of the specific software application. In FIG. 6, operation (630) is illustrated as being performed after operation (620), but the embodiments of the present disclosure are not limited thereto. For example, operation (630) may be performed simultaneously with operation (620) or may be performed before operation (620).

According to one embodiment, in operation (640), the electronic device (101) may obtain second information indicating an external electronic device. For example, the electronic device (101) may receive the second information indicating the external electronic device from the server (350). For example, the second information may include a MAC (media access control) address of the external electronic device or information for connection between the external electronic device and the electronic device (101). In the above example, for convenience of explanation, the second information indicating one external electronic device is described, but the embodiment of the present disclosure is not limited thereto. For example, the second information may include the second information indicating a plurality of external electronic devices. In this case, each of the plurality of external electronic devices indicated by the second information may be an electronic device providing the service. For example, each of the plurality of external electronic devices may generate (or update, learn) a model for the service from the reference model obtained from the server (350).

According to one embodiment, in operation (650), the electronic device (101) can obtain a second model for the service of the external electronic device. For example, the electronic device (101) and the external electronic device can establish a connection. For example, the electronic device (101) can attempt to establish the connection with the external electronic device based on the second information.

According to one embodiment, the electronic device (101) may request a model for the service from the external electronic device. For example, the electronic device (101) may transmit a signal requesting a model for the service to the external electronic device. For example, the electronic device (101) may obtain the second model for the service from the external electronic device. For example, the second model may be a model for the service generated (or learned, trained, or updated) by the external electronic device from the reference model.

According to one embodiment, in operation (660), the electronic device (101) may generate the third model for the service using the first model and the second model. For example, the electronic device (101) may generate the third model for the service using the second model from the first model. For example, the electronic device (101) may generate the third model for the service using the second model from the first model when a difference between the first model and the second model is less than a reference value. For example, the third model may be a model for the service. For example, the third model may be composed of third vector values having an average value of first vector values constituting the first model and second vector values constituting the second model. For example, each of the third vector values may have an average value of each of the first vector values and each of the second vector values.

According to one embodiment, the electronic device (101) can determine the similarity between models. For example, the electronic device (101) can determine the similarity between the first model and the second model. For example, the electronic device (101) can determine whether the difference between the first model and the second model is less than the reference value. For example, determining the similarity may include a method of comparing the first model and the second model themselves, or a method of comparing vector values output from reference data input for the first model and the second model. For specific details on the methods of determining the similarity, reference may be made to the operation based on the similarity determination module (305-1) of FIG. 3.

According to one embodiment, the electronic device (101) can provide the service by using the generated third model. For example, the electronic device (101) can obtain the user's input data while providing the service by using the third model. Thereafter, the electronic device (101) can learn (or train, update) the third model based on the input data obtained while providing the service by using the third model.

Although the above example illustrates an example of a method for the electronic device (101) to determine the similarity using the second model acquired from the external electronic device, the embodiments of the present disclosure are not limited thereto. For example, when the electronic device (101) transmits the signal requesting the model for the service, the electronic device (101) may transmit the first model of the electronic device (101) to the external electronic device together with the signal. The external electronic device may determine the similarity between the first model and the second model. Upon determining that the difference between the second model and the first model is less than the reference value, the external electronic device may transmit the second model to the electronic device (101). In this case, the electronic device (101) may generate the third model without determining the similarity between the received second model and the first model.

FIG. 7 illustrates an example of a network for generating a personalized model based on a server, according to one embodiment.

Referring to FIG. 7, an example of a network for generating a personalized model using a model acquired from external electronic devices (710, 720, 730) is illustrated. For example, FIG. 7 illustrates an example of the network including one electronic device (101), one server (750), and three external electronic devices (710, 720, 730), but embodiments of the present disclosure are not limited thereto. For example, the number of components constituting the network may be changed.

Referring to FIG. 3, an exemplary situation is illustrated in which an electronic device (101), external electronic devices (710, 720, 730), and a server (750) are connected to each other based on a wired network and/or a wireless network. In FIG. 7, an example is illustrated in which the electronic device (101) and the external electronic device (730) and the external electronic device (710) and the external electronic device (720) are not connected to each other, but this is merely for convenience of explanation, and the embodiments of the present disclosure are not limited thereto. FIG. 7 assumes that the electronic device (101) and the external electronic device (730) and the external electronic device (710) and the external electronic device (720) are connected to each other. For example, the wired network may include a network such as the Internet, a local area network (LAN), a wide area network (WAN), or a combination thereof. For example, the wireless network may include a network such as long term evolution (LTE), 5g new radio (NR), wireless fidelity (WiFi), Zigbee, near field communication (NFC), Bluetooth, Bluetooth low-energy (BLE), or a combination thereof. Although the electronic device (101), external electronic devices (710, 720, 730), and server (750) are illustrated as being directly connected, the electronic device (101), external electronic devices (710, 720, 730), and server (750) may be indirectly connected via one or more routers and/or access points (APs).

Referring to FIG. 7, according to one embodiment, the electronic device (101) may include at least one of a processor (701), a communication circuit (703), and a memory (705). The processor (701) of FIG. 7 may correspond to the processor (301) of FIG. 3. The communication circuit (703) of FIG. 7 may correspond to the communication circuit (303) of FIG. 3. The memory (705) of FIG. 7 may correspond to the memory (305) of FIG. 3. In other words, specific details of the processor (701), the communication circuit (703), and the memory (705) of FIG. 7 may be substantially identically applied to the details of the processor (01), the communication circuit (303), and the memory (305) of FIG. 3.

In the example of FIG. 7, each of the external electronic devices (710, 720, 730) may include substantially the same components of the electronic device (101) (e.g., the processor (701), the communication circuit (703), and the memory (705)). For example, the external electronic device (710) may include a processor, a communication circuit, and a memory. For example, the external electronic device (720) may include a processor, a communication circuit, and a memory. For example, the external electronic device (730) may include a processor, a communication circuit, and a memory. The description of each of the processor, the communication circuit, and the memory may be substantially the same as the description of the processor (301), the communication circuit (303), and the memory (305).

Referring to FIG. 7, the server (750) may include a processor (751), a communication circuit (753), and a memory (755). The description of each of the processor (751), the communication circuit (753), and the memory (755) may be substantially identical to the description of the processor (701), the communication circuit (703), and the memory (705). For example, the server (750) may be associated with a specific service. For example, the specific service may include a domain.

Referring to FIG. 7, according to one embodiment, in the memory (755) of the server (750), one or more instructions (or commands) representing operations and/or actions to be performed on data by the processor (751) of the server (750) may be stored. A set of one or more instructions may be referred to as a program, firmware, an operating system, a process, a routine, a sub-routine, and/or an application. Hereinafter, when an application is installed in an electronic device (e.g., the server (750)), it may mean that one or more instructions provided in the form of an application are stored in the memory (755), and that the one or more applications are stored in a format executable by the processor of the electronic device (e.g., a file having an extension specified by the operating system of the server (750)). According to one embodiment, the server (750) may execute one or more instructions stored in the memory (755) to perform the operation of FIG. 8 or FIG. 9.

According to one embodiment, the programs installed in the server (750) may be classified into one of different layers, including an application layer, a framework layer, and/or a hardware abstraction layer (HAL), based on a target. For example, programs (e.g., drivers) designed to target the hardware (e.g., communication circuit (753)) of the server (750) may be classified within the hardware abstraction layer. For example, programs (e.g., reference model generation module (755-1), similarity determination module (755-2), and/or model training module (755-3)) designed to target at least one of the hardware abstraction layer and/or the application layer may be classified within the framework layer. Programs classified into the framework layer may provide an executable API (application programming interface) based on other programs.

Referring to FIG. 7, a server (750) according to one embodiment may generate (or train) and store a reference model for FL based on the execution of a reference model generation module (755-1). For example, the server (750) may include an API (application programming interface) to be used when the electronic device (101) or external electronic devices (710, 720, 730) download the reference model. For example, the API may include a rest API.

Referring to FIG. 7, a server (750) according to an embodiment may discriminate the similarity between a model of an electronic device (101) and models obtained from external electronic devices (710, 720, 730) based on the execution of a similarity determination module (755-2). For example, the server (750) may receive a first model of the electronic device (101), a second model of the external electronic device (710), a third model of the external electronic device (720), and a fourth model of the external electronic device (730). For example, the first model may be learned from the reference model by the electronic device (101) and may be a model for the service. For example, the second model may be learned from the reference model by the external electronic device (710) and may be a model for the service. For example, the third model may be learned from the reference model by the external electronic device (720) and may be a model for the service. For example, the fourth model may be learned from the reference model by the external electronic device (730) and may be a model for the service. In the above example, it is assumed that the first model, the second model, the third model, and the fourth model are models related to the service. However, the embodiment of the present disclosure is not limited thereto. For example, the server (750) may also obtain (or store) a fifth model for another service that is not related to the service from the external electronic device (710).

According to one embodiment, the server (750) can determine the similarity between the model of the electronic device (101) and the models of the external electronic devices (710, 720, 730). For example, the server (750) can identify a similar model with respect to the first model of the electronic device (101) based on the first model of the electronic device (101). For example, the server (750) can determine the similarity between each of the models (e.g., the second model, the third model, the fourth model) of the external electronic devices (710, 720, 730) and the first model of the electronic device (101). At this time, the server (750) can recognize the second model, the third model, and the fourth model, which are models for the service related to the first model. Based on the recognition that the fifth model is a model for the other service different from the first model, the server (750) can exclude the fifth model from the model for determining similarity.

According to one embodiment, the server (750) can determine the similarity by directly comparing the model of the electronic device (101) with the model of the external electronic device, or by inputting reference data to each of the model of the electronic device (101) and the model of the external electronic device and comparing the outputs thereof. For specific details related thereto, reference may be made to the content of the similarity determination module (305-1) of FIG. 3.

Referring to FIG. 7, a server (750) according to an embodiment may train a model (e.g., the first model) of an electronic device (101) based on the execution of a model training module (755-3). For example, if the server (750) determines that the first model of the electronic device (101) and the second model of the external electronic device (710) are similar, the server (750) may generate (or learn, update) a new model for the service from the first model using the second model. According to an embodiment, the server (750) may transmit the new model for the service to the electronic device (101). Thereafter, the electronic device (101) may provide the service based on the received new model.

In the above-described example, the server (750) is described as an example of operations performed based on the model (e.g., the first model) of the electronic device (101) for the service, but the embodiments of the present disclosure are not limited thereto. For example, the server (750) may also perform the operations for each of the external electronic devices (710, 720, 730).

FIG. 8 illustrates an example signal flow for a method for generating a personalized model based on a server, according to one embodiment.

The electronic device (101) of FIG. 8 may be an example of the electronic device (101) of FIG. 7. The external electronic device (710) of FIG. 8 may be an example of the external electronic device (710) of FIG. 7. The server (750) of FIG. 8 may be an example of the server (750) of FIG. 7. In the example of FIG. 8, an example of a method for generating a personalized model based on the server through one external electronic device (710) is illustrated, but the embodiments of the present disclosure are not limited thereto. For example, the embodiments of the present disclosure may include a method for generating a personalized model based on the server through a plurality of external electronic devices.

Referring to FIG. 8, in operation (800), the server (750) can generate a reference model. For example, the server (750) can generate the reference model for FL. For example, the server (750) can store the generated reference model. For example, the reference model can be a model for a service provided by the electronic device (101).

In operation (805), the server (750) can transmit the reference model. For example, the server (750) can establish connections with a plurality of electronic devices for the FL. For example, the server (750) can establish a connection with the electronic device (101). For example, the server (750) can establish a connection with an external electronic device (710). For example, the server (750) can transmit the reference model to each of the electronic device (101) and the external electronic device (101) with which the connection is established.

In operation (810), the electronic device (101) may obtain (or collect) input data. For example, the electronic device (101) may obtain the user's input data for the service. For example, the input data may include an input performed by the user to the electronic device (101). For example, the input data may include the user's utterance, the user's touch input, or the user's usage history. However, the embodiment of the present disclosure is not limited thereto. According to one embodiment, the input data may be obtained while the user of the electronic device (101) uses the service. For example, the electronic device (101) may provide the service using the reference model. For example, the electronic device (101) may generate output data from the input data using the reference model. The output data may be used to provide the service. While the service is provided using the reference model, the electronic device (101) may obtain the input data.

In operation (815), the electronic device (101) may generate a first model. For example, the electronic device (101) may generate (or train, learn, update) the first model using the input data from the reference model. For example, the first model may be a model for the service.

In operation (820), the external electronic device (710) may obtain (or collect) input data. For example, the external electronic device (710) may obtain the user's input data for the service. For example, the input data may include an input that the user performs with respect to the external electronic device (710). For example, the input data may include the user's utterance, the user's touch input, or the user's usage history. However, the embodiment of the present disclosure is not limited thereto. According to one embodiment, the input data may be obtained while the user of the external electronic device (710) uses the service. For example, the external electronic device (710) may provide the service using the reference model. For example, the external electronic device (710) may generate output data from the input data using the reference model. The output data may be used to provide the service. While the service is provided using the reference model, the external electronic device (710) may obtain the input data.

In operation (825), the external electronic device (710) may generate a second model. For example, the external electronic device (710) may generate (or train, learn, update) the second model using the input data from the reference model. For example, the second model may be a model for the service.

In FIG. 8, operations (810) and (815) are illustrated as being performed simultaneously with operations (820) and (825), but the embodiments of the present disclosure are not limited thereto. For example, operations (810) and (815) performed by the electronic device (101) may be performed before or after operations (820) and (825) performed by the external electronic device (710).

In operation (830), the electronic device (101) can transmit the first model. For example, the electronic device (101) can transmit the first model for the service to the server (750). In operation (835), the external electronic device (710) can transmit the second model. For example, the external electronic device (710) can transmit the second model for the service to the server (750).

In the above example, the electronic device (101) is described as an example of transmitting the first model to the server (750), but the embodiments of the present disclosure are not limited thereto. For example, the electronic device (101) may transmit location information of the electronic device (101) to the server (750) together with the first model. In addition, for example, the external electronic device (710) may transmit location information of the external electronic device (710) to the server (750) together with the second model. For example, the location information of the electronic device (101) or the location information of the external electronic device (710) may include global positioning system (GPS) coordinate information of the electronic device (101) or location measurement information based on a communication technique (e.g., WiFi, BLE).

In operation (840), the server (750) can determine the similarity between the models. For example, the server (750) can determine the similarity between the first model and the second model. For example, the server (750) can select the first model as a model to be a reference for determining the similarity. For example, the server (750) can identify another model (e.g., the second model) for the service provided by the selected first model. For example, the server (750) can determine whether a difference between the first model and the second model is less than a reference value. For example, determining the similarity can include a method of comparing the first model and the second model themselves, or a method of comparing vector values output from reference data input for the first model and the second model. For specific details on methods for determining the above similarity, reference may be made to the operation based on the model training module (305-1) of FIG. 3.

According to one embodiment, the server (750) may further use the location information received from the electronic device (101) to recognize the second model, which is a model for determining similarity with respect to the first model. For example, the server (750) may recognize a plurality of external electronic devices that use the service, and determine an external electronic device (710) that can be connected (or is likely to be connected) to the electronic device (101) by using the location information of the electronic device (101) among the plurality of external electronic devices. For example, the external electronic device (710) may be included in at least one external electronic device that is within a specified distance from the location of the electronic device (101) among the plurality of external electronic devices. For example, the server (750) may recognize the external electronic device (710) based on the location information obtained from each of the plurality of external electronic devices and the location information of the electronic device (101). Accordingly, the server (750) can recognize the first model and the second model of the external electronic device (710) that provides the same service as the first model and is located in a close area to the electronic device (101). For example, the server (750) can determine the similarity between the recognized first model and the second model.

In operation (845), the server (750) can generate a third model. For example, the server (750) can generate the third model from the first model using the second model. For example, the third model can be a model for the service. For example, the third model can be composed of third vector values having an average value of first vector values constituting the first model and second vector values constituting the second model. For example, each of the third vector values can have an average value of each of the first vector values and each of the second vector values.

In operation (850), the server (750) can transmit the third model. For example, the server (750) can transmit the third model to the electronic device (101). The third model is a model generated from the first model of the electronic device (101), and the server (750) can transmit the third model to the electronic device (101).

FIG. 9 illustrates an example of an operational flow for a method of generating a personalized model based on a server, according to one embodiment.

At least some of the above methods of FIG. 9 may be performed by the server (750) of FIG. 7. For example, at least some of the above methods may be controlled by the processor (701) of the server (750). In the following embodiments, the respective operations may be performed sequentially, but are not necessarily performed sequentially. For example, the order of the respective operations may be changed, and at least two operations may be performed in parallel. In FIG. 9, the server (750) may be referred to as an electronic device.

According to one embodiment, in operation (910), the server (750) can transmit the reference model to each of the first external electronic device, the second external electronic device, and the third external electronic device. For example, the server (750) can generate the reference model for FL. For example, the server (750) can store the generated reference model. For example, the reference model can be a model for a service provided by each of the first external electronic device (e.g., electronic device (101)), the third external electronic device, and the third external electronic device.

For example, the server (750) can establish connections with a plurality of electronic devices for the FL. For example, the server (750) can establish connections with the first external electronic device, the second external electronic device, and the third external electronic device. For example, the server (750) can transmit the reference model to each of the first external electronic device, the second external electronic device, and the third external electronic device with which connections are established.

According to one embodiment, in operation (920), the server (750) may obtain a first model generated from the reference model from the first external electronic device, obtain a second model generated from the reference model from the second external electronic device, and obtain a third model generated from the reference model from the third external electronic device.

For example, the first model may be a model for the service generated from the reference model by the first external electronic device. For example, the first model may be learned based on input (or input data) performed by a user of the first external electronic device to the first external electronic device. For example, the second model may be a model for the service generated from the reference model by the second external electronic device. For example, the second model may be learned based on input (or input data) performed by a user of the second external electronic device to the second external electronic device. For example, the third model may be a model for the service generated from the reference model by the third external electronic device. For example, the third model may be learned based on input (or input data) performed by a user of the third external electronic device to the third external electronic device.

In one embodiment, at operation (930), the server (750) can recognize a first difference between the first model and the second model and a second difference between the first model and the third model.

For example, the server (750) can determine the similarity between the first model and the second model (and the third model). For example, the server (750) can select the first model as a model to be a criterion for determining the similarity. For example, the server (750) can identify other models (e.g., the second model and the third model) for the service provided by the selected first model. For example, the server (750) can determine whether the first difference between the first model and the second model is less than a criterion value. For example, the server (750) can determine whether the second difference between the first model and the third model is less than the criterion value. For example, determining the similarity may include a method of comparing the first model and the second model (or the third model) themselves, or a method of comparing vector values output from reference data input for the first model and the second model (or the third model). For specific details on methods of determining the similarity, reference may be made to the operation based on the model training module (305-1) of FIG. 3.

In one embodiment, in operation (940), the server (750) can generate a fourth model for the service from the first model using the second model. For example, the server (750) can generate (or learn, update) the fourth model for the service from the first model using the second model based on the first difference being less than the reference value and the second difference being greater than or equal to the reference value. For example, the fourth model can be composed of third vector values having an average value of the first vector values constituting the first model and the second vector values constituting the second model. For example, each of the third vector values can have an average value of each of the first vector values and each of the second vector values.

According to one embodiment, the server (750) can transmit the fourth model. For example, the server (750) can transmit the fourth model to the first external electronic device. The fourth model is a model generated from the first model of the first external electronic device, and the server (750) can transmit the fourth model to the first external electronic device.

As described above, the electronic device (101) may include a memory (305) storing instructions and including one or more storage media. The electronic device (101) may include a communication circuit (303). The electronic device (101) may include at least one processor (301) including a processing circuit. The instructions, when individually or collectively executed by the at least one processor (301), may cause the electronic device (101) to obtain a reference model from a server (350) for federated learning via the communication circuit (303). The instructions, when individually or collectively executed by the at least one processor (301), may cause the electronic device (101) to generate a first model for the service from the reference model using user input data for use of the service. The instructions, when individually or collectively executed by the at least one processor (301), may cause the electronic device (101) to transmit first information indicative of the service to the server (350). The instructions, when individually or collectively executed by the at least one processor (301), may cause the electronic device (101) to obtain, in response to the first information, second information indicative of an external electronic device (310) that uses the service from the server (350). The instructions, when individually or collectively executed by the at least one processor (301), may cause the electronic device (101) to obtain a second model for the service of the external electronic device (310) from the external electronic device (310) with which a connection is established based on the second information. The instructions, when individually or collectively executed by the at least one processor (301), may cause the electronic device (101) to generate a third model for the service using the first model and the second model based on a difference between the first model and the second model being less than a reference value.

According to one embodiment, the instructions, when individually or collectively executed by the at least one processor (301), may cause the electronic device (101) to generate output data for providing the service from the input data collected while using the service, based on the reference model. The instructions, when individually or collectively executed by the at least one processor (301), may cause the electronic device (101) to generate the first model from the reference model using the input data, based on obtaining the user's input.

According to one embodiment, the reference model may be learned by the server (350). The second model may be generated for the service by the external electronic device (310) from the reference model.

According to one embodiment, the first information may include at least one of a name of the service or a name of an application providing the service. The second information may include at least one of a MAC (media access control) address of the external electronic device (310) or information for establishing the connection.

In one embodiment, the instructions, when individually or collectively executed by the at least one processor (301), may cause the electronic device (101) to transmit a signal to the external electronic device (310) requesting a model of the external electronic device (310) learned for the service from the reference model. The instructions, when individually or collectively executed by the at least one processor (301), may cause the electronic device (101) to, in response to the signal, obtain, from the external electronic device (310), the second model for the service. The instructions, when individually or collectively executed by the at least one processor (301), may cause the electronic device (101) to calculate the difference between the first model and the second model.

According to one embodiment, the instructions, when individually or collectively executed by the at least one processor (301), may cause the electronic device (101) to, in response to the first information, obtain from the server (350) the second information further instructing another external electronic device (320) to use the service. The instructions, when individually or collectively executed by the at least one processor (301), may cause the electronic device (101) to transmit a signal to the other external electronic device (320) with which a connection is established based on the second information, requesting a model of the other external electronic device (320) learned for the service from the reference model. The signal may include the generated first model. The above instructions, when individually or collectively executed by the at least one processor (301), may cause the electronic device (101) to obtain, from the other external electronic device (320), a fourth model for the service of the other external electronic device (320). A difference between the first model and the fourth model recognized by the other external electronic device (320) may be less than the reference value. The third model may be generated from the first model by further utilizing the fourth model.

In one embodiment, the instructions, when individually or collectively executed by the at least one processor (301), may cause the electronic device (101) to recognize first vector values of the first model and second vector values of the second model. The instructions, when individually or collectively executed by the at least one processor (301), may cause the electronic device (101) to calculate distance values between the first vector values and the second vector values. Each of the distance values may be calculated based on a difference between each of the first vector values and each of the second vector values. The instructions, when individually or collectively executed by the at least one processor (301), may cause the electronic device (101) to determine whether the difference between the first model and the second model, which is an average of the distance values, is less than the reference value.

In one embodiment, the instructions, when individually or collectively executed by the at least one processor (301), may cause the electronic device (101) to obtain first reference vector values generated from reference data based on the first model. The instructions, when individually or collectively executed by the at least one processor (301), may cause the electronic device (101) to obtain second reference vector values generated from the reference data based on the second model. The instructions, when individually or collectively executed by the at least one processor (301), may cause the electronic device (101) to calculate distance values between the first reference vector values and the second reference vector values. Each of the distance values may be calculated based on a difference between each of the first reference vector values and each of the second reference vector values. The above instructions, when individually or collectively executed by the at least one processor (301), may cause the electronic device (101) to determine whether the difference between the first model and the second model, which is an average of the distance values, is less than the reference value.

In one embodiment, the instructions, when individually or collectively executed by the at least one processor (301), may cause the electronic device (101) to generate a fifth model for another service from another reference model obtained from the server (350) using the input data while generating the first model for the service. The instructions, when individually or collectively executed by the at least one processor (301), may cause the electronic device (101) to generate a seventh model for the another service from the fifth model using a sixth model for the another service obtained from the external electronic device (310) together with the second model while generating the third model for the service. The other service may be related to the service.

According to one embodiment, the instructions, when individually or collectively executed by the at least one processor (301), may cause the electronic device (101) to transmit the third model to the server (350). The third model may be configured based on third vector values that are an average between first vector values of the first model and second vector values of the second model. Each of the third vector values may be an average of each of the first vector values and each of the second vector values.

According to one embodiment, the instructions, when individually or collectively executed by the at least one processor (301), may cause the electronic device (101) to collect other input data of the user for use of the service. The instructions, when individually or collectively executed by the at least one processor (301), may cause the electronic device (101) to generate output data for providing the service from the other input data collected while using the service, based on the third model.

According to one embodiment, if the service is a voice recognition service, the other input data may include speech information of the user, and the output data may include text representing the speech information.

As described above, the electronic device (750) may include a memory (755) storing instructions and including one or more storage media. The electronic device (750) may include a communication circuit (753). The electronic device (750) may include at least one processor (751) including a processing circuit. The instructions, when individually or collectively executed by the at least one processor (751), may cause the electronic device (750) to transmit a reference model generated for federated learning to each of the first external electronic device (101), the second external electronic device (710), and the third external electronic device (720). The instructions, when individually or collectively executed by the at least one processor (751), may cause the electronic device (750) to obtain, via the communication circuit (753), a first model for a service generated from the reference model from the first external electronic device (101), a second model for the service generated from the reference model from the second external electronic device (710), and a third model for the service generated from the reference model from the third external electronic device (720). The instructions, when individually or collectively executed by the at least one processor (751), may cause the electronic device (750) to recognize a first difference between the first model and the second model and a second difference between the first model and the third model. The above instructions, when individually or collectively executed by the at least one processor (751), may cause the electronic device (750) to generate a fourth model for the service from the first model using the second model based on the first difference being less than the reference value and the second difference being greater than or equal to the reference value.

In one embodiment, the instructions, when individually or collectively executed by the at least one processor (751), may cause the electronic device (750) to obtain, from a fourth external electronic device (730) via the communication circuit (753), a fifth model for another service generated from the reference model. The instructions, when individually or collectively executed by the at least one processor (751), may cause the electronic device (750) to determine, using device information connected to the electronic device (750), the second model and the third model as models associated with the service with respect to the first model.

According to one embodiment, the device information may include first service information that can be provided by the first external electronic device (101) and a MAC (media access control) address of the first external electronic device (101), second service information that can be provided by the second external electronic device (710) and a MAC address of the second external electronic device (710), third service information that can be provided by the third external electronic device (720) and a MAC address of the third external electronic device (720), and fourth service information that can be provided by the fourth external electronic device (730) and a MAC address of the fourth external electronic device (730).

In one embodiment, the instructions, when individually or collectively executed by the at least one processor (751), may cause the electronic device (750) to recognize first vector values of the first model and second vector values of the second model. The instructions, when individually or collectively executed by the at least one processor (751), may cause the electronic device (750) to calculate distance values between the first vector values and the second vector values. Each of the distance values may be calculated based on a difference between each of the first vector values and each of the second vector values. The instructions, when individually or collectively executed by the at least one processor (751), may cause the electronic device (750) to determine whether the difference between the first model and the second model, which is an average of the distance values, is less than the reference value.

In one embodiment, the instructions, when individually or collectively executed by the at least one processor (751), may cause the electronic device (750) to obtain first reference vector values generated from reference data based on the first model. The instructions, when individually or collectively executed by the at least one processor (751), may cause the electronic device (750) to obtain second reference vector values generated from the reference data based on the second model. The instructions, when individually or collectively executed by the at least one processor (751), may cause the electronic device (750) to calculate distance values between the first reference vector values and the second reference vector values. Each of the distance values may be calculated based on a difference between each of the first reference vector values and each of the second reference vector values. The above instructions, when individually or collectively executed by the at least one processor (751), may cause the electronic device (750) to determine whether the difference between the first model and the second model, which is an average of the distance values, is less than the reference value.

According to one embodiment, the instructions, when individually or collectively executed by the at least one processor (751), may cause the electronic device (750) to transmit the fourth model to the first external electronic device (101) via the communication circuit (753). The fourth model may be configured based on third vector values that are an average between first vector values of the first model and second vector values of the second model. Each of the third vector values may be an average of each of the first vector values and each of the second vector values.

In a method performed by an electronic device (101) as described above, the method may include an operation of obtaining a reference model from a server (350) for federated learning. The method may include an operation of generating a first model for the service from the reference model by using input data of a user for using the service. The method may include an operation of transmitting first information indicating the service to the server (350). The method may include an operation of obtaining, in response to the first information, second information indicating an external electronic device (310) using the service from the server (350). The method may include an operation of obtaining, from the external electronic device (310) with which a connection is established based on the second information, a second model for the service of the external electronic device (310). The method may include an operation of generating a third model for the service using the first model and the second model, based on a difference between the first model and the second model being less than a reference value.

The non-transitory computer-readable storage medium as described above can store one or more programs including instructions that, when individually or collectively executed by at least one processor (301) of an electronic device (101) including a communication circuit (303), cause the electronic device (101) to obtain a reference model from a server (350) for federated learning via the communication circuit (303). The non-transitory computer-readable storage medium can store one or more programs including instructions that, when individually or collectively executed by the at least one processor (301), cause the electronic device (101) to generate a first model for the service from the reference model using user input data for use of the service. The non-transitory computer-readable storage medium may store one or more programs including instructions that, when individually or collectively executed by the at least one processor (301), cause the server (350) to transmit first information indicating the service. The non-transitory computer-readable storage medium may store one or more programs including instructions that, when individually or collectively executed by the at least one processor (301), cause the server (350) to acquire second information indicating an external electronic device (310) that uses the service in response to the first information. The non-transitory computer-readable storage medium may store one or more programs including instructions that, when individually or collectively executed by the at least one processor (301), cause the server (350) to acquire a second model for the service of the external electronic device (310) from the external electronic device (310) to which a connection is established based on the second information. The non-transitory computer-readable storage medium may store one or more programs including instructions that, when individually or collectively executed by the at least one processor (301), cause the third model for the service to be generated using the first model and the second model based on a difference between the first model and the second model being less than a reference value.

The method performed by the electronic device (750) as described above may include an operation in which the electronic device (750) transmits a reference model generated for federated learning to each of the first external electronic device (101), the second external electronic device (710), and the third external electronic device (720). The method may include an operation in which the first model for a service generated from the reference model is acquired from the first external electronic device (101), a second model for the service generated from the reference model is acquired from the second external electronic device (710), and a third model for the service generated from the reference model is acquired from the third external electronic device (720) via the communication circuit (753). The method may include an operation in which the first difference between the first model and the second model and the second difference between the first model and the third model are recognized. The method may include an operation of generating a fourth model for the service from the first model using the second model, based on the first difference being less than the reference value and the second difference being greater than or equal to the reference value.

The non-transitory computer-readable storage medium as described above can store one or more programs including instructions that, when individually or collectively executed by at least one processor (751) of an electronic device (750) including a communication circuit (753), cause the electronic device (750) to transmit a reference model generated for federated learning to each of a first external electronic device (101), a second external electronic device (710), and a third external electronic device (720). The non-transitory computer-readable storage medium may store one or more programs including instructions that, when individually or collectively executed by the at least one processor (751), cause the electronic device (750) to obtain a first model for a service generated from the reference model from the first external electronic device (101), a second model for the service generated from the reference model from the second external electronic device (710), and a third model for the service generated from the reference model from the third external electronic device (720), via the communication circuit (753). The non-transitory computer-readable storage medium may store one or more programs including instructions that, when individually or collectively executed by the at least one processor (751), cause the electronic device (750) to recognize a first difference between the first model and the second model and a second difference between the first model and the third model. The non-transitory computer-readable storage medium may store one or more programs including instructions that, when individually or collectively executed by the at least one processor (751), cause the fourth model for the service to be generated from the first model using the second model based on the first difference being less than the reference value and the second difference being greater than or equal to the reference value.

The electronic devices according to various embodiments disclosed in this document may be devices of various forms. The electronic devices may include, for example, portable communication devices (e.g., smartphones), computer devices, portable multimedia devices, portable medical devices, cameras, wearable devices, or home appliance devices. The electronic devices according to embodiments of this document are not limited to the above-described devices.

It should be understood that the various embodiments of this document and the terminology used herein are not intended to limit the technical features described in this document to specific embodiments, but include various modifications, equivalents, or substitutes of the embodiments. In connection with the description of the drawings, similar reference numerals may be used for similar or related components. The singular form of a noun corresponding to an item may include one or more of the items, unless the context clearly dictates otherwise. In this document, each of the phrases "A or B", "at least one of A and B", "at least one of A or B", "A, B, or C", "at least one of A, B, and C", and "at least one of A, B, or C" can include any one of the items listed together in the corresponding phrase, or all possible combinations thereof. Terms such as "first", "second", or "first" or "second" may be used merely to distinguish one component from another, and do not limit the components in any other respect (e.g., importance or order). When a component (e.g., a first component) is referred to as "coupled" or "connected" to another (e.g., a second component), with or without the terms "functionally" or "communicatively," it means that the component can be connected to the other component directly (e.g., wired), wirelessly, or through a third component.

The term "module" used in various embodiments of this document may include a unit implemented in hardware, software or firmware, and may be used interchangeably with terms such as logic, logic block, component, or circuit, for example. A module may be an integrally configured component or a minimum unit of the component or a portion thereof that performs one or more functions. For example, according to one embodiment, a module may be implemented in the form of an application-specific integrated circuit (ASIC).

Various embodiments of the present document may be implemented as software (e.g., a program (140)) including one or more instructions stored in a storage medium (e.g., an internal memory (136) or an external memory (138)) readable by a machine (e.g., an electronic device (101)). For example, a processor (e.g., a processor (120)) of the machine (e.g., an electronic device (101)) may call at least one instruction among the one or more instructions stored from the storage medium and execute it. This enables the machine to operate to perform at least one function according to the called at least one instruction. The one or more instructions may include code generated by a compiler or code executable by an interpreter. The machine-readable storage medium may be provided in the form of a non-transitory storage medium. Here, 'non-transitory' simply means that the storage medium is a tangible device and does not contain signals (e.g. electromagnetic waves), and the term does not distinguish between cases where data is stored semi-permanently or temporarily on the storage medium.

According to one embodiment, the method according to various embodiments disclosed in the present document may be provided as included in a computer program product. The computer program product may be traded between a seller and a buyer as a commodity. The computer program product may be distributed in the form of a machine-readable storage medium (e.g., a compact disc read only memory (CD-ROM)), or may be distributed online (e.g., downloaded or uploaded) via an application store (e.g., Play Store™) or directly between two user devices (e.g., smart phones). In the case of online distribution, at least a part of the computer program product may be at least temporarily stored or temporarily generated in a machine-readable storage medium, such as a memory of a manufacturer's server, a server of an application store, or an intermediary server.

According to various embodiments, each component (e.g., a module or a program) of the above-described components may include a single or multiple entities, and some of the multiple entities may be separately arranged in other components. According to various embodiments, one or more components or operations of the above-described components may be omitted, or one or more other components or operations may be added. Alternatively or additionally, the multiple components (e.g., a module or a program) may be integrated into one component. In such a case, the integrated component may perform one or more functions of each of the multiple components identically or similarly to those performed by the corresponding component of the multiple components before the integration. According to various embodiments, the operations performed by the module, program, or other component may be executed sequentially, in parallel, repeatedly, or heuristically, or one or more of the operations may be executed in a different order, omitted, or one or more other operations may be added.

Claims (15)

In an electronic device (101), A memory (305) storing instructions and including one or more storage media; a communication circuit (303); and At least one processor (301) comprising a processing circuit, The above instructions, when individually or collectively executed by the at least one processor (301), cause the electronic device (101) to: Through the above communication circuit (303), a reference model is obtained from a server (350) for federated learning; Using user input data for use of the service, a first model for the service is created from the reference model; Transmitting first information instructing the above service to the server (350); In response to the first information, second information is obtained from the server (350) that instructs an external electronic device (310) to use the service; Obtaining a second model for the service of the external electronic device (310) from the external electronic device (310) with which a connection is established based on the second information; and Based on the difference between the first model and the second model which is less than a reference value, causing a third model for the service to be generated using the first model and the second model. Electronic device (101). In claim 1, The above instructions, when individually or collectively executed by the at least one processor (301), cause the electronic device (101) to: Based on the above reference model, generating output data for providing the service from the input data collected while using the service; and Based on obtaining the input of the user, causing the first model to be generated from the reference model using the input data. Electronic device (101). In claim 1, The above reference model is learned by the server (350), and The second model is generated for the service by the external electronic device (310) from the reference model. Electronic device (101). In claim 1, The first information includes at least one of the name of the service or the name of the application providing the service, and The second information includes at least one of the MAC (media access control) address of the external electronic device (310) or information for establishing the connection. Electronic device (101). In claim 1, The above instructions, when individually or collectively executed by the at least one processor (301), cause the electronic device (101) to: Transmitting a signal to the external electronic device (310) requesting a model of the external electronic device (310) learned for the service from the reference model; In response to the signal, obtaining the second model for the service from the external electronic device (310); and causing the difference between the first model and the second model to be calculated, Electronic device (101). In claim 1, The above instructions, when individually or collectively executed by the at least one processor (301), cause the electronic device (101) to: In response to the first information, the second information is obtained from the server (350) to further instruct another external electronic device (320) to use the service; Transmitting a signal to the other external electronic device (320) with which a connection is established based on the second information, requesting a model of the other external electronic device (320) learned for the service from the reference model, the signal including the generated first model; and To cause the other external electronic device (320) to obtain a fourth model for the service of the other external electronic device (320), The difference between the first model and the fourth model recognized by the other external electronic device (320) is less than the reference value, and The third model is generated by further utilizing the fourth model from the first model. Electronic device (101). In claim 1, The above instructions, when individually or collectively executed by the at least one processor (301), cause the electronic device (101) to: Recognize first vector values of the first model and second vector values of the second model; Calculating distance values between the first vector values and the second vector values, each of the distance values being calculated based on a difference between each of the first vector values and each of the second vector values; and causing a determination as to whether the difference between the first model and the second model, which is an average of the distance values, is less than the reference value; Electronic device (101). In claim 1, The above instructions, when individually or collectively executed by the at least one processor (301), cause the electronic device (101) to: Based on the first model, first reference vector values generated from reference data are obtained; Based on the second model, second reference vector values generated from the reference data are obtained; Calculating distance values between the first reference vector values and the second reference vector values, each of the distance values being calculated based on a difference between each of the first reference vector values and each of the second reference vector values; and causing a determination as to whether the difference between the first model and the second model, which is an average of the distance values, is less than the reference value; Electronic devices (101). In claim 1, The above instructions, when individually or collectively executed by the at least one processor (301), cause the electronic device (101) to: While generating the first model for the above service, generating a fifth model for another service from another reference model obtained from the server (350) using the input data; and While generating the third model for the above service, cause the seventh model for the other service to be generated from the fifth model by using the sixth model for the other service obtained together with the second model from the external electronic device (310). The above other services, related to the above services, Electronic devices (101). In claim 1, The above instructions, when individually or collectively executed by the at least one processor (301), cause the electronic device (101) to: Causing the above third model to be transmitted to the server (350), The third model is configured based on third vector values which are an average value between the first vector values of the first model and the second vector values of the second model, and Each of the above third vector values is an average of each of the above first vector values and each of the above second vector values. Electronic devices (101). In claim 1, The above instructions, when individually or collectively executed by the at least one processor (301), cause the electronic device (101) to: Collect other input data of the user for use of the above service; and Based on the third model, causing the output data for providing the service to be generated from the other input data collected while using the service. Electronic devices (101). In claim 11, If the above service is a voice recognition service, the other input data includes the user's speech information, and the output data includes text representing the speech information. Electronic devices (101). In an electronic device (750), A memory (755) storing instructions and including one or more storage media; a communication circuit (753); and At least one processor (751) comprising a processing circuit, The above instructions, when individually or collectively executed by the at least one processor (751), cause the electronic device (750) to: A reference model generated for federated learning is transmitted to each of the first external electronic device (101), the second external electronic device (710), and the third external electronic device (720); Obtaining a first model for a service generated from the above-mentioned reference model from the first external electronic device (101), obtaining a second model for the service generated from the above-mentioned reference model from the second external electronic device (710), and obtaining a third model for the service generated from the above-mentioned reference model from the third external electronic device (720) through the communication circuit (753); Recognizing a first difference between the first model and the second model and a second difference between the first model and the third model; and Based on the first difference being less than the reference value and the second difference being greater than the reference value, causing a fourth model for the service to be generated from the first model using the second model. Electronic devices (750). A method performed by an electronic device (101), the method comprising: An operation of obtaining a reference model from a server (350) for federated learning; An operation of generating a first model for the service from the reference model using user input data for using the service; An action of transmitting first information instructing the above service to the server (350); In response to the first information, an operation of obtaining second information from the server (350) that instructs an external electronic device (310) to use the service; An operation of obtaining a second model for the service of the external electronic device (310) from the external electronic device (310) to which a connection is established based on the second information; and An operation of generating a third model for the service using the first model and the second model, based on a difference between the first model and the second model that is less than a reference value, method. A non-transitory computer-readable storage medium, when individually or collectively executed by at least one processor (301) of an electronic device (101) including a communication circuit (303), causes the electronic device (101) to: Through the above communication circuit (303), a reference model is obtained from a server (350) for federated learning; Using user input data for use of the service, a first model for the service is created from the reference model; Transmitting first information instructing the above service to the server (350); In response to the first information, second information is obtained from the server (350) that instructs an external electronic device (310) to use the service; Obtaining a second model for the service of the external electronic device (310) from the external electronic device (310) with which a connection is established based on the second information; and storing one or more programs including instructions causing a third model for the service to be generated using the first model and the second model, based on a difference between the first model and the second model that is less than a reference value; A non-transitory computer-readable storage medium.

Images