Electronic device and method for delegating training of artificial intelligence model, and non-transitory computer-readable recording medium

Abstract

An electronic device is disclosed. The electronic device may receive, from a server through a communication circuit, a request to train an artificial intelligence (AI) model. The electronic device may determine whether to delegate training of the AI model. The electronic apparatus, on the basis of determining to delegate the training of the AI model, may request, through the communication circuit, the training of the AI model from at least one electronic apparatus capable of training the AI model among one or more candidate electronic apparatuses having a relationship of specified conditions with respect to the electronic apparatus. The electronic device may perform the training of the AI model on the basis of determining not to delegate the training of the AI model. The electronic device, on the basis of determining not to delegate the training of the AI model, may transmit, to the server through the communication circuit, data indicating the result of performing the training.

Description

Electronic device, method, and non-transitory computer-readable recording medium for delegating training of an artificial intelligence model

[0001] The following descriptions relate to electronic devices, methods, and non-transitory computer-readable recording media for delegating the training of an artificial intelligence (AI) model.

[0002] Artificial intelligence (AI) systems can learn through machine learning. Machine learning can refer to algorithms that classify and / or learn the characteristics of input data on their own.

[0003] One example of a method for performing machine learning is federated learning (FL). For federated learning, each of multiple electronic devices can train an AI model using training data and transmit the updated AI model parameters to a server. For federated learning, the server can update the AI ​​model based on parameters acquired from multiple electronic devices.

[0004] The above information may be provided as background art to aid 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.

[0005] An electronic device is disclosed. The electronic device may include a communication circuit. The electronic device may include at least one processor including a processing circuit. The electronic device may include a memory storing instructions and including one or more storage media. The instructions, when individually or collectively executed by the at least one processor, may cause the electronic device to receive a request for training an artificial intelligence (AI) model from a server via the communication circuit. The instructions, when individually or collectively executed by the at least one processor, may cause the electronic device to determine whether to delegate training of the AI ​​model. The instructions, when individually or collectively executed by the at least one processor, may cause the electronic device to request training of the AI ​​model from at least one electronic device capable of training the AI ​​model among one or more candidate electronic devices having a relationship with a specified condition to the electronic device via the communication circuit, based on the electronic device's decision to delegate training of the AI ​​model. The instructions, when individually or collectively executed by the at least one processor, may cause the electronic device to perform the training of the AI ​​model based on a determination not to delegate the training of the AI ​​model. The instructions, when individually or collectively executed by the at least one processor, may cause the electronic device to transmit data representing a result of performing the training to the server via the communication circuit based on a determination not to delegate the training of the AI ​​model.

[0006] An electronic device is disclosed. The electronic device may include a communication circuit. The electronic device may include at least one processor including a processing circuit. The electronic device may include a memory storing instructions and including one or more storage media. The instructions, when individually or collectively executed by the at least one processor, may cause the electronic device to receive a request for learning an AI model from another electronic device having a relationship with the electronic device that has a specified conditional relationship with the electronic device via the communication circuit. The instructions, when individually or collectively executed by the at least one processor, may cause the electronic device to determine whether learning of the AI ​​model is possible. The instructions, when individually or collectively executed by the at least one processor, may cause the electronic device to perform learning of the AI ​​model based on a determination that learning of the AI ​​model is possible. The above instructions, when individually or collectively executed by the at least one processor, may cause the electronic device to transmit, through the communication circuit, data representing the results of the performance of the learning to a server that requested the learning of the AI ​​model.

[0007] A method is disclosed. The method may be performed by an electronic device including a communication circuit. The method may include an operation of receiving a request for training an AI model from a server via the communication circuit. The method may include an operation of determining whether to delegate training of the AI ​​model. Based on a decision to delegate the training of the AI ​​model, the method may include an operation of requesting, via the communication circuit, training of the AI ​​model to at least one electronic device capable of training the AI ​​model among one or more candidate electronic devices having a relationship with a specified condition to the electronic device. Based on a decision not to delegate the training of the AI ​​model, the method may include an operation of performing the training of the AI ​​model. The method may include an operation of transmitting, via the communication circuit, data indicating a result of performing the training to the server.

[0008] A method is disclosed. The method can be performed by an electronic device including a communication circuit. The method can include receiving a request for learning an AI model from another electronic device having a relationship with a specified condition to the electronic device through the communication circuit. The method can include determining whether learning of the AI ​​model is possible. The method can include performing the learning of the AI ​​model based on determining that the learning of the AI ​​model is possible. The method can include transmitting data representing the result of performing the learning to a server that requested the learning of the AI ​​model through the communication circuit.

[0009] A non-transitory computer-readable storage medium is disclosed. The non-transitory computer-readable storage medium may store a program including instructions. The instructions, when individually or collectively executed by at least one processor of an electronic device including a communication circuit, may cause the electronic device to receive a request for training an AI model from a server via the communication circuit. The instructions, when individually or collectively executed by the at least one processor, may cause the electronic device to determine whether to delegate training of the AI ​​model. The instructions, when individually or collectively executed by the at least one processor, may cause the electronic device to request training of the AI ​​model from at least one electronic device capable of training the AI ​​model among one or more candidate electronic devices having a relationship with a specified condition to the electronic device via the communication circuit, based on the electronic device's decision to delegate training of the AI ​​model. The instructions, when individually or collectively executed by the at least one processor, may cause the electronic device to perform the training of the AI ​​model based on a determination not to delegate the training of the AI ​​model. The instructions, when individually or collectively executed by the at least one processor, may cause the electronic device to transmit data representing a result of performing the training to the server via the communication circuit based on a determination not to delegate the training of the AI ​​model.

[0010] A non-transitory computer-readable recording medium is disclosed. The non-transitory computer-readable recording medium can store a program including instructions. The instructions, when individually or collectively executed by at least one processor of an electronic device including a communication circuit, can cause the electronic device to receive a request for learning an AI model from another electronic device having a relationship with a specified condition to the electronic device through the communication circuit. The instructions, when individually or collectively executed by the at least one processor, can cause the electronic device to determine whether learning of the AI ​​model is possible. The instructions, when individually or collectively executed by the at least one processor, can cause the electronic device to perform learning of the AI ​​model based on determining that learning of the AI ​​model is possible. The above instructions, when individually or collectively executed by the at least one processor, may cause the electronic device to transmit, through the communication circuit, data representing the results of the performance of the learning to a server that requested the learning of the AI ​​model.

[0011] In connection with the description of the drawings, the same or similar reference numerals may be used for the same or similar components.

[0012] FIG. 1 is a block diagram of an electronic device within a network environment according to various embodiments.

[0013] FIG. 2a illustrates an example of a network environment for learning an artificial intelligence (AI) model according to one embodiment.

[0014] FIG. 2b illustrates an example of electronic devices of a trust relationship for delegating learning of an AI model, according to one embodiment.

[0015] FIG. 3 illustrates an example of programs for learning an AI model according to one embodiment.

[0016] Figure 4 is a flowchart of operations for learning an AI model according to one embodiment.

[0017] FIG. 5 is a flowchart of operations for delegating learning of an AI model performed by an electronic device according to an embodiment.

[0018] FIG. 6 is a flowchart of operations for learning an AI model assigned to an electronic device according to one embodiment.

[0019] FIG. 7 is a flowchart of operations for transmitting learning results of an AI model performed by an electronic device according to an embodiment.

[0020] FIG. 8 is a flowchart of operations for transmitting learning results of an AI model assigned to an electronic device according to an embodiment.

[0021] FIG. 9 illustrates an example of a user interface (UI) displayed in electronic devices according to one embodiment.

[0022] Hereinafter, embodiments of the present disclosure will be described in detail with reference to the drawings so that those skilled in the art can easily implement the present disclosure. However, the present disclosure may be implemented in various different forms and is not limited to the embodiments described herein. In connection with the description of the drawings, the same or similar reference numerals may be used for identical or similar components. Furthermore, in the drawings and related descriptions, descriptions of well-known functions and configurations may be omitted for clarity and conciseness.

[0023] FIG. 1 is a block diagram of an electronic device (101) within a network environment (100) according to various embodiments.

[0024] 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)).

[0025] The processor (120) may, for example, execute software (e.g., a program (140)) to control at least one other component (e.g., a hardware or software component) of the electronic device (101) connected to the processor (120) and perform various data processing or operations. According to one embodiment, as at least a part of the data processing or operations, the processor (120) may store commands or data received from other components (e.g., a sensor module (176) or a communication module (190)) in a volatile memory (132), process the commands or data stored in the volatile memory (132), and store result data in a non-volatile 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 graphics 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 with the main processor (121). For example, when the electronic device (101) includes the main processor (121) and the auxiliary processor (123), the auxiliary processor (123) may be configured to use less power than the main processor (121) or to be specialized for a given function. The auxiliary processor (123) may be implemented separately from the main processor (121) or as a part thereof.

[0026] The auxiliary processor (123) may control at least a portion of functions or states associated with at least one component (e.g., a display module (160), a sensor module (176), or a communication module (190)) of the electronic device (101), 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 unit) may include a hardware structure specialized for processing artificial intelligence models. The artificial intelligence models may be generated through machine learning. This learning can be performed, for example, on the electronic device (101) itself where the artificial intelligence model is executed, or can be performed through a separate server (e.g., server (108)). The learning algorithm can 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 can include multiple 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), a deep Q-network, or a combination of two or more of the above, but is not limited to the examples described above. In addition to, or alternatively to, a hardware structure, an artificial intelligence model may include a software structure.

[0027] 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 non-volatile memory (134).

[0028] 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).

[0029] 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).

[0030] The audio output module (155) can output audio signals 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 incoming calls. In one embodiment, the receiver can be implemented separately from the speaker or as part of the speaker.

[0031] 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) may 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) may 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.

[0032] 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 acquire sound through the input module (150), output sound through the sound output module (155), or an external electronic device (e.g., electronic device (102)) (e.g., speaker or headphone) directly or wirelessly connected to the electronic device (101).

[0033] The sensor module (176) can detect the operating status (e.g., power or temperature) of the electronic device (101) or the external environmental status (e.g., user status) and generate an electrical signal or data value corresponding to the detected status. 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.

[0034] 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)). In 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.

[0035] 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., 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).

[0036] The haptic module (179) can convert electrical signals into mechanical stimuli (e.g., vibration or movement) or electrical stimuli that a user can perceive through tactile or kinesthetic sensations. According to one embodiment, the haptic module (179) can include, for example, a motor, a piezoelectric element, or an electrical stimulation device.

[0037] The camera module (180) can capture still images and videos. According to one embodiment, the camera module (180) may include one or more lenses, image sensors, image signal processors, or flashes.

[0038] 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).

[0039] A battery (189) may power at least one component of the electronic device (101). In one embodiment, the battery (189) may include, for example, a non-rechargeable primary battery, a rechargeable secondary battery, or a fuel cell.

[0040] The communication module (190) may support the 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., electronic device (102), electronic device (104), or server (108)), and the performance of communication through the established communication channel. The communication module (190) may operate independently from the processor (120) (e.g., 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 global navigation satellite system (GNSS) 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, the corresponding communication module can 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 can 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) can verify or authenticate the electronic device (101) within a communication network such as the first network (198) or the second network (199) by using subscriber information (e.g., an international mobile subscriber identity (IMSI)) stored in the subscriber identification module (196).

[0041] The wireless communication module (192) can support 5G networks and next-generation communication technologies following the 4G network, such as NR access technology (new radio access technology). The NR access technology can support high-speed transmission of high-capacity data (eMBB (enhanced mobile broadband)), minimization of terminal power 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) can 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) can support various requirements specified in the electronic device (101), an external electronic device (e.g., the electronic device (104)), or a network system (e.g., the second network (199)). According to one embodiment, the wireless communication module (192) can support a peak data rate (e.g., 20 Gbps or more) for realizing eMBB, a loss coverage (e.g., 664 dB or less) for realizing mMTC, or a U-plane latency (e.g., 0.5 ms or less for downlink (DL) and uplink (UL), or 6 ms or less for round trip) for realizing URLLC.

[0042] The antenna module (197) can transmit or receive signals or power to or from an external device (e.g., an external electronic device). In one embodiment, the antenna module (197) may include an antenna including a radiator formed of a conductor or a conductive pattern formed on a substrate (e.g., a PCB). In one embodiment, the antenna module (197) may 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), may be selected from the plurality of antennas, for example, by the communication module (190). A signal or power may be transmitted or received between the communication module (190) and an external electronic device via the at least one selected antenna. In some embodiments, in addition to the radiator, another component (e.g., a radio frequency integrated circuit (RFIC)) may be additionally formed as a part of the antenna module (197).

[0043] According to various embodiments, the antenna module (197) may form a mmWave antenna module. In one embodiment, the mmWave antenna module may include a printed circuit board, an RFIC disposed 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) disposed on or adjacent a second side (e.g., a top side or a side side) of the printed circuit board and capable of transmitting or receiving signals in the designated high-frequency band.

[0044] At least some of the above components can be interconnected and exchange signals (e.g., commands or data) with each other via a communication method between peripheral devices (e.g., a bus, GPIO (general purpose input and output), SPI (serial peripheral interface), or MIPI (mobile industry processor interface)).

[0045] According to 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). According to 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 the function or at least a part of the service. One or more external electronic devices that receive the request may execute at least a portion 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 portion of a response to the request. For this purpose, cloud computing, distributed computing, mobile edge computing (MEC), or client-server computing technology may be used, for example. The electronic device (101) may provide an ultra-low latency service by using distributed computing or mobile edge computing, for example. In another embodiment, the external electronic device (104) may include an Internet of Things (IoT) device. The server (108) may be an intelligent server utilizing 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.

[0046] FIG. 2a illustrates an example of a network environment for training an artificial intelligence (AI) model, according to one embodiment. FIG. 2b illustrates an example of electronic devices in a trust relationship for delegating AI model training, according to one embodiment.

[0047] Referring to FIG. 2A, a network environment for learning an AI model may include a server (108) and electronic devices (101, 201, 205, 211, 213, 215). In one embodiment, the server (108) and electronic devices (101, 201, 205, 211, 213, 215) may include at least some of the components of the electronic device (101) of FIG. 1 (e.g., a processor (120), a memory (130), and a communication module (190)).

[0048] Referring to FIG. 2B, the electronic device (101) may include a processor (120), a memory (130), and a communication circuit (290). In one embodiment, each of the electronic devices (211, 213, 215) may include a processor (221, 223, 225), a memory (231, 233, 235), and a communication circuit (291, 293, 295).

[0049] In one embodiment, each of the processors (120, 221, 223, 225) may be utilized to execute the operations exemplified in the descriptions of FIGS. 4 to 8. For example, each of the processors (221, 223, 225) may include at least a portion of the processor (120) of FIG. 1 or may correspond to at least a portion of the processor (120) of FIG. 1. For example, each of the processors (120, 221, 223, 225) may include one or more processors including an application processor (AP) and / or a communication processor (CP). For example, each of the processors (120, 221, 223, 225) may be implemented as a single chip, such as a system on chip (SoC), or may be implemented as multiple chips. For example, each of the processors (120, 221, 223, 225) may be implemented as a single integrated circuit or may be implemented as multiple integrated circuits. For example, each of the processors (120, 221, 223, 225) may be distributedly arranged within an electronic device. For example, the processors (120, 221, 223, 225) may correspond to multiple processors that collectively perform multiple operations by dividing them among the processors.

[0050] In one embodiment, each of the memories (130, 231, 233, 235) may (at least temporarily) store instructions for executing the operations exemplified in the descriptions of FIGS. 4 through 8. The instructions may be executed by the processor (120, 221, 223, 225). The instructions may be included in one or more programs stored in each of the memories (130, 231, 233, 235). For example, each of the memories (231, 233, 235) may include at least a portion of the memory (130) of FIG. 1 (or at least a portion of the non-volatile memory (134)) or may correspond to at least a portion of the memory (130) of FIG. 1 (or at least a portion of the non-volatile memory (134)). For example, each of the memories (130, 231, 233, 235) may include a main memory (e.g., a random access memory (RAM), a register for the processor (120, 221, 223, 225), a cache for the processor (120, 221, 223, 225), a register for the communication circuit (290, 291, 293, 295), a buffer (or soft buffer) for the communication circuit (290, 291, 293, 295), and / or an auxiliary memory (e.g., a hard disk drive (HDD), a solid state drive (SSD)). For example, each of the memories (130, 231, 233, 235) may be implemented as a single chip or may be implemented as multiple chips. For example, the memories (130, 231, 233, 235) Each of these may be implemented as a single integrated circuit or as multiple integrated circuits. For example, each of the memories (130, 231, 233, 235) may be distributed within the electronic device.

[0051] In one embodiment, each of the communication circuits (290, 291, 293, 295) may be used for a communication connection between devices (e.g., electronic devices (101, 211, 213, 215) and a server (108)). For example, each of the communication circuits (290, 291, 293, 295) may include at least a portion of the communication module (190) (or wireless communication module (192)) of FIG. 1, or may correspond to at least a portion of the communication module (190) (or wireless communication module (192)) of FIG. 1. For example, each of the communication circuits (290, 291, 293, 295) may include a communication circuit for a long-distance communication network. For example, each of the communication circuits (290, 291, 293, 295) may be used to establish a communication link. For example, each of the communication circuits (290, 291, 293, 295) may be implemented as a single chip or as multiple chips. For example, each of the communication circuits (290, 291, 293, 295) may be implemented as a single integrated circuit or as multiple integrated circuits. For example, each of the communication circuits (290, 291, 293, 295) may be arranged in a distributed manner within an electronic device.

[0052] In one embodiment, the server (108) may be a server for learning (or federated learning (FL)) of an AI model. In one embodiment, the AI ​​model may be utilized to learn data to identify patterns and use them to make predictions or classify new data. When the AI ​​model is trained through data, supervised learning, unsupervised learning, clustering algorithms, or a combination thereof may be utilized. In one embodiment, the AI ​​model may be implemented based on an artificial neural network (mimicking the structure of the human brain). In one embodiment, the AI ​​model may be composed of a deep neural network composed of multiple layers and may be trained using a deep learning method. In one embodiment, the AI ​​model may include a plurality of parameters related to a neural network having a structure based on an encoder and a decoder, such as a transformer. In one embodiment, the AI ​​model may include a bidirectional model based on learning for an encoder (e.g., bidirectional encoder representations from transformers (BERT)) or an auto-encoding model (e.g., a diffusion model). In one embodiment, the AI ​​model may include an auto-regressor model based on learning for a decoder (e.g., a generative pre-trained transformer (GPT)). In one embodiment, the AI ​​model may include a sequence-to-sequence model based on learning for an encoder and a decoder (e.g., stable diffusion, DALL-E 2). In one embodiment, the AI ​​model may include, but is not limited to, a large language model (LLM) for processing natural language based on massive parameters.In one embodiment, the AI ​​model may include parameters for driving a neural network, such as a convolutional neural network (CNN), a recurrent neural network (RNN), a feedforward neural network (FNN), and / or a long short-term memory (LSTM).

[0053] In one embodiment, the server (108) may select a plurality of electronic devices (101, 201, 205) to perform AI model training in a round for AI model training. In one embodiment, a round for AI model training may refer to one cycle in which a plurality of clients (e.g., electronic devices (101, 201, 205)) individually train an AI model, and the server (108) collects (or synthesizes) the training results to update the AI ​​model. Here, a group (or set) of a plurality of clients may be referred to as a cluster of a round for AI model training.

[0054] Here, multiple clients (e.g., electronic devices (101, 201, 205)) individually train an AI model, which may include updating one or more parameters of the AI ​​model using each client's local data (or its own data). In one embodiment, the server (108) may repeat the rounds for training the AI ​​model multiple times, but is not limited thereto.

[0055] In one embodiment, the server (108) may select a plurality of electronic devices (101, 201, 205) to perform AI model learning based on information (e.g., device performance information, device identification information, device usage information, and / or user information) of each of the plurality of electronic devices (101, 201, 205). In one embodiment, the device performance information may include information about the computational capabilities (e.g., computational capabilities of a CPU, a GPU, and / or an NPU) of each of the plurality of electronic devices (101, 201, 205). In one embodiment, the device identification information may include information about the ID (identification) (or product identification information) and / or type (e.g., smartphone, tablet, PC, TV, refrigerator, robot vacuum cleaner, AI speaker, or air conditioner) of each of the plurality of electronic devices (101, 201, 205). In one embodiment, the device usage information may include information about the network connection status (e.g., whether connected via a local area network (e.g., WiFi), battery status (e.g., whether charged and / or remaining battery power), and / or whether idle (or whether performing a task) of each of the plurality of electronic devices (101, 201, 205). In one embodiment, the user information may include information about gender, age, nationality, race, religion, and / or region of residence.

[0056] In one embodiment, the server (108) may request training of an AI model to selected electronic devices (101, 201, 205). For example, the server (108) may transmit an AI model for training (or information on a path for downloading the AI ​​model) to the electronic devices (101, 201, 205). For example, the server (108) may transmit a condition of a round for training the AI ​​model to the electronic devices (101, 201, 205). In one embodiment, the condition of the round may include an expiration time of the round, a size of data required for training, a type of data representing a training result to be transmitted to the server (108) (e.g., parameters of an updated AI model, and / or evaluation results of an updated AI model), whether training of the AI ​​model can be delegated, and / or a condition for delegating training of the AI ​​model.

[0057] In one embodiment, the server (108) may obtain data representing learning results from electronic devices (101, 201, 205) (or electronic devices that have approved the delegation of learning of the AI ​​model) before the end of learning (or the end of transmission of data representing learning results to be transmitted to the server (108).

[0058] In one embodiment, the server (108) may collect data representing learning results to update (or refine) the AI ​​model. In one embodiment, the server (108) may update (or refine) the AI ​​model by averaging (or weighting averaging) parameters of the AI ​​model in the data representing learning results obtained from the electronic devices (101, 201, 205). In one embodiment, the server (108) may update (or refine) the AI ​​model based on data (or metrics) related to the evaluation of the AI ​​model in terms of accuracy and / or loss in the data representing learning results. However, the present invention is not limited thereto.

[0059] In one embodiment, the server (108) may provide services to electronic devices (101, 201, 205, 211, 213, 215) using an updated (or improved) AI model.

[0060] In one embodiment, each of the electronic devices (101, 201, 205) may train an AI model based on a training request for an AI model from the server (108). For example, each of the electronic devices (101, 201, 205) may train an AI model based on data stored in each of the electronic devices (101, 201, 205) (or data collected by each of the electronic devices (101, 201, 205)) (or data generated by each of the electronic devices (101, 201, 205)) (hereinafter, referred to as local data). For example, each of the electronic devices (101, 201, 205) may update a plurality of parameters included in the AI ​​model based on the local data. For example, each of the electronic devices (101, 201, 205) may evaluate the AI ​​model based on the local data. For example, each of the electronic devices (101, 201, 205) may generate data (or metrics) related to the accuracy and / or loss of the updated AI model based on local data, but is not limited thereto.

[0061] In one embodiment, each of the electronic devices (101, 201, 205) may transmit data representing learning results to the server (108) before the end of learning (or the end of transmission of data representing learning results to be transmitted to the server (108).

[0062] In order to learn and / or evaluate the AI ​​model in each of the electronic devices (101, 201, 205), the resources of each of the electronic devices (101, 201, 205) may be consumed. Furthermore, in each of the electronic devices (101, 201, 205), the learning and / or evaluation of the AI ​​model may be delayed due to competition with other tasks. Accordingly, the learning and / or evaluation of the AI ​​model may not be completed within the deadline specified by the server (108) due to the limited resources of each of the electronic devices (101, 201, 205).

[0063] In addition, even if learning and / or evaluation of the AI ​​model is completed, due to a lack of resources (e.g., remaining battery power, network disconnection) for transmitting data representing learning results to the server (108) in each of the electronic devices (101, 201, 205), the data representing learning results may not be transmitted to the server (108) within the deadline specified by the server (108).

[0064] Therefore, a method may be required to ensure that learning and / or evaluation of an AI model is completed in a timely manner and that data indicating learning results are transmitted to a server (108) in a timely manner.

[0065] Hereinafter, a delegated operation performed by the electronic device (101) may be described when the electronic device (101) fails to complete learning and / or evaluation of the AI ​​model and / or fails to transmit data representing the learning result to the server (108).

[0066] Delegation of AI model learning

[0067] In one embodiment, the electronic device (101) can determine whether delegation of AI model learning is necessary. In one embodiment, the electronic device (101) can determine whether delegation of AI model learning is necessary based on device performance information and / or device usage information of the electronic device (101). For example, the electronic device (101) can determine whether AI model learning and / or evaluation can be terminated before the end of a round.

[0068] In one embodiment, the electronic device (101) can continuously determine whether delegation of AI model learning is necessary. For example, the electronic device (101) can determine whether delegation of AI model learning is necessary before initiating AI model learning after receiving an AI model learning request from the server (108). For example, the electronic device (101) can determine whether delegation of AI model learning is necessary after initiating AI model learning but before completing AI model learning. For example, the electronic device (101) can determine whether delegation of AI model learning is necessary before transmitting data representing learning results to the server (108) after completing AI model learning.

[0069] In one embodiment, the electronic device (101) may decide to delegate the learning of the AI ​​model if it is expected that the learning and / or evaluation of the AI ​​model will not be completed before the expiration of the round.

[0070] In one embodiment, the electronic device (101) can delegate learning of an AI model among electronic devices (211, 213, 215) of a trust relationship (210). For example, an electronic device having the same user account (e.g., Samsung account) as the user account of the electronic device (101) can be evaluated as having a trust relationship (210) with the electronic device (101). For example, an electronic device that is communicatively connected to the electronic device (101) can be evaluated as having a trust relationship (210) with the electronic device (101). For example, an electronic device that is communicatively connected to the electronic device (101) through a designated short-range network (e.g., Bluetooth, WiFi, P2P (peer to peer)) can be evaluated as having a trust relationship (210) with the electronic device (101). For example, an electronic device that has authorized a communication connection with an electronic device (101) to provide a designated function (e.g., screen mirroring) may be evaluated as having a trust relationship (210) with the electronic device (101). For example, an electronic device that is communicatively connected with a designated peripheral device (e.g., a mouse, a keyboard, a printer) that is communicatively connected with the electronic device (101) may be evaluated as having a trust relationship (210) with the electronic device (101). For example, an electronic device that is separately designated for model training may be evaluated as having a trust relationship (210) with the electronic device (101). Here, the electronic device being separately designated for model training may include being designated by the electronic device (101) and / or the server (108). For example, the electronic device being separately designated for model training may include being designated by the electronic device (101) and / or the server (108) for delegation of learning of an AI model. For example, an electronic device identified by an electronic device (101) (or a user of the electronic device (101)) for model training may be evaluated as having a trust relationship (210) with the electronic device (101).Here, the electronic device may be selected from among the scanned electronic devices after the electronic device (101) receives a request for training an AI model from the server (108) (or before initiating training of the AI ​​model).

[0071] In one embodiment, the electronic device (101) may delegate the learning of an AI model to an electronic device selected from among the electronic devices (211, 213, 215) of the trust relationship (210). For example, the electronic device (101) may select an electronic device to which to delegate the learning of the AI ​​model based on information (e.g., device performance information, device identification information, and / or device usage information) of each of the electronic devices (211, 213, 215). In one embodiment, the device performance information may include information about the computing capability (e.g., the computing capability of a CPU, GPU, and / or NPU) of each of the electronic devices (211, 213, 215). In one embodiment, the device identification information may include information about the ID (or product identification information) and / or type (e.g., smartphone, tablet, PC, TV, refrigerator, robot vacuum cleaner, AI speaker, or air conditioner) of each of the electronic devices (211, 213, 215). In one embodiment, the device usage information may include information about the network connection status of each of the electronic devices (211, 213, 215) (e.g., whether connected via a local area network (e.g., WiFi), battery status (e.g., whether charging, and / or remaining battery capacity), and / or whether idle (or whether performing a task).

[0072] In one embodiment, the electronic device (101) may select an electronic device among the electronic devices (211, 213, 215) that has performance (e.g., computing power of a CPU, GPU, or NPU) higher than the performance for learning an AI model, as an electronic device to which learning of the AI ​​model will be delegated.

[0073] In one embodiment, the electronic device (101) may select an electronic device that is used less by the user (or that has no CPU, GPU, or NPU occupied by other tasks) (or that has a high expected idle time) and / or that has a high battery capacity (or remaining amount) (or charging frequency) as an electronic device to which AI model learning is delegated, based on the usage history (or usage pattern) (or frequency) (e.g., CPU usage, display active time) and battery capacity (or charging frequency) of each of the electronic devices (211, 213, and 215). In one embodiment, the electronic device (101) may select an electronic device that is being charged as an electronic device to which AI model learning is delegated, based on whether the batteries of each of the electronic devices (211, 213, and 215) are charged. In one embodiment, the electronic device (101) may identify the average idle time (or the most recent idle time) based on the usage history (or usage pattern) of each of the electronic devices (211, 213, 215) as the expected idle time.

[0074] In one embodiment, the electronic device (101) may select an electronic device with a network connection status (e.g., a connection status via WiFi) of each of the electronic devices (211, 213, 215) as the electronic device to which the AI ​​model learning will be delegated.

[0075] In one embodiment, the electronic device (101) may select an electronic device to which AI model learning is delegated based on the location of each of the electronic devices (211, 213, 215). For example, the electronic device (101) may select an electronic device among the electronic devices (211, 213, 215) that is connected to the same network (or the same access point) as the electronic device (101) as the electronic device to which AI model learning is delegated.

[0076] In one embodiment, the electronic device (101) may identify an electronic device capable of learning and / or evaluating an AI model before the expiration time of a round among the electronic devices (211, 213, 215). In one embodiment, the electronic device (101) may calculate (or identify) a time required for learning and / or evaluating the AI ​​model of each of the electronic devices (211, 213, 215) based on information about each of the electronic devices (211, 213, 215). In one embodiment, the electronic device (101) may identify an electronic device capable of learning and / or evaluating an AI model before the expiration time based on the required time of each of the electronic devices (211, 213, 215).

[0077] In one embodiment, the electronic device (101) may request (or delegate) AI model training to an electronic device (e.g., electronic device (211)) selected from among the electronic devices (211, 213, 215) (hereinafter, referred to as a delegated device). For example, the electronic device (101) may transmit an AI model for training and local data of the electronic device (101) to the delegated device. For example, the electronic device (101) may transmit a delegation condition to the delegated device. In one embodiment, the delegation condition may include information about the expiration time of a round, the size of data required for training, and / or the type of data representing the learning result to be transmitted to the server (108) (e.g., parameters of an updated AI model, and / or evaluation results of an updated AI model). In one embodiment, the delegation condition may include information about the destination of the data representing the learning result to be transmitted (e.g., electronic device (101) or server (108)). In one embodiment, the delegation conditions may include information about whether learning based on local data of the delegated device is necessary.

[0078] In one embodiment, the commissioning device may transmit data representing learning results to the server (108) (via (or via) the electronic device (101)) before the expiration of the round.

[0079] According to an embodiment, the electronic device (101) may continuously determine whether the AI ​​model's learning can be completed before the end of the round during AI model learning. According to an embodiment, based on the determination that the AI ​​model's learning cannot be completed during AI model learning, the training of the partially learned AI model may be requested (or delegated) to a selected delegate device among the electronic devices (211, 213, 215).

[0080] As described above, the electronic device (101) can transmit data indicating learning results in a timely manner (before the end of the round). In addition, the server (108) can timely obtain the learning results of the AI ​​model performed by the electronic devices (101, 201, 205) included in the cluster of rounds for learning. Accordingly, since the federated learning is not delayed, the need for additional rounds requiring computing resources of the server (108) and the electronic devices (101, 201, 205) can be reduced. Here, a group (or set) of multiple clients may be referred to as a cluster of rounds for learning the AI ​​model.

[0081] Delegation of transmission of data representing learning results

[0082] In one embodiment, the electronic device (101) may determine whether delegation of data transmission is required. In one embodiment, the electronic device (101) may determine whether delegation of transmission of data representing learning results is required based on device performance information and / or device usage information of the electronic device (101). For example, the electronic device (101) may determine whether data representing learning results can be transmitted to the server (108) before the end of the round.

[0083] In one embodiment, the electronic device (101) can continuously determine whether delegation of data transmission is necessary. For example, the electronic device (101) can determine whether delegation of data transmission is necessary before initiating AI model training after receiving an AI model training request from the server (108). For example, the electronic device (101) can determine whether delegation of data transmission is necessary after initiating AI model training but before completing AI model training. For example, the electronic device (101) can determine whether delegation of data transmission is necessary before transmitting data representing the training results to the server (108) after completing AI model training.

[0084] In one embodiment, the electronic device (101) may decide to delegate transmission of data representing learning results if it is expected that transmission of data representing learning results of the AI ​​model will not be completed before the expiration of the round.

[0085] In one embodiment, the electronic device (101) may delegate transmission of data representing learning results among the electronic devices (211, 213, 215) of the trust relationship (210).

[0086] In one embodiment, the electronic device (101) may delegate transmission of data indicating learning results to an electronic device selected from among the electronic devices (211, 213, 215) in the trust relationship (210). For example, the electronic device (101) may select an electronic device to delegate transmission of data indicating learning results to based on information (e.g., device performance information, device identification information, and / or device usage information) of each of the electronic devices (211, 213, 215). In one embodiment, the electronic device (101) may display a user interface (UI) representing a list (e.g., 910 of FIG. 9) of the electronic devices (211, 213, 215) in the trust relationship (210) through a display (e.g., 260 of FIG. 9). In one embodiment, the electronic device (101) may display a UI (e.g., 920 of FIG. 9) through the display (260) that queries whether to delegate transmission of data to an electronic device selected (by a user) from among electronic devices (211, 213, 215) of a trust relationship (210). In one embodiment, the electronic device (101) may delegate transmission of data to the selected electronic device based on an input requesting transmission of data to the selected electronic device.

[0087] In one embodiment, the electronic device (101) can identify an electronic device capable of transmitting data indicating a learning result before the expiration time of a round among the electronic devices (211, 213, 215). In one embodiment, the electronic device (101) can calculate (or identify) a time required for transmitting data indicating a learning result of each of the electronic devices (211, 213, 215) based on information about each of the electronic devices (211, 213, 215). In one embodiment, the electronic device (101) can identify an electronic device capable of transmitting data indicating a learning result before the expiration time based on the required time of each of the electronic devices (211, 213, 215).

[0088] In one embodiment, the electronic device (101) may request (or delegate) transmission of data representing learning results to an electronic device (e.g., electronic device (211)) selected from among the electronic devices (211, 213, 215) (hereinafter, referred to as a delegated device). For example, the electronic device (101) may transmit data representing learning results to the delegated device. For example, the electronic device (101) may transmit delegation conditions to the delegated device. In one embodiment, the delegation conditions may include information about the expiration time of a round and / or the type of data representing learning results to be transmitted to the server (108) (e.g., parameters of an updated AI model and / or evaluation results of an updated AI model). In one embodiment, the delegation conditions may include information about the destination of the data representing learning results to be transmitted (e.g., server (108)). In one embodiment, the delegation conditions may include information about whether additional learning is necessary based on local data of the delegated device.

[0089] In one embodiment, the commissioning device may transmit data representing learning results to the server (108) before the expiration of the round.

[0090] As described above, the electronic device (101) can transmit data indicating learning results in a timely manner (before the end of the round). Furthermore, the server (108) can obtain the learning results of the AI ​​model performed by the electronic devices (101, 201, 205) included in the cluster for the round of learning in a timely manner. Accordingly, since federated learning is not delayed, the need for additional rounds requiring computing resources of the server (108) and the electronic devices (101, 201, 205) can be reduced. Furthermore, since the learning of the AI ​​model is not delayed, the AI ​​model after learning can be quickly distributed.

[0091] Consent of delegation

[0092] In one embodiment, the electronic device (101) may, before delegating the learning of the AI ​​model to the electronic devices (211, 213, 215) of the trust relationship (210), check (or inquire) with the server (108) whether delegation is possible. In one embodiment, if the electronic device (101) is expected to not complete the learning and / or evaluation of the AI ​​model before the expiration of the round, the electronic device (101) may inquire with the server (108) whether delegation is possible.

[0093] In one embodiment, the electronic device (101) may inquire whether delegation of AI model learning is possible among the electronic devices (211, 213, 215) of the trust relationship (210). In one embodiment, the electronic device (101) may inquire whether delegation of AI model learning is possible for an electronic device selected for AI model learning among the electronic devices (211, 213, 215).

[0094] In one embodiment, the electronic device (101) may delegate the learning of the AI ​​model to an electronic device selected for learning the AI ​​model based on the server (108) confirming (or agreeing) that the delegation of learning of the AI ​​model is possible among the electronic devices (211, 213, 215) of the trust relationship (210).

[0095] According to an embodiment, the server (108) may select an electronic device for training the AI ​​model among the electronic devices (211, 213, 215) in a trust relationship (210) with the electronic device (101). In one embodiment, the server (108) may transmit a signal requesting training of the AI ​​model to the electronic device (101) (or the selected electronic device) through the electronic device selected among the electronic devices (211, 213, 215).

[0096] As described above, the electronic device (101) can inquire with the server (108) whether delegation of AI model learning is possible. Accordingly, the server (108) can determine if AI model learning is being performed by an electronic device not included in the cluster. Furthermore, if delegation of AI model learning is inappropriate, the server (108) can maintain the benefits of clustering by not agreeing to delegation of AI model learning.

[0097] Question of delegation

[0098] In one embodiment, the electronic device (101) may output a user interface (UI) that queries the user of the electronic device (101) to delegate learning of the AI ​​model before requesting learning of the AI ​​model to the selected electronic device.

[0099] In one embodiment, the electronic device (101) may output a UI including a list of multiple electronic devices to which AI model learning can be delegated among the electronic devices (211, 213, 215) of the trust relationship (210). In one embodiment, after outputting the UI, the electronic device (101) may, in response to a user input for selecting one electronic device, request the selected electronic device to learn the AI ​​model. However, the present invention is not limited thereto. In one embodiment, the electronic device (101) may output a UI for querying AI model learning through an electronic device selected by the electronic device (101) among the electronic devices (211, 213, 215) of the trust relationship (210). In one embodiment, after outputting the UI, the electronic device (101) may, in response to a user input for consent (or confirmation), request the selected electronic device to learn the AI ​​model.

[0100] Question of appointment

[0101] In one embodiment, an electronic device (e.g., electronic device (211)) selected for AI model learning may output a user interface (UI) that queries the user of the selected electronic device (e.g., electronic device (211)) about the number of AI model learning requests received from the electronic device (101).

[0102] In one embodiment, the selected electronic device (e.g., electronic device (211)) may, in response to a request for learning an AI model from an electronic device (101) of a trust relationship (210), transmit a signal to the electronic device (101) approving learning of the AI ​​model in response to a user input for consent (or confirmation) after outputting a UI.

[0103] According to an embodiment, an electronic device (e.g., electronic device (211)) selected for AI model learning may determine whether AI model learning is possible based on usage history (or usage pattern) (or frequency) (e.g., CPU usage, display activation time) and battery capacity (or charging frequency). The electronic device (e.g., electronic device (211)) selected for AI model learning may output a UI that queries the user of the selected electronic device (e.g., electronic device (211)) about the possibility of AI model learning based on the determination that AI model learning is possible.

[0104] Queries on local data

[0105] In one embodiment, the electronic device (101) may check (or inquire) to the server (108) whether AI model learning is possible based on local data of electronic devices (211, 213, 215) of the trust relationship (210). In one embodiment, when the size of the local data of the electronic device (101) is smaller than the size of the data for learning the AI ​​model, the electronic device (101) may check (or inquire) to the server (108) whether AI model learning is possible based on local data of electronic devices (211, 213, 215) of the trust relationship (210). However, the present invention is not limited thereto. For example, before delegating the learning of the AI ​​model to the electronic devices (211, 213, 215) of the trust relationship (210), the electronic device (101) may check (or inquire) to the server (108) whether learning of the AI ​​model is possible based on the local data of the entrusted device (e.g., the electronic device (211)). Alternatively, the entrusted device (e.g., the electronic device (211)) may check (or inquire) to the server (108) whether learning (or additional learning) of the AI ​​model is possible based on the local data of the entrusted device (e.g., the electronic device (211)) before learning (or transmitting data representing the learning result) of the AI ​​model from the electronic device (101).

[0106] In one embodiment, the server (108) may transmit a signal to the electronic device (101) indicating that AI model learning is possible based on local data of the electronic devices (211, 213, 215) of the trust relationship (210) with the electronic device (101), based on the determination that AI model learning is possible based on local data of the electronic devices (211, 213, 215) of the trust relationship (210). According to an embodiment, the server (108) may transmit a signal indicating that AI model learning is possible based on local data of a specific electronic device (e.g., a delegated electronic device (e.g., electronic device (211))) among the electronic devices (211, 213, 215) of the trust relationship (210). According to an embodiment, the server (108) may transmit a signal to the assigned device (e.g., the electronic device (211)) indicating that learning of the AI ​​model is possible based on local data of the assigned device (e.g., the electronic device (211)).

[0107] Changes to the cluster

[0108] In one embodiment, the server (108) can identify whether electronic devices (101, 201, 205) included in a cluster for federated learning of an AI model are replaced. For example, the server (108) can query the electronic device (101) for replacement (or exclusion) from the cluster in a subsequent round based on receiving a signal from the electronic device (101) for confirming (or querying) whether delegation is possible. In one embodiment, replacement in the cluster can include an electronic device included in the cluster being changed from an electronic device (101) to an entrusting device (e.g., electronic device (211)). In one embodiment, exclusion from the cluster can include an electronic device (101) being excluded from the electronic devices (101, 201, 205) included in the cluster.

[0109] In one embodiment, the server (108) may receive a response from the electronic device (101) to a query regarding replacement (or exclusion) from the cluster in a subsequent round. In one embodiment, the server (108) may replace (or exclude) the electronic device (101) from the cluster for training the AI ​​model based on the response from the electronic device (101).

[0110] In an embodiment, the server (108) may query the assigned device (e.g., electronic device (211)) for inclusion in the cluster. In one embodiment, the server (108) may add the assigned device (e.g., electronic device (211)) to the cluster in a subsequent round based on a response from the assigned device (e.g., electronic device (211)).

[0111] In an embodiment, the electronic device (101) may request the server (108) to include the assigned device (e.g., the electronic device (211)) in the cluster. In one embodiment, the server (108) may determine whether to include the assigned device (e.g., the electronic device (211)) in the cluster based on the request.

[0112] FIG. 3 illustrates an example of programs for learning an AI model according to one embodiment.

[0113] Programs (310, 320, 330, 340, 350, 360, 370, 380) for learning the AI ​​model of FIG. 3 may be stored in the memory of each of the electronic devices (101, 201, 205, 211, 213, 215). In one embodiment, the programs (310, 320, 330, 340, 350, 360, 370, 380) for learning the AI ​​model may be executed by the processor of each of the electronic devices (101, 201, 205, 211, 213, 215).

[0114] In one embodiment, a federated learning (FL) manager (310) can control and / or manage a task (or task job) for AI learning (or the entire workflow for a task from start to finish). In one embodiment, the FL manager (310) can check the status of an electronic device (e.g., electronic device (101) and electronic devices (211, 213, 215)). In one embodiment, the FL manager (310) can identify (or select) an electronic device to which AI model learning will be delegated among the electronic devices (211, 213, 215).

[0115] In one embodiment, the data collector (320) can collect local data. In one embodiment, the data collector (320) can collect local data through a sensor module (e.g., 176 of FIG. 1). In one embodiment, the data collector (320) can collect local data through a program (e.g., an application (146), middleware (144), and / or an operating system (142) of FIG. 1). However, the present invention is not limited thereto. In one embodiment, the data collector (320) can transmit the collected local data to the FL manager (310).

[0116] In one embodiment, the device analyzer (330) may collect information of an electronic device related to learning of an AI model (e.g., information about the performance (or metric) of the electronic device and / or information about the learning participation rate). In one embodiment, the device analyzer (330) may transmit the collected information to the FL manager (310).

[0117] In one embodiment, the data transceiver (340) may obtain an AI model from the server (108). In one embodiment, the data transceiver (340) may transmit data representing the learning results of the AI ​​model to the server (108). In one embodiment, the data transceiver (340) may transmit the AI ​​model to the assigned device. In one embodiment, the data transceiver (340) may transmit data (e.g., local data) for learning the AI ​​model to the assigned device. In one embodiment, the data transceiver (340) may receive data representing the learning results of the AI ​​model from the assigned device.

[0118] In one embodiment, the model learner (350) can train an AI model (e.g., an AI model obtained from a server (108)) based on local data (e.g., data obtained through a data collector (320). In one embodiment, the model learner (350) can evaluate the trained AI model.

[0119] In one embodiment, the device status manager (360) may obtain information indicating the status of the electronic device (e.g., whether AI model learning is in progress, whether it is idle, whether it is connected to a local area network, and / or whether the battery is charged). In one embodiment, the device status manager (360) may transmit information indicating the status of the electronic device to other programs (310, 320, 330, 340, 350, 370, 380) (e.g., FL manager (310)).

[0120] In one embodiment, the security module (370) can identify (or maintain) (or manage) a trust relationship (210) between the electronic device (101) and electronic devices (211, 213, 215). In one embodiment, the FL manager (310) can identify an electronic device to which AI model learning will be delegated among the electronic devices (211, 213, 215) having a trust relationship (210) identified through the security module (370).

[0121] In one embodiment, the inference module (380) may perform a function related to prediction through the learned AI model after learning the AI ​​model.

[0122] Figure 4 is a flowchart of operations for learning an AI model according to one embodiment.

[0123] FIG. 4 can be explained with reference to FIG. 1, FIG. 2a, FIG. 2b, and FIG. 3.

[0124] In FIG. 4, the electronic device (401) may be an electronic device that has formed a trust relationship with the electronic device (101). For example, the electronic device (401) may be one of the electronic devices (211, 213, or 215) of FIGS. 2A and 2B.

[0125] In the following examples, the operations may be performed sequentially, but are not necessarily sequential. For example, the order of the operations may be changed, and at least two operations may be performed in parallel.

[0126] Referring to FIG. 4, in operation 410, the server (108) may request training of an AI model from the electronic device (101). In one embodiment, the server (108) may request training of an AI model from an electronic device (101) selected as an electronic device to perform training of an AI model in a round for training an AI model. For example, the server (108) may transmit an AI model for training (or information on a path for downloading the AI ​​model) to the electronic device (101). For example, the server (108) may transmit conditions of a round for training an AI model to the electronic device (101). In one embodiment, the conditions of the round may include an expiration time of the round, a size of data required for training, a type of data representing a training result to be transmitted to the server (108) (e.g., parameters of an updated AI model, and / or evaluation results of an updated AI model), whether training of an AI model can be delegated, and / or conditions for delegating training of an AI model.

[0127] In operation 420, the electronic device (101) can determine whether self-learning of the AI ​​model is possible. In one embodiment, the electronic device (101) can determine whether to delegate the AI ​​model.

[0128] In one embodiment, the electronic device (101) may determine whether delegation of AI model learning is necessary based on device performance information and / or device usage information of the electronic device (101). For example, the electronic device (101) may determine whether AI model learning and / or evaluation can be terminated before the end of a round.

[0129] In one embodiment, the electronic device (101) may determine whether self-learning of the AI ​​model is possible (or whether delegation of AI model learning is required) based on the time required for AI model learning and / or evaluation. In one embodiment, the electronic device (101) may determine that self-learning of the AI ​​model is not possible (or delegation of AI model learning is required) if it is expected that the AI ​​model learning and / or evaluation will not be completed before the end of the round.

[0130] In one embodiment, in operation 420, based on the determination that self-learning is possible (or, delegation of learning of an AI model is not necessary), the electronic device (101) may perform operation 430. In one embodiment, in operation 420, based on the determination that self-learning is not possible (or, delegation of learning of an AI model is necessary), the electronic device (101) may perform operation 450.

[0131] In operation 430, the electronic device (101) may train an AI model. For example, the electronic device (101) may train the AI ​​model based on local data stored in the electronic device (101). For example, the electronic device (101) may update a plurality of parameters included in the AI ​​model based on the local data.

[0132] In operation 435, the electronic device (101) may evaluate the AI ​​model. For example, the electronic device (101) may evaluate the updated AI model based on local data. For example, the electronic device (101) may generate data (or metrics) related to the accuracy and / or loss of the updated AI model based on the local data. However, the present invention is not limited thereto.

[0133] In operation 440, the electronic device (101) may transmit the learning result. In one embodiment, the electronic device (101) may transmit data representing the learning result to the server (108) before the learning expiration time (or the expiration time of transmission of data representing the learning result to be transmitted to the server (108). In one embodiment, the data representing the learning result may include updated parameters of the AI ​​model. In one embodiment, the data representing the learning result may include data related to the evaluation of the AI ​​model (e.g., data related to the accuracy and / or loss of the AI ​​model). However, the present invention is not limited thereto.

[0134] Depending on the embodiment, operation 435 may be omitted. For example, the electronic device (101) may not perform evaluation of the AI ​​model. In this case, the data transmitted by the electronic device (101) to the server (108) may include updated parameters of the AI ​​model.

[0135] In operation 450, the electronic device (101) may identify a learnable electronic device. In one embodiment, the electronic device (101) may identify a learnable electronic device among candidate electronic devices (211, 213, 215) having a trust relationship (210). For example, an electronic device having the same user account as the user account of the electronic device (101) (e.g., a Samsung account) may be evaluated as having a trust relationship (210) with the electronic device (101). For example, an electronic device that is communicatively connected to the electronic device (101) may be evaluated as having a trust relationship (210) with the electronic device (101). For example, an electronic device that is communicatively connected to the electronic device (101) through a designated short-range network (e.g., Bluetooth, WiFi, P2P (peer to peer)) may be evaluated as having a trust relationship (210) with the electronic device (101). For example, an electronic device that has authorized a communication connection with an electronic device (101) to provide a specified function (e.g., screen mirroring) may be evaluated as having a trust relationship (210) with the electronic device (101). For example, an electronic device that is communicatively connected to a specified peripheral device (e.g., mouse, keyboard, printer) that is communicatively connected to the electronic device (101) may be evaluated as having a trust relationship (210) with the electronic device (101).

[0136] In one embodiment, the electronic device (101) can identify an electronic device capable of learning from among the electronic devices (211, 213, 215) of the trust relationship (210). In one embodiment, the electronic device (101) can select an electronic device capable of learning an AI model based on information (e.g., device performance information, device identification information, and / or device usage information) of each of the electronic devices (211, 213, 215). In one embodiment, the device performance information may include information about the computational capabilities (e.g., the computational capabilities of a CPU, a GPU, and / or an NPU) of each of the electronic devices (211, 213, 215). In one embodiment, the device identification information may include information about the ID (or product identification information) and / or type (e.g., a smartphone, a tablet, a PC, a TV, a refrigerator, a robot vacuum cleaner, an AI speaker, or an air conditioner) of each of the electronic devices (211, 213, 215). In one embodiment, the device usage information may include information about the network connection status of each of the electronic devices (211, 213, 215) (e.g., whether connected via a local area network (e.g., WiFi), battery status (e.g., whether charging, and / or remaining battery capacity), and / or whether idle (or whether performing a task).

[0137] In one embodiment, the electronic device (101) may identify an electronic device capable of learning and / or evaluating an AI model before the expiration time of a round among the electronic devices (211, 213, 215). In one embodiment, the electronic device (101) may calculate (or identify) a time required for learning and / or evaluating the AI ​​model of each of the electronic devices (211, 213, 215) based on information about each of the electronic devices (211, 213, 215). In one embodiment, the electronic device (101) may identify an electronic device capable of learning and / or evaluating an AI model before the expiration time based on the required time of each of the electronic devices (211, 213, 215).

[0138] In operation 460, the electronic device (101) may request learning. In one embodiment, the electronic device (101) may request (or delegate) learning of an AI model to an electronic device (401) capable of learning an AI model among candidate electronic devices (211, 213, 215). For example, the electronic device (101) may transmit an AI model for learning and local data of the electronic device (101) to the electronic device (401). For example, the electronic device (101) may transmit local data of the electronic device (101) to the electronic device (401) as data for learning. For example, the electronic device (101) may transmit delegation conditions to the electronic device (401). In one embodiment, the delegation condition may include information about the expiration time of the round, the size of data required for learning, and / or the type of data representing the learning result to be transmitted to the server (108) (e.g., parameters of the updated AI model, and / or evaluation results of the updated AI model). In one embodiment, the delegation condition may include information about the destination of the data representing the learning result to be transmitted (e.g., the electronic device (101) or the server (108)). In one embodiment, the delegation condition may include information about whether learning based on local data of the electronic device (401) is necessary.

[0139] In operation 470, the electronic device (401) may train an AI model. For example, the electronic device (401) may train the AI ​​model based on local data of the electronic device (101) acquired from the electronic device (101). For example, the electronic device (401) may update a plurality of parameters included in the AI ​​model based on the local data of the electronic device (101). However, the present invention is not limited thereto. For example, the electronic device (401) may train the AI ​​model based on local data of the electronic device (401) stored in the electronic device (401). In one embodiment, the size of the local data of the electronic device (401) and the local data of the electronic device (101) used for the AI ​​model may not exceed the size of the data required for training according to the conditions of the round. In one embodiment, the electronic device (401) can train the AI ​​model based on the local data of the electronic device (401) by the amount of the difference between the size of data required for learning according to the conditions of the round (or the size of the local data of the electronic device (101) minus the size of data required for learning according to the conditions of the round).

[0140] In operation 475, the electronic device (401) may evaluate the AI ​​model. For example, the electronic device (401) may evaluate an updated AI model. For example, the electronic device (401) may generate data (or metrics) related to the accuracy and / or loss of the updated AI model. However, the present invention is not limited thereto.

[0141] Depending on the embodiment, operation 475 may be omitted. For example, the electronic device (401) may not perform evaluation of the AI ​​model. In this case, the data transmitted by the electronic device (401) to the server (108) may include updated parameters of the AI ​​model.

[0142] In operation 480, the electronic device (401) may transmit the learning result to the electronic device (101). The learning result may be transmitted. In one embodiment, the electronic device (401) may transmit data representing the learning result to the electronic device (101) based on the delegation condition. In one embodiment, the data representing the learning result may include updated parameters of the AI ​​model. In one embodiment, the data representing the learning result may include data related to the evaluation of the AI ​​model (e.g., data related to the accuracy and / or loss of the AI ​​model). However, the present invention is not limited thereto.

[0143] In operation 485, the electronic device (101) may transmit data indicating learning results to the server (108) before the expiration of the round.

[0144] In operation 490, the electronic device (401) may transmit learning results to the server (108). In one embodiment, the electronic device (401) may transmit data representing learning results to the server (108) based on the delegation conditions.

[0145] According to an embodiment, the electronic device (401) may perform one of operations 480 and 490. For example, the electronic device (401) may transmit data representing learning results to the electronic device (101) or the server (108) based on the delegation conditions.

[0146] FIG. 5 is a flowchart of operations for delegating learning of an AI model performed by an electronic device according to an embodiment.

[0147] FIG. 5 may be described with reference to FIGS. 1, 2A, 2B, 3, and 4. Operation 450 of FIG. 5 may be included in operation 450 of FIG. 4. Operations 510, 520, and 530 of FIG. 5 may be included in operation 460 of FIG. 4.

[0148] In FIG. 5, the electronic device (401) may be an electronic device that has formed a trust relationship with the electronic device (101). For example, the electronic device (401) may be one of the electronic devices (211, 213, or 215) of FIGS. 2A and 2B.

[0149] In the following examples, the operations may be performed sequentially, but are not necessarily sequential. For example, the order of the operations may be changed, and at least two operations may be performed in parallel.

[0150] In operation 450, the electronic device (101) can identify a learnable electronic device (401). For example, the electronic device (101) can identify an electronic device (401) capable of learning and / or evaluating an AI model before the end of a round among the electronic devices (211, 213, 215).

[0151] In operation 510, the electronic device (101) may determine whether learning based on local data of the receiving device (e.g., electronic device (401)) is required. For example, if the size of the local data of the electronic device (101) is less than the size of the data required for learning, it may be determined that learning based on local data of the receiving device (e.g., electronic device (401)) is required.

[0152] In one embodiment, the electronic device (101) may determine whether learning based on local data of the entrusted device (e.g., electronic device (401)) is required based on conditions when the server (108) approves learning delegation from the electronic device (101) to the electronic device (401). For example, the electronic device (101) may determine that learning based on local data of the entrusted device (e.g., electronic device (401)) is required based on the server (108) instructing (or requesting) learning based on local data of the entrusted device (e.g., electronic device (401)) in response to a query for learning delegation of the electronic device (101).

[0153] In one embodiment, based on the determination that learning based on local data of the assigned device (e.g., electronic device (401)) is not required in operation 510, the electronic device (101) may perform operation 520. In one embodiment, based on the determination that learning based on local data of the assigned device (e.g., electronic device (401)) is required in operation 510, the electronic device (101) may perform operation 530.

[0154] In operation 520, the electronic device (101) may request the electronic device (401) to learn an AI model based on local data of the delegated device (e.g., the electronic device (101)). In one embodiment, the electronic device (101) may prohibit the electronic device (401) from using local data of the delegated device (e.g., the electronic device (401)) when learning the AI ​​model.

[0155] In operation 530, the electronic device (101) may request training of an AI model based on local data of the receiving device. In one embodiment, the electronic device (101) may request training of an AI model from the electronic device (401) using local data of the receiving device (e.g., the electronic device (401)) when training the AI ​​model.

[0156] In one embodiment, training of an AI model based on local data of a delegated device does not exclude training based on local data of a delegated device (e.g., electronic device (101)). For example, when requesting training of an AI model based on local data of a delegated device, the electronic device (101) may also request training based on local data of a delegated device (e.g., electronic device (101)). For example, when training an AI model, the electronic device (101) may request training of an AI model from the electronic device (401) using local data of the delegated device (e.g., electronic device (101)) and local data of the delegated device (e.g., electronic device (401)).

[0157] FIG. 6 is a flowchart of operations for learning an AI model assigned to an electronic device according to one embodiment.

[0158] FIG. 6 may be described with reference to FIGS. 1, 2A, 2B, 3, and 4. Operations 620, 630, and 640 of FIG. 6 may be included in operation 470 of FIG. 4. Operation 650 of FIG. 6 may be included in operation 475 of FIG. 4. Operation 660 of FIG. 6 may correspond to operation 480 or operation 490 of FIG. 4.

[0159] In FIG. 6, the electronic device (401) may be an electronic device that has formed a trust relationship with the electronic device (101). For example, the electronic device (401) may be one of the electronic devices (211, 213, or 215) of FIGS. 2A and 2B.

[0160] In the following examples, the operations may be performed sequentially, but are not necessarily sequential. For example, the order of the operations may be changed, and at least two operations may be performed in parallel.

[0161] In operation 610, the electronic device (401) may receive a learning request from the electronic device (101) of the trust relationship (210). For example, the electronic device (401) may receive an AI model for learning and local data of the electronic device (101) from the electronic device (101). For example, the electronic device (401) may receive a delegation condition from the electronic device (101). In one embodiment, the delegation condition may include information about an expiration time of a round, a size of data required for learning, and / or a type of data representing a learning result to be transmitted to the server (108) (e.g., parameters of an updated AI model, and / or evaluation results of an updated AI model). In one embodiment, the delegation condition may include information about a destination of the data representing a learning result to be transmitted (e.g., the electronic device (101) or the server (108)). In one embodiment, the delegation conditions may include information about whether learning based on local data of the electronic device (401) is necessary.

[0162] In operation 620, the electronic device (401) may determine whether learning based on local data of the receiving device (e.g., the electronic device (401)) is required. For example, the electronic device (401) may determine that learning based on local data of the electronic device (401) is required when the size of the local data of the receiving device (e.g., the electronic device (401)) is less than the size of the data required for learning.

[0163] In one embodiment, the electronic device (401) may determine whether learning based on local data of the electronic device (401) is possible based on a delegation condition. For example, based on the delegation condition indicating that learning based on local data of the electronic device (401) is possible, the electronic device (401) may determine that learning based on local data of the electronic device (401) is necessary.

[0164] In one embodiment, the electronic device (401) may query the server (108) as to whether learning based on local data of the electronic device (401) is required (or possible). For example, based on the server (108) responding to the query by instructing (or requesting) learning based on local data of the electronic device (401), the electronic device (401) may determine that learning based on local data of the electronic device (401) is required (or possible).

[0165] In one embodiment, at operation 620, based on determining that learning based on local data of the electronic device (401) is necessary, the electronic device (401) may perform operation 630. In one embodiment, at operation 620, based on determining that learning based on local data of the electronic device (401) is not necessary, the electronic device (401) may perform operation 640.

[0166] In operation 630, the electronic device (401) may train an AI model based on local data of the delegated device (e.g., the electronic device (401)). In one embodiment, the electronic device (401) may train an AI model based on local data of the delegated device (e.g., the electronic device (101)) and local data of the delegated device (e.g., the electronic device (401)).

[0167] In operation 640, the electronic device (401) may train an AI model based on local data of the delegated device (e.g., electronic device (101)). In one embodiment, the electronic device (401) may train the AI ​​model based on local data of the delegated device (e.g., electronic device (101)) without using local data of the delegated device (e.g., electronic device (401)) when training the AI ​​model.

[0168] In operation 650, the electronic device (401) may evaluate the AI ​​model. For example, the electronic device (401) may evaluate an updated AI model. For example, the electronic device (401) may generate data (or metrics) related to the accuracy and / or loss of the updated AI model. However, this is not limited thereto.

[0169] Depending on the embodiment, operation 650 may be omitted. For example, the electronic device (401) may not perform evaluation of the AI ​​model. In this case, the data transmitted by the electronic device (401) to the server (108) (or the electronic device (101)) may include updated parameters of the AI ​​model.

[0170] In operation 660, the electronic device (401) may transmit the learning result to the server (108) or the electronic device (101). The electronic device (401) may transmit data indicating the learning result to the server (108) or the electronic device (101) before the expiration of the round. For example, the electronic device (401) may transmit the learning result to the electronic device (101). The electronic device (101) that receives the learning result from the electronic device (401) may transmit data indicating the learning result to the server (108) before the expiration of the round. For example, the electronic device (401) may directly transmit the learning result to the server (108). The electronic device (401) may directly transmit data indicating the learning result to the server (108) before the expiration of the round.

[0171] FIG. 7 is a flowchart of operations for transmitting learning results of an AI model performed by an electronic device according to an embodiment.

[0172] FIG. 7 can be explained with reference to FIG. 1, FIG. 2a, FIG. 2b, FIG. 3, and FIG. 4.

[0173] In FIG. 7, the electronic device (401) may be an electronic device that has formed a trust relationship with the electronic device (101). For example, the electronic device (401) may be one of the electronic devices (211, 213, or 215) of FIGS. 2A and 2B. Operations 410, 430, 435, and 440 of FIG. 7 may correspond to operations 410, 430, 435, and 440 of FIG. 4, respectively.

[0174] In the following examples, the operations may be performed sequentially, but are not necessarily sequential. For example, the order of the operations may be changed, and at least two operations may be performed in parallel.

[0175] Referring to FIG. 7, in operation 410, the server (108) may request training of an AI model from an electronic device (101). In one embodiment, in a round for training an AI model, the server (108) may request training of an AI model from an electronic device (101) selected as an electronic device to perform training of the AI ​​model.

[0176] In operation 430, the electronic device (101) may train an AI model. In one embodiment, the electronic device (101) may train the AI ​​model based on the determination that the AI ​​model is capable of self-learning (or that delegation of the AI ​​model is not required).

[0177] At operation 435, the electronic device (101) may evaluate the updated AI model based on local data.

[0178] In operation 710, the electronic device (101) can determine whether self-transmission is possible. In one embodiment, the electronic device (101) can determine whether delegation of transmission of data representing learning results is required. In one embodiment, the electronic device (101) can determine whether delegation of transmission of data representing learning results is required based on device performance information and / or device usage information of the electronic device (101). For example, the electronic device (101) can transmit data representing learning results to the server (108) before the expiration of the round.

[0179] In one embodiment, if the electronic device (101) is expected to fail to transmit data indicating learning results to the server (108) before the end of the round, it may determine that self-transmission is not possible (or that delegation of transmission is required).

[0180] In one embodiment, based on the determination that self-transmission is not possible (or that delegation of transmission is required) in operation 710, the electronic device (101) may perform operation 720. In one embodiment, based on the determination that self-transmission is possible (or that delegation of transmission is not required) in operation 710, the electronic device (101) may perform operation 440.

[0181] In operation 440, the electronic device (101) may transmit data indicating learning results to the server (108) before the end of learning (or the end of transmission of data indicating learning results to be transmitted to the server (108).

[0182] In operation 720, the electronic device (101) may identify a transmittable electronic device. In one embodiment, the electronic device (101) may identify a learnable electronic device among candidate electronic devices (211, 213, 215) having a trust relationship (210). For example, an electronic device having the same user account as the user account of the electronic device (101) (e.g., a Samsung account) may be evaluated as having a trust relationship (210) with the electronic device (101). For example, an electronic device that is communicatively connected to the electronic device (101) may be evaluated as having a trust relationship (210) with the electronic device (101). For example, an electronic device that is communicatively connected to the electronic device (101) through a designated short-range network (e.g., Bluetooth, WiFi, P2P (peer to peer)) may be evaluated as having a trust relationship (210) with the electronic device (101). For example, an electronic device that has authorized a communication connection with an electronic device (101) to provide a specified function (e.g., screen mirroring) may be evaluated as having a trust relationship (210) with the electronic device (101). For example, an electronic device that is communicatively connected to a specified peripheral device (e.g., mouse, keyboard, printer) that is communicatively connected to the electronic device (101) may be evaluated as having a trust relationship (210) with the electronic device (101).

[0183] In one embodiment, the electronic device (101) can identify an electronic device capable of learning from among the electronic devices (211, 213, 215) of the trust relationship (210). In one embodiment, the electronic device (101) can select an electronic device capable of learning an AI model based on information (e.g., device performance information, device identification information, and / or device usage information) of each of the electronic devices (211, 213, 215). In one embodiment, the device performance information may include information about the computational capabilities (e.g., the computational capabilities of a CPU, a GPU, and / or an NPU) of each of the electronic devices (211, 213, 215). In one embodiment, the device identification information may include information about the ID (or product identification information) and / or type (e.g., a smartphone, a tablet, a PC, a TV, a refrigerator, a robot vacuum cleaner, an AI speaker, or an air conditioner) of each of the electronic devices (211, 213, 215). In one embodiment, the device usage information may include information about the network connection status of each of the electronic devices (211, 213, 215) (e.g., whether connected via a local area network (e.g., WiFi), battery status (e.g., whether charging, and / or remaining battery capacity), and / or whether idle (or whether performing a task).

[0184] In one embodiment, the electronic device (101) may identify an electronic device capable of learning and / or evaluating an AI model before the expiration time of a round among the electronic devices (211, 213, 215). In one embodiment, the electronic device (101) may calculate (or identify) a time required for learning and / or evaluating the AI ​​model of each of the electronic devices (211, 213, 215) based on information about each of the electronic devices (211, 213, 215). In one embodiment, the electronic device (101) may identify an electronic device capable of learning and / or evaluating an AI model before the expiration time based on the required time of each of the electronic devices (211, 213, 215).

[0185] In operation 730, the electronic device (101) may request transmission. In one embodiment, the electronic device (101) may request (or delegate) transmission of data representing learning results to an electronic device (401) selected from among the electronic devices (211, 213, and 215). For example, when requesting transmission of data representing learning results to the electronic device (401), the electronic device (101) may also transmit delegation conditions. In one embodiment, the delegation conditions may include information about the expiration time of a round and / or the type of data representing learning results to be transmitted to the server (108) (e.g., parameters of an updated AI model and / or evaluation results of an updated AI model). In one embodiment, the delegation conditions may include information about the destination of the data representing learning results to be transmitted (e.g., server (108)). In one embodiment, the delegation conditions may include information about whether additional learning is necessary based on local data of the delegated device.

[0186] In operation 740, the electronic device (401) may transmit the learning result. In one embodiment, the electronic device (401) may transmit data indicating the learning result to the server (108) before the end of the round.

[0187] In one embodiment, the electronic device (401) may train an AI model based on the determination that additional learning of the AI ​​model based on local data of the electronic device (401) is necessary. Thereafter, the electronic device (401) may transmit data representing the learning results of the additionally trained AI model to the server (108).

[0188] FIG. 8 is a flowchart of operations for transmitting learning results of an AI model assigned to an electronic device according to an embodiment.

[0189] FIG. 8 may be described with reference to FIGS. 1, 2a, 2b, 3, 4, and 7. Operations 820, 830, 840, and 850 of FIG. 8 may be included in operation 740 of FIG. 7.

[0190] In FIG. 8, the electronic device (401) may be an electronic device that has formed a trust relationship with the electronic device (101). For example, the electronic device (401) may be one of the electronic devices (211, 213, 215) of FIGS. 2A and 2B.

[0191] In operation 810, the electronic device (401) may receive a transmission request. In one embodiment, the electronic device (401) may receive a transmission request of data representing learning results from the electronic device (101) of the trust relationship (210). For example, the electronic device (401) may receive a transmission request of data including parameters of an updated AI model and / or evaluation results of an updated AI model from the electronic device (101).

[0192] In operation 820, the electronic device (401) may determine whether learning based on local data of the delegated device (or the electronic device (401)) is required. For example, the electronic device (401) may determine whether learning based on local data of the electronic device (401) is required based on the delegation condition. For example, based on the delegation condition indicating that learning based on local data of the electronic device (401) is required, the electronic device (401) may determine that learning based on local data of the electronic device (401) is required.

[0193] In one embodiment, the electronic device (401) may query the server (108) as to whether learning based on local data of the electronic device (401) is required (or possible). For example, based on the server (108) responding to the query by instructing (or requesting) learning based on local data of the electronic device (401), the electronic device (401) may determine that learning based on local data of the electronic device (401) is required (or possible).

[0194] In one embodiment, based on the determination that learning based on local data of the electronic device (401) is required in operation 820, the electronic device (401) may perform operation 830. In one embodiment, based on the determination that learning based on local data of the electronic device (401) is not required in operation 820, the electronic device (401) may perform operation 850.

[0195] In operation 830, the electronic device (401) may train an AI model based on local data of the receiving device. In one embodiment, the electronic device (401) may further train the AI ​​model trained in the electronic device (101) based on local data of the receiving device (e.g., the electronic device (401)). In one embodiment, the electronic device (401) may further train the AI ​​model based on local data of the receiving device (e.g., the electronic device (401)).

[0196] In operation 840, the electronic device (401) may evaluate the AI ​​model. For example, the electronic device (401) may evaluate an updated AI model. For example, the electronic device (401) may generate data (or metrics) related to the accuracy and / or loss of the updated AI model. However, the present invention is not limited thereto.

[0197] In operation 850, the electronic device (401) may transmit the learning result to the server (108). The electronic device (401) may transmit data indicating the learning result to the server (108) before the end of the round.

[0198] FIG. 9 illustrates an example of a user interface (UI) displayed in electronic devices according to one embodiment.

[0199] In one embodiment, the electronic device (101) may determine whether delegation of AI model learning is required. In one embodiment, based on the determination that delegation of AI model learning is required, the electronic device (101) may display a list including electronic devices (211, 213, 215) of the trust relationship (210) (or electronic devices capable of AI model learning in the electronic devices (211, 213, 215) of the trust relationship (210)) through the display (260). For example, referring to FIG. 9, in state (901), the electronic device (101) may display a list (910) including A cell phone, B TV, C refrigerator, and D laptop on the display (260). In one embodiment, the electronic devices included in the list (910) may be electronic devices that have a trust relationship (210) with the electronic device (101), or may be electronic devices capable of learning an AI model among the electronic devices in the trust relationship (210).

[0200] In one embodiment, the electronic device (101) may display a UI (920) indicating an electronic device to which AI model learning is to be delegated on the display (260). In one embodiment, the electronic device (101) may display a UI (920) for inquiring about delegation of AI model learning to mobile phone A on the display (260).

[0201] In one embodiment, the electronic device (101) may, in response to a user input selecting a menu (931) for confirmation after outputting the UI (920), request training of the AI ​​model from the selected mobile phone A. In one embodiment, the electronic device (101) may, in response to a user input selecting a menu (935) for rejection after outputting the UI (920), display another UI on the display (260) that queries the delegation of training of the AI ​​model through another electronic device. However, the present invention is not limited thereto. In one embodiment, the electronic device (101) may, in response to a user input selecting a menu (935) for rejection, display a UI on the display (260) that requests a selection from among electronic devices included in the menu (910).

[0202] In one embodiment, the electronic device (401) may output a user interface (UI) that queries the user about the number of AI model training requests received from the electronic device (101). For example, referring to FIG. 9 , in state (903), the electronic device (401) may display a UI (950) indicating that a request for AI model training has been received from the electronic device (101) on the display (940).

[0203] In one embodiment, the electronic device (101) may display a UI (960) on the display (940) that queries whether to approve a request for delegation of learning of an AI model.

[0204] In one embodiment, the electronic device (401) may, in response to a user input selecting a menu (971) for confirmation after outputting the UI (960), transmit a signal to the electronic device (101) indicating approval for the AI ​​model's learning request.

[0205] In one embodiment, the electronic device (401) may, in response to a user input selecting a menu (975) for rejection after outputting the UI (960), transmit a signal to the electronic device (101) indicating rejection of the AI ​​model's learning request.

[0206] As described above, the electronic device (101) may include a communication circuit (290). The electronic device (101) may include at least one processor (120) including a processing circuit. The electronic device (101) may include a memory (130) that stores instructions and includes one or more storage media. The instructions, when individually or collectively executed by the at least one processor (120), may cause the electronic device (101) to receive a request for training an artificial intelligence (AI) model from a server (108) via the communication circuit (290). The instructions, when individually or collectively executed by the at least one processor (120), may cause the electronic device (101) to determine whether to delegate training of the AI ​​model. The instructions, when individually or collectively executed by the at least one processor (120), may cause the electronic device (101) to request, through the communication circuit (290), the learning of the AI ​​model from at least one electronic device (401) capable of learning the AI ​​model among one or more candidate electronic devices (211, 213, 215) having a relationship (210) of a specified condition with respect to the electronic device (101), based on a determination by the electronic device (101) to delegate the learning of the AI ​​model. The instructions, when individually or collectively executed by the at least one processor (120), may cause the electronic device (101) to perform the learning of the AI ​​model based on a determination by the electronic device (101) not to delegate the learning of the AI ​​model.The above instructions, when individually or collectively executed by the at least one processor (120), may cause the electronic device (101) to transmit data representing the result of the learning to the server (108) via the communication circuit (290) based on a decision not to delegate the learning of the AI ​​model.

[0207] The instructions, when individually or collectively executed by the at least one processor (120), may cause the electronic device (101) to identify the performance of each of the one or more candidate electronic devices (211, 213, 215) based on device information of each of the one or more candidate electronic devices (211, 213, 215). The instructions, when individually or collectively executed by the at least one processor (120), may cause the electronic device (101) to identify, based on the performance of each of the one or more candidate electronic devices (211, 213, 215), the at least one electronic device (401) having a performance higher than or equal to the performance for learning the AI ​​model.

[0208] The instructions, when individually or collectively executed by the at least one processor (120), may cause the electronic device (101) to identify an expected idle time of each of the one or more candidate electronic devices (211, 213, 215) based on a usage pattern of each of the one or more candidate electronic devices (211, 213, 215). The instructions, when individually or collectively executed by the at least one processor (120), may cause the electronic device (101) to identify, based on the expected idle time of each of the one or more candidate electronic devices (211, 213, 215), the at least one electronic device (401) having the expected idle time greater than or equal to the time required for training the AI ​​model.

[0209] The instructions, when individually or collectively executed by the at least one processor (120), may cause the electronic device (101) to request the learning of the AI ​​model to the at least one selected electronic device (401) via the communication circuit (290). The instructions, when individually or collectively executed by the at least one processor (120), may cause the electronic device (101) to transmit data for learning the AI ​​model to the at least one selected electronic device (401) based on obtaining approval for the request from the at least one selected electronic device (401) via the communication circuit (290).

[0210] The instructions, when individually or collectively executed by the at least one processor (120), may cause the electronic device (101) to determine whether the learning of the AI ​​model is possible by a designated learning expiration time. The instructions, when individually or collectively executed by the at least one processor (120), may cause the electronic device (101) to decide to delegate the learning of the AI ​​model based on the determination that the learning is not possible by the designated learning expiration time. The instructions, when individually or collectively executed by the at least one processor (120), may cause the electronic device (101) to decide not to delegate the learning of the AI ​​model based on the determination that the learning is not possible by the designated learning expiration time.

[0211] The instructions, when individually or collectively executed by the at least one processor (120), may cause the electronic device (101) to determine whether to perform the learning of the AI ​​model using data of the at least one selected electronic device (401) based on a determination that the learning of the AI ​​model is delegated. The instructions, when individually or collectively executed by the at least one processor (120), may cause the electronic device (101) to request the learning of the AI ​​model such that the AI ​​model is learned using the data of the at least one selected electronic device (401) based on a determination that the learning of the AI ​​model is to use the data of the at least one selected electronic device (401). The instructions, when individually or collectively executed by the at least one processor (120), may cause the electronic device (101) to request the learning of the AI ​​model such that the AI ​​model is learned using data other than the data of the at least one selected electronic device (401) based on a determination that the learning of the AI ​​model does not use the data of the at least one selected electronic device (401). The data other than the data of the at least one selected electronic device (401) may be provided from the electronic device (101) to the at least one selected electronic device (401).

[0212] The instructions, when individually or collectively executed by the at least one processor (120), may cause the electronic device (101) to perform the learning of the AI ​​model. The instructions, when individually or collectively executed by the at least one processor (120), may cause the electronic device (101) to evaluate the learned AI model using designated data. The instructions, when individually or collectively executed by the at least one processor (120), may cause the electronic device (101) to transmit data representing a result of performing the learning, including at least one parameter for updating the AI ​​model and an evaluation result of the learned AI model, to the server (108).

[0213] The instructions, when individually or collectively executed by the at least one processor (120), may cause the electronic device (101) to perform the learning of the AI ​​model. The instructions, when individually or collectively executed by the at least one processor (120), may cause the electronic device (101) to determine, before the learning of the AI ​​model is completed, whether the learning of the AI ​​model is possible by a designated learning completion time. The instructions, when individually or collectively executed by the at least one processor (120), may cause the electronic device (101) to request the learning of the partially learned AI model to the at least one electronic device (401) capable of learning the AI ​​model based on the determination that the learning is not possible by the designated learning completion time.

[0214] The instructions, when individually or collectively executed by the at least one processor (120), may cause the electronic device (101) to receive, through the communication circuit (290), other data representing a result of performing the learning from the at least one selected electronic device (401). The instructions, when individually or collectively executed by the at least one processor (120), may cause the electronic device (101) to transmit, through the communication circuit (290), other data representing a result of performing the learning to the server (108).

[0215] The instructions, when individually or collectively executed by the at least one processor (120), may cause the electronic device (101) to determine whether the data representing the result of performing the learning can be transmitted to the server (108). The instructions, when individually or collectively executed by the at least one processor (120), may cause the electronic device (101) to transmit the data to the server (108) via the communication circuit (290) based on the determination that the data can be transmitted to the server (108). The instructions, when individually or collectively executed by the at least one processor (120), may cause the electronic device (101) to request, through the communication circuit (290), at least one electronic device (401) selected from the one or more candidate electronic devices (211, 213, 215) to transmit the data to the server (108) based on a determination that the electronic device (101) is not capable of transmitting the data to the server (108).

[0216] The one or more candidate electronic devices (211, 213, 215) may be an electronic device (101) logged in with the same user account as the user account of the electronic device (101), an electronic device (101) connected to the electronic device (101) via a short-range network, or a combination thereof.

[0217] As described above, the method may be performed by an electronic device (101) including a communication circuit (290). The method may include an operation of receiving a request for learning an artificial intelligence (AI) model from a server (108) through the communication circuit (290). The method may include an operation of determining whether to delegate learning of the AI ​​model. Based on a determination to delegate the learning of the AI ​​model, the method may include an operation of requesting, through the communication circuit (290), the learning of the AI ​​model to at least one electronic device (401) capable of learning the AI ​​model among one or more candidate electronic devices (211, 213, 215) having a relationship with a specified condition to the electronic device (101). Based on a determination not to delegate the learning of the AI ​​model, the method may include an operation of performing the learning of the AI ​​model. The above method may include an operation of transmitting data indicating the results of the learning to the server (108) through the communication circuit (290).

[0218] The method may include an operation of identifying performance of each of the one or more candidate electronic devices (211, 213, 215) based on device information of each of the one or more candidate electronic devices (211, 213, 215). The method may include an operation of identifying at least one electronic device (401) having performance higher than or equal to performance for learning the AI ​​model based on the performance of each of the one or more candidate electronic devices (211, 213, 215).

[0219] The method may include an operation of identifying an expected idle time of each of the one or more candidate electronic devices (211, 213, 215) based on a usage pattern of each of the one or more candidate electronic devices (211, 213, 215). The method may include an operation of identifying, based on the expected idle time of each of the one or more candidate electronic devices (211, 213, 215), at least one electronic device (401) having the expected idle time longer than a time required for training the AI ​​model.

[0220] The method may include an operation of requesting the learning of the AI ​​model to at least one selected electronic device (401) through the communication circuit (290). The method may include an operation of transmitting data for learning the AI ​​model to the at least one selected electronic device (401) based on obtaining approval for the request from the at least one selected electronic device (401) through the communication circuit (290).

[0221] The method may include an operation for determining whether the learning of the AI ​​model is possible until a designated learning completion point. The method may include an operation for deciding to delegate the learning of the AI ​​model based on the determination that the learning is not possible until the designated learning completion point. The method may include an operation for deciding not to delegate the learning of the AI ​​model based on the determination that the learning is not possible until the designated learning completion point.

[0222] As described above, a non-transitory computer readable storage medium can store a program including instructions. The instructions, when individually or collectively executed by at least one processor (120) of an electronic device including a communication circuit (290), can cause the electronic device (101) to receive a request for training an artificial intelligence (AI) model from a server (108) via the communication circuit (290). The instructions, when individually or collectively executed by the at least one processor (120), can cause the electronic device (101) to determine whether to delegate training of the AI ​​model. The instructions, when individually or collectively executed by the at least one processor (120), may cause the electronic device (101) to request, through the communication circuit (290), the learning of the AI ​​model from at least one electronic device (401) capable of learning the AI ​​model among one or more candidate electronic devices (211, 213, 215) having a relationship (210) of a specified condition with respect to the electronic device (101), based on a determination by the electronic device (101) to delegate the learning of the AI ​​model. The instructions, when individually or collectively executed by the at least one processor (120), may cause the electronic device (101) to perform the learning of the AI ​​model based on a determination by the electronic device (101) not to delegate the learning of the AI ​​model.The above instructions, when individually or collectively executed by the at least one processor (120), may cause the electronic device (101) to transmit data representing the result of the learning to the server (108) via the communication circuit (290) based on a decision not to delegate the learning of the AI ​​model.

[0223] As described above, the electronic device (401) may include a communication circuit. The electronic device (401) may include at least one processor including a processing circuit. The electronic device (401) may include a memory that stores instructions and includes one or more storage media. The instructions, when individually or collectively executed by the at least one processor, may cause the electronic device (401) to receive a request for learning an artificial intelligence (AI) model from another electronic device (101) having a relationship with a specified condition to the electronic device (401) through the communication circuit. The instructions, when individually or collectively executed by the at least one processor, may cause the electronic device (401) to determine whether learning of the AI ​​model is possible. The instructions, when individually or collectively executed by the at least one processor, may cause the electronic device (401) to perform the learning of the AI ​​model based on determining that the learning of the AI ​​model is possible. The instructions, when individually or collectively executed by the at least one processor, may cause the electronic device (401) to transmit data representing the result of performing the learning to the server (108) that requested the learning of the AI ​​model through the communication circuit.

[0224] The electronic device (401) may include a display (940). The instructions, when individually or collectively executed by the at least one processor, may cause the electronic device (401) to display a screen through the display (940) for inquiring whether to approve the request for learning of the AI ​​model based on the request for learning of the AI ​​model. The instructions, when individually or collectively executed by the at least one processor, may cause the electronic device (401) to transmit a signal approving the request for learning to the other electronic device (101) through the communication circuit based on obtaining an input approving the request for learning after displaying the screen.

[0225] The instructions, when individually or collectively executed by the at least one processor, may cause the electronic device (401) to receive data for training the AI ​​model from the other electronic device (101) via the communication circuit. The instructions, when individually or collectively executed by the at least one processor, may cause the electronic device (401) to train the AI ​​model based on the data for training the AI ​​model and other data for training the AI ​​model generated by the electronic device (401).

[0226] The above instructions, when individually or collectively executed by the at least one processor, may cause the electronic device (401) to transmit the data representing the result of performing the learning to the server (108) via the other electronic device (101).

[0227] As described above, the method can be performed by an electronic device (401) including a communication circuit. The method can include an operation of receiving a request for learning an artificial intelligence (AI) model from another electronic device (101) having a relationship with a specified condition with respect to the electronic device (401) through the communication circuit. The method can include an operation of determining whether learning of the AI ​​model is possible. The method can include an operation of performing the learning of the AI ​​model based on a determination that the learning of the AI ​​model is possible. The method can include an operation of transmitting data representing a result of performing the learning to a server (108) that requested learning of the AI ​​model through the communication circuit.

[0228] As described above, a non-transitory computer readable storage medium can store a program including instructions. The instructions, when individually or collectively executed by at least one processor of an electronic device including a communication circuit, can cause the electronic device (401) to receive a request for learning an artificial intelligence (AI) model from another electronic device (101) having a relationship with a specified condition to the electronic device (401) through the communication circuit. The instructions, when individually or collectively executed by the at least one processor, can cause the electronic device (401) to determine whether learning of the AI ​​model is possible. The instructions, when individually or collectively executed by the at least one processor, can cause the electronic device (401) to perform learning of the AI ​​model based on determining that learning of the AI ​​model is possible. The above instructions, when individually or collectively executed by the at least one processor, may cause the electronic device (401) to transmit data representing the result of the learning to the server (108) that requested the learning of the AI ​​model through the communication circuit.

[0229] Electronic devices according to the various embodiments disclosed in this document may take various forms. 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 appliances. Electronic devices according to the embodiments of this document are not limited to the aforementioned devices.

[0230] The various embodiments of this document and the terminology used therein are not intended to limit the technical features described in this document to specific embodiments, but should be understood to 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 indicates 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 among those phrases, 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 component (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.

[0231] 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. A module may be an integral component, or a minimum unit or part of such a component 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).

[0232] 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 at least one called 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.

[0233] According to one embodiment, the method according to various embodiments disclosed in this document may be provided as included in a computer program product. The computer program product may be traded as a product between a seller and a buyer. 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., by download or upload) through 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 portion of the computer program product may be temporarily stored or temporarily generated in a machine-readable storage medium, such as the memory of a manufacturer's server, an application store's server, or an intermediary server.

[0234] According to various embodiments, each component (e.g., a module or a program) of the above-described components may include one or more entities, and some of the entities may be separated and arranged in other components. According to various embodiments, one or more components or operations of the aforementioned components may be omitted, or one or more other components or operations may be added. Alternatively or additionally, a plurality of components (e.g., a module or a program) may be integrated into a single component. In such a case, the integrated component may perform one or more functions of each of the plurality of components identically or similarly to those performed by the corresponding component among the plurality of components prior to the integration. According to various embodiments, the operations performed by a module, program, or other component may be executed sequentially, in parallel, iteratively, 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

1. In an electronic device (101), Communication circuit (290), At least one processor (120) including a processing circuit, and A memory (130) storing instructions and including one or more storage media, The above instructions, when individually or collectively executed by the at least one processor (120), cause the electronic device (101) to: Receive a learning request for an AI (artificial intelligence) model from a server (108) through the above communication circuit (290), Decide whether to delegate the learning of the above AI model, Based on the decision to delegate the learning of the AI model, through the communication circuit (290), requesting the learning of the AI model to at least one electronic device (401) capable of learning the AI model among one or more candidate electronic devices (211, 213, 215) having a relationship (210) of a specified condition with respect to the electronic device (101), and Based on the decision not to delegate the above learning of the above AI model, Performing the above learning of the above AI model, and Causing data indicating the results of the learning to be transmitted to the server (108) through the above communication circuit (290). Electronic devices.

2. In claim 1, The above instructions, when individually or collectively executed by the at least one processor (120), cause the electronic device (101) to: Based on the device information of each of the one or more candidate electronic devices (211, 213, 215), the performance of each of the one or more candidate electronic devices (211, 213, 215) is identified, Based on the performance of each of the one or more candidate electronic devices (211, 213, 215), causing identification of at least one electronic device (401) having a performance higher than the performance for learning the AI model. Electronic devices.

3. In claim 1 or claim 2, The above instructions, when individually or collectively executed by the at least one processor (120), cause the electronic device (101) to: Based on the usage pattern of each of the one or more candidate electronic devices (211, 213, 215), identifying the expected idle time of each of the one or more candidate electronic devices (211, 213, 215), Based on the expected idle time of each of the one or more candidate electronic devices (211, 213, 215), causing identification of at least one electronic device (401) having the expected idle time longer than the time required for training the AI model. Electronic devices.

4. In any one of claims 1 to 3, The above instructions, when individually or collectively executed by the at least one processor (120), cause the electronic device (101) to: Through the above communication circuit (290), requesting the learning of the AI model to the at least one selected electronic device (401), Through the communication circuit (290), based on obtaining approval for the request from the at least one selected electronic device (401), causing data for learning the AI model to be transmitted to the at least one selected electronic device (401). Electronic devices.

5. In any one of claims 1 to 4, The above instructions, when individually or collectively executed by the at least one processor (120), cause the electronic device (101) to: Determine whether the above learning of the above AI model is possible until the specified learning expiration time, Based on the determination that the above learning is not possible by the above designated learning expiration time, it is decided to delegate the above learning of the AI model, Causing a decision not to delegate the learning of the AI model based on the determination that the learning is not possible by the designated learning expiration time. Electronic devices.

6. In any one of claims 1 to 5, The above instructions, when individually or collectively executed by the at least one processor (120), cause the electronic device (101) to: Based on the decision to delegate the learning of the AI model, whether to perform the learning of the AI model is determined using the data of at least one selected electronic device (401), Based on the determination that the learning of the AI model utilizes the data of the at least one selected electronic device (401), the learning of the AI model is requested so that the AI model is learned using the data of the at least one selected electronic device (401). Based on the determination that the learning of the AI model does not use the data of the at least one selected electronic device (401), the learning of the AI model is requested so that the AI model is learned using data other than the data of the at least one selected electronic device (401), and the data other than the data of the at least one selected electronic device (401) is provided from the electronic device (101) to the at least one selected electronic device (401). Electronic devices.

7. In any one of claims 1 to 6, The above instructions, when individually or collectively executed by the at least one processor (120), cause the electronic device (101) to: Perform the above learning of the above AI model, Using the specified data, evaluate the learned AI model, and Causing the server (108) to transmit data representing the performance result of the learning, including at least one parameter for updating the AI model and an evaluation result of the learned AI model. Electronic devices.

8. In any one of claims 1 to 7, The above instructions, when individually or collectively executed by the at least one processor (120), cause the electronic device (101) to: Perform the above learning of the above AI model, Before the learning of the AI model is completed, it is determined whether the learning of the AI model is possible by the designated learning expiration time, Based on the determination that completion of the above learning is not possible by the designated learning expiration time, causing the at least one electronic device (401) capable of learning the AI model to request learning of the partially learned AI model. Electronic devices.

9. In any one of claims 1 to 8, The above instructions, when individually or collectively executed by the at least one processor (120), cause the electronic device (101) to: Through the above communication circuit (290), other data indicating the performance result of the learning is received from at least one selected electronic device (401), Causing the other data representing the result of the learning to be transmitted to the server (108) through the communication circuit (290). Electronic devices.

10. In any one of claims 1 to 9, The above instructions, when individually or collectively executed by the at least one processor (120), cause the electronic device (101) to: Determine whether the data representing the results of the above learning can be transmitted to the server (108), Based on the determination that the above data can be transmitted to the server (108), the data is transmitted to the server (108) through the communication circuit (290), Based on the determination that the data cannot be transmitted to the server (108), causing a request to be made to at least one electronic device (401) selected from among the one or more candidate electronic devices (211, 213, 215) to transmit the data to the server (108) through the communication circuit (290). Electronic devices.

11. In any one of claims 1 to 10, The one or more candidate electronic devices (211, 213, 215) are an electronic device (101) logged in with the same user account as the user account of the electronic device (101), an electronic device (101) connected to the electronic device (101) via a short-range network, or a combination thereof. Electronic devices.

12. In the electronic device (401), communication circuit, At least one processor comprising a processing circuit, and A memory comprising one or more storage media for storing instructions, The above instructions, when individually or collectively executed by the at least one processor, cause the electronic device (401) to: Receive a learning request for an AI (artificial intelligence) model from another electronic device (101) having a relationship with a specified condition for the electronic device (401) through the above communication circuit, Determine whether the above AI model can be learned, Based on the determination that the above learning of the above AI model is possible, Performing the above learning of the above AI model, and Causing data representing the performance result of the learning to be transmitted to the server (108) that requested the learning of the AI model through the above communication circuit. Electronic devices.

13. In claim 12, Includes a display (940), The above instructions, when individually or collectively executed by the at least one processor, cause the electronic device (401) to: Based on the request for the learning of the AI model, a screen is displayed through the display (940) to inquire whether the request for the learning of the AI model is approved, After displaying the above screen, based on obtaining an input approving the request for learning, a signal approving the request for learning is transmitted to the other electronic device (101) through the communication circuit. Electronic devices.

14. In claim 12 or claim 13, The above instructions, when individually or collectively executed by the at least one processor, cause the electronic device (401) to: Through the above communication circuit, data for learning the AI model is received from the other electronic device (101), Causing the AI model to be trained based on the data for training the AI model and other data for training the AI model generated by the electronic device (401). Electronic devices.

15. In any one of claims 12 to 14, The above instructions, when individually or collectively executed by the at least one processor, cause the electronic device (401) to: Causing the data representing the result of the above learning to be transmitted to the server (108) via the other electronic device (101). Electronic devices.

Images