Method for selecting ai training data on basis of user type

Abstract

An operation method of an electronic device, according to an embodiment of the present invention, may comprise the steps of: acquiring first user data for identifying a user type on the basis of a plurality of preset measurement items; identifying the user type on the basis of the acquired first user data; acquiring second user data on the basis of a professional evaluation index required for at least one industrial category; extracting target data, which matches the user type, from among the second user data according to a data selection criterion of the industrial category; and adding the target data to a training data group matching the industrial category and the identified user type.

Description

How to Select AI Training Data Based on User Type

[0001] The present invention relates to a method for selecting AI training data based on user types. Through a method for selecting and collecting AI training data from a user perspective, it is possible to enhance the efficiency of AI models and implement a human-centered, ethical AI system.

[0002] The existing data acquisition and selection processes for training artificial intelligence models have primarily relied on utilizing large-scale spatial data (3D) or vast amounts of general user data. This approach is applied to technologies such as Large Language Models (LLM) and Physical AI, and requires the essential process of collecting and analyzing large volumes of data to improve model performance. However, this data selection and processing process has the following problems.

[0003] First, existing data collection methods often fail to adequately reflect the individual characteristics and environments of users. Second, while the data distillation process can remove unnecessary or biased information, there is a possibility of information loss. Third, there is a potential for high-cost and low-efficiency issues, such as model performance degradation and high power consumption, during the processing of large-scale data. Fourth, there is a shortage of high-quality data supply relative to demand, and latecomer countries in AI face difficulties in building large-scale data-based models. Fifth, despite the existence of specialized evaluation criteria required by specific industrial sectors, existing methods have failed to effectively reflect these factors, resulting in limitations in selecting data optimized for specific industries.

[0004] Accordingly, there is a need for data selection methods to train more precise and efficient artificial intelligence models, and in particular, there is a demand for optimal data selection techniques that reflect individual user characteristics and domain-specific industrial requirements.

[0005] Various embodiments of the present invention are designed to establish selection criteria for artificial intelligence learning data by analyzing more precise user types based on five major measurement items such as disposition, virtue, personality, cognitive faculty, and personal environments.

[0006] In addition, the present invention is designed to enable the selection of data suitable for a specific industry by reflecting specialized evaluation indicators required by industrial sector.

[0007] Furthermore, the present invention is designed to minimize information loss occurring in the existing data distillation process and to effectively extract important data by deriving data selection criteria through the combination of user type analysis and industry evaluation criteria.

[0008] The purposes of the present disclosure are not limited to those mentioned above, and other purposes and advantages of the present disclosure not mentioned may be understood from the following description and will be more clearly understood from the embodiments of the present disclosure. Furthermore, it will be readily apparent that the purposes and advantages of the present disclosure can be realized by the means and combinations thereof set forth in the claims.

[0009] A method of operation of an electronic device for acquiring user-customized artificial intelligence data according to an embodiment of the present invention may include the steps of: acquiring first user data to identify a user type based on a plurality of preset measurement items; identifying a user type based on the acquired first user data; acquiring second user data based on professional evaluation indicators required for at least one industry category; extracting target data that matches the user type among the second user data according to data selection criteria of the industry category; and adding the target data to a training data group that matches the industry category and the identified user type.

[0010] At this time, the above multiple measurement items may consist of disposition, virtue, personality, cognitive faculty, and personal environments.

[0011] And the step of identifying the user type includes the step of identifying the user as one of three user types based on the evaluation results for each of the aforementioned preset multiple measurement items, and the three user types may be set to be distinguished according to the degree of proactiveness and willpower regarding problem solving.

[0012] Since the present invention enables precise data selection based on five major measurement items, it has the effect of allowing artificial intelligence models to learn with more reliable data.

[0013] The present invention has the effect of improving the learning quality of a model by effectively removing unnecessary or biased data.

[0014] Since the present invention provides high precision in the initial data selection process, it can reduce unnecessary repetitive work compared to existing methods that require multiple data distillation steps.

[0015] FIG. 1 is a diagram illustrating the operation sequence of an electronic device according to an embodiment of the present invention.

[0016] FIG. 2 is a diagram illustrating the processor configuration of an electronic device according to an embodiment of the present invention.

[0017] FIG. 3 is a drawing for explaining the operation according to an embodiment of the present invention.

[0018] FIG. 4 is a diagram illustrating the configuration of an electronic device according to an embodiment of the present invention.

[0019] The advantages and features of the present invention and the methods for achieving them will become clear by referring to the embodiments described below in detail together with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below but can be implemented in various different forms. These embodiments are provided merely to ensure that the disclosure of the present invention is complete and to fully inform those skilled in the art of the scope of the present invention, and the present invention is defined only by the scope of the claims.

[0020] The terms used in this specification are for describing embodiments and are not intended to limit the invention. In this specification, the singular form includes the plural form unless specifically stated otherwise in the text. The terms "comprises" and / or "comprising" used in this specification do not exclude the presence or addition of one or more other components in addition to the components mentioned. Throughout the specification, the same reference numerals refer to the same components, and "and / or" includes each of the mentioned components and all combinations of one or more. Although terms such as "first," "second," etc., are used to describe various components, these components are not limited by these terms. These terms are used merely to distinguish one component from another. Therefore, the first component mentioned below may be the second component within the technical scope of the invention.

[0021] Unless otherwise defined, all terms used herein (including technical and scientific terms) may be used in a meaning commonly understood by those skilled in the art to which the present invention pertains. Additionally, terms defined in commonly used dictionaries are not to be interpreted ideally or excessively unless explicitly and specifically defined otherwise.

[0022] The terms “part” or “module” as used in the specification refer to software or hardware components, such as FPGAs or ASICs, and “part” or “module” perform certain roles. However, “part” or “module” is not limited to software or hardware. “Part” or “module” may be configured to reside in an addressable storage medium or configured to run on one or more processors. Thus, by example, “part” or “module” includes components such as software components, object-oriented software components, class components, and task components, as well as processes, functions, attributes, procedures, subroutines, segments of program code, drivers, firmware, microcode, circuits, data, databases, data structures, tables, arrays, and variables. The functions provided within the components and “parts” or “modules” may be combined into a smaller number of components and “parts” or “modules,” or further separated into additional components and “parts” or “modules.”

[0023] Spatially relative terms such as "below," "beneath," "lower," "above," and "upper" may be used to facilitate the description of the relationship between one component and other components as illustrated in the drawings. Spatially relative terms should be understood as encompassing different orientations of components during use or operation, in addition to the orientations depicted in the drawings. For example, if a component depicted in a drawing is inverted, a component described as "below" or "beneath" of another component may be placed "above" of that component. Therefore, the exemplary term "below" may encompass both the lower and upper directions. Components may also be oriented in other directions, and accordingly, spatially relative terms may be interpreted according to the orientation.

[0024] Embodiments of the present invention will be described below with reference to the attached drawings.

[0025] FIG. 1 is a diagram illustrating the sequence of operations of an electronic device according to an embodiment of the present invention. FIG. 1 illustrates the process of the electronic device (100) of the present invention performing customized data acquisition and optimization for learning an artificial intelligence model.

[0026] As illustrated in FIG. 1, an electronic device (100) according to an embodiment of the present invention may perform step S110 of acquiring first user data to identify a user type.

[0027] At this time, the first user data collected by the electronic device (100) may be collected based on measurement items including disposition, virtue, personality, cognitive faculty, and personal environments.

[0028] The explanation for each of these five measurement items is as follows. First, the aforementioned 'dignity' is a latent variable that measures the strength and darkness of innate temperament. The aforementioned 'dignity' is an item designed to measure a person's consistency as strength and weakness, and activity as brightness or darkness. For example, in the present invention, the value corresponding to 'dignity' can be measured as a larger value as the user is judged to be steadfast and consistent, and possesses a bright personality.

[0029] Next, the aforementioned 'Simdeok' is a variable that measures an individual's usual practice of ethical norms and fidelity (consistency between words and actions) established through their acquired efforts; the value of 'Simdeok' can be measured as higher the more a user is judged to possess ethical norms and fidelity.

[0030] Furthermore, the aforementioned 'Cheyong' is an item designed to measure consideration for others, catholicity, and social adjustment; the higher a user is judged to have high catholicity and social adjustment, the higher the value for 'Cheyong' may be measured.

[0031] Furthermore, the aforementioned 'knowledge' is a variable that measures a person's problem-solving ability, and the higher a user is evaluated to have excellent problem-solving ability, the higher the value of the aforementioned 'knowledge' may be measured.

[0032] Finally, the above 'situation' is a variable that measures a person's social position and the degree of social activity. In addition, the above 'situation' is a variable that measures the degree of social support from organizational members or people around them, and the social sphere valued in community activities (social achieved status and social sphere). The value of the 'situation' item may be measured as high as the user's social support or social sphere ability is judged to be excellent.

[0033] The electronic device (100) may generate a question capable of evaluating each of the five measurement items, for example, in order to collect user data for the five measurement items. Thus, the electronic device (100) may obtain user response information for the generated question. That is, according to one embodiment, the first user data may refer to response information for the question generated by the electronic device (100).

[0034] However, the first user data is not limited to responses to a survey. According to various embodiments, the electronic device (100) can acquire user behavior information through devices such as wearable devices, smartphone sensors, cameras, and AI speakers, and can perform an evaluation of the five measurement items based on the acquired user behavior information. In addition, the electronic device (100) can collect the first user data in various ways.

[0035]

[0036] After step S110, the electronic device (100) can perform step S120, which identifies a user type based on the first user data.

[0037] At this time, the electronic device (100) can classify users into three types based on evaluation scores for the values ​​of the five measurement items.

[0038] The above three types may correspond, for example, to avoidance type, compromise type, and problem-solving type. In this case, the electronic device (100) can evaluate the user's level of proactiveness and willpower regarding problem-solving based on evaluation scores for five measurement items, and can identify the user as one of the avoidance type, compromise type, or problem-solving type according to the evaluated level.

[0039] The aforementioned avoidance type is the one with the lowest proactiveness and willpower regarding problem-solving, while the problem-solving type is the one with the highest proactiveness and willpower and is full of confidence in overcoming problems. Additionally, the aforementioned compromise type may fall under the category of an intermediate type, with proactiveness regarding problem-solving higher than the avoidance type but lower than the problem-solving type.

[0040] Specifically, the electronic device (100) can identify the user as an avoidance type if the measured values ​​for all five measurement items are below a threshold, identify the user as a compromise type if the measured values ​​for disposition and cognitive faculty among the five measurement items are below a threshold, and identify the user as a problem-solving type if the measured values ​​for all five measurement items are above a threshold.

[0041] After step S120, the electronic device (100) can perform step S130 of acquiring second user data based on professional evaluation indicators required for an industry category.

[0042] Specifically, the aforementioned industry categories may include various items such as healthcare, education, and industrial accident prevention. Furthermore, the specialized evaluation indicators required for these industry categories may, for example, correspond to heart rate, blood pressure, and body temperature in the healthcare field, while in the education field, they may correspond to learning speed and feedback response data.

[0043] After step S130, the electronic device (100) can perform step S140 of extracting target data that matches the user type among the second user data according to the industrial category data selection criteria.

[0044] For example, if the industry category is healthcare, the electronic device (100) may exclude basic health guide data instead of extracting high-intensity exercise data and heart rate data as target data for problem-solving users. On the other hand, for avoidance users, the electronic device (100) may select basic health information and low-intensity exercise data as target data and exclude high-intensity exercise data.

[0045] In this way, the electronic device (100) can select target data that matches the user type from among user data (second user data) corresponding to the industry category.

[0046] According to various embodiments, the electronic device (100) can derive reference information for second user data that matches a user type.

[0047]

[0048] After step S140, the electronic device (100) may perform step S150 of adding the target data to a group of training data that matches an industry category and an identified user type.

[0049] The electronic device (150) can finally add the target data extracted in step S140 to a training data group so that it can be used for training an AI model, and at this time, different data groups can be created for different user types even within the same industry.

[0050] In addition, at this time, the electronic device (100) can perform a data cleaning process for each data group to minimize data redundancy and reduce data size.

[0051]

[0052] FIG. 2 is a diagram illustrating the configuration of an electronic device according to an embodiment of the present invention.

[0053] The processor (140) of the electronic device (100) according to an embodiment of the present invention may be configured to include a data collection module (141), a user type analysis module (142), a data selection module (143), and a model learning module (144).

[0054] The above data collection module (141) can perform a preparation operation (e.g., generating survey questions) and a user data acquisition operation to collect data from the user.

[0055] Specifically, the data collection module (141) can collect first user data for five measurement items (dignity, character, physical ability, knowledge, and circumstances). At this time, the data collection module (141) can collect first user data through a user survey method.

[0056] For example, the data collection module (141) can generate a question about 'dignity' among the five measurement items. 'Dignity' is an item that measures the strength and weakness and brightness of innate temperament.

[0057] The above data collection module (141) can generate questions such as "Have you continued a planned exercise for more than one month?" as questions to measure the "dignity" item, and obtain responses from the user.

[0058] Next, the data collection module (141) can generate questions for measuring 'virtue'. 'Virtue' is a variable that measures an individual's usual practice of ethical norms and fidelity (consistency between words and actions) built through the individual's acquired efforts, and the value of 'virtue' can be measured as higher as the user is judged to have ethical norms and fidelity.

[0059] Depending on the characteristics of this item called 'virtue', the data collection module (141) can generate questions such as, for example, 'Do you usually act morally and in accordance with reason and common sense?', 'Do you usually receive evaluations that you lack humility?', 'Do you tend to be consistent with your words and actions?', and can obtain responses to these from the user.

[0060] The data collection module (141) above can generate questions for measuring 'personality' among five measurement items. 'Personality' is an item designed to measure consideration for others, catholicity, and social adjustment, and the value of 'personality' can be measured as high as the user is judged to have high catholicity and social adjustment.

[0061] Depending on the characteristics of this item called 'consideration,' the question generation unit (141) can generate a question such as, for example, 'Do you usually have a high level of consideration and tolerance for others?' and obtain a response from the user.

[0062] Next, the data collection module (141) can generate a question for measuring 'cognitive faculty' among five measurement items. The 'cognitive faculty' is a variable that measures a person's problem-solving ability, and the value of the 'cognitive faculty' can be measured as higher for users who are evaluated as having excellent problem-solving ability.

[0063] Depending on the characteristics of the item 'knowledge' mentioned above, the data collection module (141) can generate a question such as, for example, 'Do you tend to make good use of your own experiential knowledge and advice from people around you?' and obtain a user response corresponding to it.

[0064] Finally, the data collection module (141) can generate a question for measuring 'personal environments' among the five measurement items. 'Personal environments' is a variable that measures a person's social position and degree of social activity. 'Personal environments' is a variable that measures the degree of social support from organizational members or people around them, and the social organization (social achieved status and social sphere) that is valued in community activities, and the value of the 'personal environments' item can be measured as high as the user's social support or social organization ability is judged to be excellent.

[0065] Depending on the characteristics of the item 'situation' mentioned above, the data collection module (141) can generate questions such as, for example, 'are you generally expected to be respected and admired by members of the organization or people around you?' and obtain user responses to these questions.

[0066] In this manner, the data collection module (141) can generate questions for each of the five measurement items, conduct a survey, and obtain the corresponding user responses as first user data.

[0067] The above data collection module (141) may collect first user data based on this survey method, but is not limited thereto.

[0068] The above data collection module (141) can obtain first user data for user type identification through various data collection methods, such as behavior detection sensors and mobile log analysis.

[0069] And the above data collection module (141) can collect second user data, which is data for training AI models by industry category (e.g., healthcare field, education field, etc.).

[0070] For example, the data collection module (141) can obtain data corresponding to heart rate, blood pressure, body temperature, sleep patterns, etc. from the user as second user data for training an AI model related to healthcare.

[0071] In addition, the above data collection module (141) can obtain data from the user, such as learning speed, problem-solving ability (score), and concentration analysis data, as second user data for learning an AI model corresponding to the education field.

[0072] The above user type analysis module (142) can perform an operation of classifying user types based on the collected first user data.

[0073] The user type analysis module (142) can measure scores for each of the measurement items of dignity, moral character, physical ability, knowledge, and treatment based on the acquired first user data. Based on this, the user type analysis module (142) can identify the user type as one of three types.

[0074] At this time, the user type analysis module (142) can evaluate the level of the user's proactiveness and willpower regarding problem solving based on the evaluation score, and can identify the user as one of the avoidance type, compromise type, or problem-solving type according to the evaluated level.

[0075] The aforementioned avoidance type is the one with the lowest proactiveness and willpower regarding problem-solving, while the problem-solving type is the one with the highest proactiveness and willpower and is full of confidence in overcoming problems. Additionally, the aforementioned compromise type may fall under the category of an intermediate type, with proactiveness regarding problem-solving higher than the avoidance type but lower than the problem-solving type.

[0076] Alternatively, according to various embodiments, the user type analysis module (142) may identify the user type through a preset AI model for user type identification.

[0077] In this way, the AI ​​model for user type identification (hereinafter, type identification model) can perform data operations to select one of three types of user types.

[0078] In this case, the user type analysis module (142) may use a type identification model generated by supervised learning based on response data to a survey (first user data) and user type information (correct answer label) of the user.

[0079] In this way, when the user type analysis module (142) identifies a user type using a previously trained type identification model, the user type can be predicted with only a reduced amount of the first user data.

[0080] Alternatively, the user type analysis module (142) can classify user types by inputting the first user data collected in real time into a pre-generated type identification model, even when it intends to identify user types using first user data collected in a method other than a survey method (e.g., a method of acquiring user data through a sensor of a user terminal, an AI speaker, etc.).

[0081] Accordingly, the user type analysis module (142) can input user data (first user data) obtained in various ways, as shown in FIG. 3, into the type identification model to identify the target user as one of three user types.

[0082] The above data selection module (143) can perform the operation of extracting target data from the second user data according to the identified user type and adding the extracted target data to the training data.

[0083] The above second user data corresponds to data obtained based on pre-established professional evaluation indicators in a specific industry category. However, the above second user data may include data that is meaningful to one type of user but meaningless to other types of users.

[0084] Accordingly, the data selection module (143) can select to extract only meaningful data from the entire data (second user data) as target data according to the user type, and not use the second user data other than the target data as training data for the AI ​​model. Accordingly, since the data selection module (143) does not perform model training using the entire data, it can prevent the waste of resources, such as costs and time, required for unnecessary data computation.

[0085] The data selection module (143) extracts target data that matches the user type among the second user data according to the data selection criteria of the industry category. At this time, whether it matches the user type is determined by measuring the contribution in the user type model to be created for each of the multiple data categories constituting the second user data, and if the contribution is less than the threshold, it can be determined that the data category does not match the user type.

[0086] Specifically, the second user data may include multiple data categories and be composed of parameters for each category. For example, data categories constituting the second user data for the education field may include learning speed, attention time, frequency of repetitive learning, dropout rate, number of answer changes, utilization of learning tools, learning fatigue, preferred learning type, stress index, etc.

[0087] And the data selection module (143) can extract only the data corresponding to the category in which the contribution is identified as being greater than or equal to a threshold value as target data by measuring the degree of contribution for each data category constituting the acquired second user data.

[0088] In this case, the contribution can be determined based on the difference in the result values ​​when data from a specific data category (e.g., stress index) is included in the target AI model to be created (e.g., a model related to education for avoidant users (learning solution recommendation)) versus when it is not.

[0089] In this case, the artificial intelligence model used to verify the contribution may refer to a sample model trained on only a portion of the data.

[0090] The above data selection module (143) can perform an analysis using at least one of Information Gain (IG), Feature Importance (using Random Forest, etc.), and SHAP (Shapley Additive Explanations) when measuring the contribution of each data category to the model based on the result of applying data categories to the training of the sample model. The aforementioned analysis methods can quantify the contribution of each data category.

[0091] If the difference in results between a sample model A, which is not trained on data of category A, and a sample model B, which is trained on data of all categories, is small, the data of category A can be identified as having a low contribution.

[0092] Thus, the above data selection module (143) can select data in categories with high contribution as data (target data) needed for AI learning, and determine data with low contribution as unnecessary data.

[0093] For example, in an education-related model for avoidant users, 'dropout rate' and 'frequency of repeated learning' may be identified as important data categories, while 'attention time' and 'learning speed' may be identified as relatively unimportant data categories.

[0094] The data selection module (143) can identify a data category as unimportant if the contribution value for each data category is below a threshold, and can determine a data category as important if the contribution value is above the threshold, thereby determining the data of that category as target data. However, it is not limited thereto, and the data selection module (143) can determine the criteria for setting target data by comprehensively considering not only the contribution derived through a sample model but also the importance determined through other evaluation indicators (e.g., statistically confirmed association).

[0095] Alternatively, the data selection module (143) may determine the setting criteria for the target data through various other methods.

[0096] For example, the data selection module (143) can perform target data matching by user type based on the correlation values ​​between each data category and the willpower to solve the problem.

[0097] Specifically, the data selection module (143) can quantify the attributes of each data category and convert them into vectors, and measure the similarity between the vector values ​​by comparing them with the vector value for willpower, which serves as the classification criterion for user types. Furthermore, based on the measured similarity, the data selection module (143) can determine that the higher the similarity, the more likely it is to be matched with a user type that has higher willpower. For example, if the measured similarity is a value corresponding to the first level (highest), the data category may be determined to be matched with the problem-solving type, which is the type with the highest level of willpower, and accordingly, it may be set as target data.

[0098] According to various embodiments, the data selection module (143) can determine the setting criteria for target data by considering both the contribution calculated by applying to the sample model and the vector value similarity.

[0099] In addition, the data selection module (143) can select target data that matches the user type among the second user data and add it as a training data group for training the target AI model later.

[0100] In addition, the data selection module (143) can perform various data selection operations to remove additional noise in addition to the target data selection operation.

[0101] The above model training module (144) can train an AI model for each industry category based on the final derived training data group.

[0102] In addition, the above model learning module (144) can generate an AI model corresponding to a specific industry category for each user type through model training using the above learning data group.

[0103] In summary, the method of operation of an electronic device for acquiring user-customized artificial intelligence data according to an embodiment of the present invention may include the steps of: acquiring first user data to identify a user type based on a plurality of preset measurement items; identifying a user type based on the acquired first user data; acquiring second user data based on professional evaluation indicators required for at least one industry category; extracting target data that matches the user type among the second user data according to data selection criteria of the industry category; and adding the target data to a training data group that matches the industry category and the identified user type.

[0104] At this time, the above multiple measurement items may consist of disposition, virtue, personality, cognitive faculty, and personal environments.

[0105] And the step of identifying the user type includes the step of identifying the user as one of three user types based on the evaluation results for each of the aforementioned preset multiple measurement items, and the three user types may be set to be distinguished according to the degree of proactiveness and willpower regarding problem solving.

[0106]

[0107] An electronic device (100) according to an embodiment of the present invention may include a memory (110), an input / output interface (120), a communication interface (130), a processor (140), etc., as shown in FIG. 4.

[0108] The electronic device may be a server, a user terminal, or a device that performs various other operations and functions. If the electronic device is a server, it may be an FTP server (File Transfer Protocol Server), a web server, a database server, or a cloud server, but is not limited thereto. The server may perform one or more functions and operations, and may be implemented as a single device or distributed across multiple devices to implement each function and operation.

[0109] Memory can store various programs and data necessary for the operation of electronic devices. Memory can be implemented as non-volatile memory, volatile memory, flash memory, hard disk drives (HDD), or solid-state drives (SSD).

[0110] An input / output interface may be a configuration in which an electronic device receives data, signals, information, etc. from another device or transmits data, signals, information, etc. to another device.

[0111] The input / output interface may be a port or terminal capable of transmitting and receiving various types of data (e.g., text, audio, image, video, etc.) and configured to be connected to another device provided separately from the electronic device. For example, the input / output interface may be a Universal Serial Bus (USB) terminal, and may be any one of HDMI (High Definition Multimedia Interface), MHL (Mobile High-Definition Link), USB (Universal Serial Bus), DP (Display Port), Thunderbolt, VGA (Video Graphics Array) port, RGB port, D-SUB (D-subminiature), or DVI (Digital Visual Interface), but is not limited thereto.

[0112] The communication interface may include at least one of a wireless communication interface, a wired communication interface, and an input / output interface. The wireless communication interface may communicate with various external devices using wireless communication technology or mobile communication technology. Such wireless communication technologies may include, for example, Bluetooth, Bluetooth Low Energy, CAN communication, Wi-Fi, Wi-Fi Direct, ultrawide band (UWB), Zigbee, infrared data association (IrDA), or near field communication (NFC), and mobile communication technologies may include 3GPP, Wi-Max, LTE (Long Term Evolution), 5G, etc.

[0113] A wireless communication interface can be implemented using an antenna, a communication chip, a substrate, etc., capable of transmitting electromagnetic waves to the outside or receiving electromagnetic waves transmitted from the outside.

[0114] A wired communication interface can communicate with various devices based on a wired communication network. Here, the wired communication network can be implemented using physical cables, such as, for example, pair cables, coaxial cables, fiber optic cables, or Ethernet cables.

[0115] The processor can control the overall operation of a user device using various programs stored in memory. The processor may be composed of RAM, ROM, a graphics processing unit, a main CPU, first to n interfaces, and a bus. At this time, the RAM, ROM, graphics processing unit, main CPU, first to n interfaces, etc., may be connected to each other through a bus.

[0116] RAM stores the operating system and application programs. Specifically, when an electronic device boots up, the operating system is stored in RAM, and various application data selected by the user can be stored in RAM.

[0117] The ROM stores a set of instructions for booting the system, etc. When a turn-on command is input and power is supplied, the main CPU copies the O / S stored in memory (200) to RAM according to the instructions stored in the ROM, and runs the O / S to boot the system. When booting is complete, the main CPU copies various application programs stored in memory to RAM, and runs the application programs copied to RAM to perform various operations.

[0118] The main CPU accesses memory and performs operations, including booting and execution, using the OS stored in memory. Additionally, the main CPU performs various operations using various programs, content, and data stored in memory.

[0119] The first to n interfaces are connected to the various components described above. One of the first to n interfaces may be a network interface connected to an external device through a network.

[0120] Furthermore, the processor can control the artificial intelligence model. In this case, the control unit may, of course, include a dedicated graphics processor (e.g., GPU) for controlling the artificial intelligence model.

[0121] The processor may include one or more cores (not shown) and a graphics processing unit (not shown) and / or a connection channel (e.g., a bus, etc.) for transmitting and receiving signals with other components.

[0122] A processor according to one embodiment performs the method described in connection with the present invention by executing one or more instructions stored in memory.

[0123] Meanwhile, the processor may further include RAM (Random Access Memory, not shown) and ROM (Read-Only Memory, not shown) for temporarily and / or permanently storing signals (or data) processed within the processor. Additionally, the processor (130) may be implemented in the form of a system-on-chip (SoC) that includes at least one of a graphics processing unit, RAM, and ROM.

[0124] Memory can store programs (one or more instructions) for processing and controlling the processor. Programs stored in the storage unit can be divided into multiple modules according to their function.

[0125] The steps of the method or algorithm described in connection with embodiments of the present invention may be implemented directly in hardware, implemented as a software module executed by hardware, or implemented by a combination thereof. The software module may reside in RAM (Random Access Memory), ROM (Read Only Memory), EPROM (Erasable Programmable ROM), EEPROM (Electrically Erasable Programmable ROM), Flash Memory, a hard disk, a removable disk, a CD-ROM, or any form of computer-readable recording medium well known in the art to which the present invention belongs.

[0126] The components of the present invention may be implemented as a program (or application) and stored on a medium to be executed in combination with a computer, which is hardware. The components of the present invention may be implemented as software programming or software elements, and similarly, embodiments may be implemented in programming or scripting languages ​​such as C, C++, Java, assembler, etc., including various algorithms implemented as combinations of data structures, processes, routines, or other programming configurations. Functional aspects may be implemented as algorithms executed on one or more processors.

[0127] Although the present invention has been described in detail with reference to the examples above, those skilled in the art may make modifications, changes, and variations to these examples without departing from the scope of the invention. In short, it should be noted that in order to achieve the intended effect of the present invention, it is not necessary to separately include all functional blocks shown in the drawings or follow all sequences shown in the drawings exactly as shown, and that such matters may fall within the technical scope of the invention as described in the claims.

Claims

1. A method of operation of an electronic device for acquiring user-customized artificial intelligence data, A step of obtaining first user data to identify a user type based on a plurality of preset measurement items; A step of identifying a user type based on the first user data obtained above; A step of acquiring second user data based on professional evaluation indicators required for at least one industry category; A step of extracting target data that matches the user type among the second user data according to the data selection criteria of the above industry category; and A method of operating an electronic device comprising the step of adding the target data to a training data group that matches the above industry category and the above identified user type.

2. In Paragraph 1, The above plurality of measurement items are A method of operation of an electronic device comprising disposition, virtue, personality, cognitive faculty, and personal environments.

3. In Paragraph 1, The step of identifying the above user type is The method includes the step of identifying a user as one of three user types based on the evaluation results for each of the aforementioned preset multiple measurement items; A method of operation of an electronic device characterized by the fact that the above three user types are set to be distinguished according to the degree of activeness and willpower regarding problem solving.

4. In Paragraph 1, The step of extracting target data matching the above user type is The method includes the step of analyzing the contribution of each data category constituting the second user data and selecting a data category with a contribution greater than or equal to a threshold as the target data. The above contribution is A method of operating an electronic device characterized by being performed using at least one of Information Gain, Feature Importance, and SHAP (Shapley Additive Explanations) techniques.

Images