User-customized cognitive function training method using generative artificial intelligence model, and device thereof

Abstract

A user-customized cognitive function training device using a generative artificial intelligence model, according to one aspect, comprises: a memory on which at least one program is stored; and at least one processor for executing the at least one program, wherein the at least one processor determines the difficulty level of cognitive training on the basis of an evaluation result before training of a user, determines a training word from a word category on the basis of the difficulty level of the cognitive training, provides, to the user, the cognitive training based on the determined training word by using the generative artificial intelligence model, and provides cognitive training feedback on the basis of a cognitive training performance result of the user.

Description

User-customized cognitive function training method using generative AI models and the device thereof

[0001] The present disclosure relates to a user-customized cognitive function training method and apparatus utilizing a generative artificial intelligence model.

[0002] In general, human cognitive function consists of various sub-components such as attention, memory, judgment, and language ability, and these functions can decline due to causes such as aging, disease, and trauma. Accordingly, various training programs have been proposed to prevent or improve cognitive decline.

[0003] However, conventional cognitive training methods generally provide uniform content to users, which limits their ability to offer appropriate feedback or training paths tailored to individual cognitive abilities or responses. Furthermore, even when some methods collect user response data, they lack the functionality to analyze or incorporate it in real-time, which reduces training efficiency.

[0004] Therefore, there is a growing need for a more sophisticated and personalized method of providing cognitive training that can accurately evaluate the user's current cognitive state and dynamically adjust training content based on the evaluation results.

[0005] The invention provides a method and device for user-customized cognitive function training utilizing a generative artificial intelligence model. Additionally, the invention provides a computer-readable recording medium storing a program for executing the above method on a computer. The technical problems to be solved are not limited to those described above, and other technical problems may exist.

[0006] According to one aspect of the present disclosure, a user-customized cognitive function training method utilizing a generative artificial intelligence model can be provided, comprising: a step of determining the difficulty of cognitive training based on the user's pre-training evaluation results; a step of determining a training word from a word category based on the difficulty of the cognitive training; a step of providing the cognitive training based on the determined training word to the user using a generative artificial intelligence model; and a step of providing cognitive training feedback based on the user's performance results of the cognitive training.

[0007] An apparatus according to another aspect of the present disclosure comprises: a memory in which at least one program is stored; and at least one processor for executing said at least one program, wherein the at least one processor determines the difficulty of cognitive training based on the results of a user’s pre-training evaluation, determines a training word from a word category based on the difficulty of said cognitive training, provides said cognitive training based on said determined training word to said user using a generative artificial intelligence model, and provides said cognitive training feedback based on the results of said cognitive training performance by said user.

[0008] A computer-readable recording medium according to another aspect of the present disclosure includes a recording medium that records a program for executing the above-described method on a computer.

[0009] According to one embodiment of the present invention, by using a generative artificial intelligence model to communicate with the user in real-time and bidirectionally, it is possible to help improve the user's cognitive function.

[0010] In addition, cognitive reserve can be effectively activated by selectively providing a cognitive training method suitable for the user based on the results of the user's cognitive training performance.

[0011] The objectives of the present invention are not limited to those mentioned above, and other unmentioned objectives will be clearly understood by those skilled in the art from the description below.

[0012] FIG. 1 is a diagram illustrating an example of a user-customized cognitive function training method using a generative artificial intelligence model according to one embodiment.

[0013] FIG. 2 is a block diagram illustrating an example of a user-customized cognitive function training device utilizing a generative artificial intelligence model according to one embodiment.

[0014] FIG. 3 is a flowchart illustrating an example of a user-customized cognitive function training method using a generative artificial intelligence model according to one embodiment.

[0015] FIG. 4 is a diagram illustrating an example of a method for determining the difficulty level of cognitive training according to one embodiment.

[0016] FIG. 5 is a diagram illustrating an example of a method for determining training words according to one embodiment.

[0017] FIG. 6 is a diagram illustrating an example of a method for providing an introductory sentence or an introductory picture to a user according to one embodiment.

[0018] FIG. 7 is a diagram illustrating an example of a method for providing cognitive training to a user according to one embodiment.

[0019] FIGS. 8 and 9 are drawings for illustrating other examples of a method for providing cognitive training to a user according to one embodiment.

[0020] FIG. 10 is a diagram illustrating another example of a method for providing cognitive training to a user according to one embodiment.

[0021] A user-customized cognitive function training device utilizing a generative artificial intelligence model according to one aspect comprises: a memory in which at least one program is stored; and at least one processor that executes the at least one program, wherein the at least one processor determines the difficulty of cognitive training based on the user's pre-training evaluation results, determines training words from word categories based on the difficulty of cognitive training, provides the cognitive training based on the determined training words to the user using a generative artificial intelligence model, and provides cognitive training feedback based on the user's cognitive training performance results.

[0022] The terms used in the embodiments have been selected to be as close as possible to currently widely used general terms; however, these may vary depending on the intent of those skilled in the art, case law, the emergence of new technologies, etc. Additionally, in specific cases, terms have been selected at the applicant's discretion, and in such cases, their meanings will be described in detail in the relevant description section. Therefore, terms used in the specification must be defined not merely by their names, but based on their meanings and the content throughout the specification.

[0023] When a part of the specification is described as "comprising" a certain component, this means that, unless specifically stated otherwise, it does not exclude other components but may include additional components. Furthermore, terms such as "~ unit" or "~ module" as used in the specification refer to a unit that processes at least one function or operation, and this may be implemented in hardware or software, or as a combination of hardware and software.

[0024] Additionally, terms including ordinal numbers, such as "first" or "second," used in the specification may be used to describe various components, but said components should not be limited by said terms. Such terms may be used for the purpose of distinguishing one component from another.

[0025] The present disclosure will be described in detail below with reference to the attached drawings. Specifically, a user-customized cognitive function training method utilizing a generative artificial intelligence model according to one embodiment will be described in more detail with reference to FIGS. 1 to 10. However, the embodiment may be implemented in various different forms and is not limited to the examples described herein.

[0026]

[0027] FIG. 1 is a diagram illustrating an example of a user-customized cognitive function training method using a generative artificial intelligence model according to one embodiment.

[0028] Hereinafter, with reference to FIG. 1, an example of a user-customized cognitive function training method using a generative artificial intelligence model will be described.

[0029] Referring to FIG. 1, a user (1) can perform cognitive training using a user-customized cognitive function training device (10) utilizing a generative artificial intelligence model.

[0030] A user-customized cognitive function training device (10) utilizing a generative artificial intelligence model can provide cognitive training to a user (1) and provide feedback on the input of the user (1) performing the provided cognitive training. Here, the input of the user (1) may include all forms of input, such as voice, text, touch, motion, and biosignals (e.g., brainwaves, gaze, heart rate, etc.).

[0031] In the present disclosure, a user-customized cognitive function training device (10) utilizing a generative artificial intelligence model can provide cognitive training to a user (1) using a generative artificial intelligence model.

[0032] For example, a user-customized cognitive function training device (10) utilizing a generative artificial intelligence model can provide cognitive training to a user (1) and obtain input from the user (1) performing the provided cognitive training. Additionally, the user-customized cognitive function training device (10) utilizing a generative artificial intelligence model can input the user's (1) input obtained as input data for the generative artificial intelligence model into the artificial intelligence model, analyze the user's (1) input to derive feedback as output data for the generative artificial intelligence model, and provide it to the user (1).

[0033] Here, an artificial intelligence model refers to a set of machine learning algorithms that use a layered algorithmic structure based on deep neural networks in machine learning technology and cognitive science.

[0034] For example, an artificial intelligence model may be composed of an input layer that receives input signals or data from the outside, an output layer that outputs output signals or data corresponding to the input data, and at least one hidden layer located between the input layer and the output layer, which receives a signal from the input layer, extracts a feature, and transmits it to the output layer. The output layer receives a signal or data from the hidden layer and outputs it to the outside.

[0035] More specifically, the generative artificial intelligence model in the present disclosure may refer to a model that generates text, images, audio, etc., as output data using input data. That is, the generative artificial intelligence model in the present disclosure is an artificial intelligence model capable of understanding and generating human language patterns, and can determine whether emotional words included in the input of the user (1) are used appropriately, whether the context of the text is natural, and whether a predetermined word is used in an appropriate position. In addition, the generative artificial intelligence model in the present disclosure may generate text, images, etc. based on training words, emotional themes, etc.

[0036] Accordingly, a user (1) performing cognitive training can receive real-time feedback by communicating bidirectionally with a user-customized cognitive function training device (10) utilizing a generative artificial intelligence model, and can also receive cognitive training feedback after the cognitive training is completed.

[0037] Meanwhile, the cognitive training, training words, etc. provided to the user in the present disclosure may be provided to the user through a predetermined input / output device, such as a display, interface, or speaker, included in a user-customized cognitive function training device (10) utilizing a generative artificial intelligence model, and may be provided visually or audibly.

[0038]

[0039] FIG. 2 is a block diagram illustrating an example of a user-customized cognitive function training device utilizing a generative artificial intelligence model according to one embodiment.

[0040] Referring to FIG. 2, a user-customized cognitive function training device (hereinafter referred to as the 'device') (200) utilizing a generative artificial intelligence model may include a communication unit (210), a processor (220), and a memory (230). Only the components related to the embodiment are illustrated in the device (200) of FIG. 2. Therefore, it is obvious to a person skilled in the art that other general-purpose components may be included in addition to the components illustrated in FIG. 2.

[0041] The communication unit (210) may include one or more components that enable wired / wireless communication with an external server or external device. For example, the communication unit (210) may include a short-range communication unit (not shown) and a mobile communication unit (not shown) for communication with an external server or external device.

[0042] The memory (230) is hardware that stores various data processed within the device (200) and can store a program for processing and controlling the processor (220).

[0043] For example, various data may be stored in the memory (230), such as the user's medical data, data evaluating the user's cognitive function, the user's previous cognitive training performance data, the user's cognitive training performance data, cognitive training-related data, word category data, training words, user input, real-time feedback data, cognitive training feedback data, and data generated according to the operation of the processor (220). Additionally, the memory (230) may store an operating system (OS) and at least one program (e.g., a program required for the processor (220) to operate).

[0044] The memory (230) may include RAM (random access memory), such as DRAM (dynamic random access memory) and SRAM (static random access memory), ROM (read-only memory), EEPROM (electrically erasable programmable read-only memory), CD-ROM, Blu-ray or other optical disc storage, HDD (hard disk drive), SSD (solid state drive), or flash memory.

[0045] The processor (220) controls the overall operation of the device (200). For example, the processor (220) can control the input unit (not shown), display (not shown), communication unit (210), memory (230), etc., by executing programs stored in memory (230).

[0046] The processor (220) may be implemented using at least one of ASICs (application specific integrated circuits), DSPs (digital signal processors), DSPDs (digital signal processing devices), PLDs (programmable logic devices), FPGAs (field programmable gate arrays), controllers, micro-controller units, microprocessors, and other electrical units for performing functions.

[0047] The processor (220) can control the operation of the device (200) by executing programs stored in memory (230). For example, the processor (220) can perform at least some of the user-customized cognitive function training methods using a generative artificial intelligence model described with reference to FIGS. 3 to 10.

[0048]

[0049] FIG. 3 is a flowchart illustrating an example of a user-customized cognitive function training method using a generative artificial intelligence model according to one embodiment.

[0050] Hereinafter, with reference to FIG. 3, an example of a user-customized cognitive function training method using a generative artificial intelligence model will be described.

[0051] Referring to FIG. 3, a user-customized cognitive function training method using a generative artificial intelligence model may include steps 310 to 340. However, it is not limited thereto, and other general operations in addition to those shown in FIG. 3 may be further included in the user-customized cognitive function training method using a generative artificial intelligence model. Additionally, as described above with reference to FIG. 1 and FIG. 2, at least one of the operations of the flowchart shown in FIG. 3 may be processed by a processor (220).

[0052] First, in step 310, the processor can determine the difficulty of the cognitive training based on the user's pre-training evaluation results.

[0053] For example, the processor can provide the user with a pre-training evaluation based on previous training words provided in previous cognitive training, and determine the difficulty of the cognitive training according to the number of training words based on the user's pre-training evaluation results.

[0054] In step 320, the processor can determine training words from word categories based on the difficulty of cognitive training.

[0055] For example, the processor can obtain user input selecting at least one word category from among at least one candidate word category, and determine a training word from the word category based on the difficulty of cognitive training.

[0056] In step 330, the processor can use a generative artificial intelligence model to provide cognitive training based on determined training words to the user.

[0057] First, the processor can present the determined training words to the user in the form of an introductory sentence or an introductory picture.

[0058] For example, the processor may use a generative artificial intelligence model to generate at least one of an introductory sentence or an introductory picture based on training words, and provide at least one of the introductory sentence or an introductory picture to the user.

[0059] The processor may provide cognitive training to a user that includes multiple detailed training exercises. Here, the multiple detailed training exercises may include intensive training, association training, and associative training, and the multiple detailed training exercises may be classified based on the purpose of the cognitive training, the user's cognitive function state, etc. Additionally, the processor may provide the multiple detailed training exercises to the user sequentially in the order of intensive training, association training, and associative training, but may also provide them to the user in a different order or randomly depending on the user's cognitive function state, cognitive training evaluation results, etc.

[0060] The processor can provide intensive training to the user. Here, intensive training may refer to training that allows the user to improve cognitive function by distinguishing between training words and non-training words.

[0061] For example, the processor can provide the user with either a training word or a non-training word to obtain user input for classifying the training word or the non-training word, and provide real-time feedback to the user based on the user input.

[0062] The processor may provide association training to the user. Here, association training may refer to training that allows the user to improve cognitive function by sensorially describing training words. Additionally, the processor may provide the user with at least one of unit training using pictures and unit training using quizzes as association training.

[0063] First, the processor can provide the user with unit training using pictures as association training.

[0064] For example, the processor can acquire user input for drawing a first picture based on training words, calculate a score based on the degree of correspondence between the training words and the first picture using a generative artificial intelligence model, and provide real-time feedback to the user based on whether the score is greater than or equal to a preset score.

[0065] In addition, if the score is below a preset score, the processor may sequentially provide real-time drawing feedback to the user to draw a second picture based on the training words.

[0066] The processor can provide the user with unit training using quizzes as association training.

[0067] For example, the processor can obtain user input regarding the question by providing the user with a question about training words generated using a generative artificial intelligence model, derive a correct answer word based on the user's input, and provide real-time feedback to the user based on whether the correct answer word is the same as the training word.

[0068] The processor can provide associated training to the user. Here, associated training may refer to training that allows the user to improve cognitive function by generating emotion-based stories using training words.

[0069] For example, the processor may use a generative artificial intelligence model to generate story pictures based on training words and a predetermined emotional theme, provide the story pictures to the user to obtain user input in which the user inputs a sentence based on the story pictures, and provide real-time feedback to the user based on whether the user's input corresponds to the training words or the predetermined emotional theme.

[0070] As one example, the processor can determine whether there are missing training words in the user's input and provide real-time feedback to the user based on the missing training words. As another example, the processor can use a generative artificial intelligence model to derive an emotional theme from the user's input and provide real-time feedback to the user based on whether the emotional theme of the user's input is the same as a predetermined emotional theme.

[0071] In step 340, the processor can provide cognitive training feedback based on the user's performance of the above cognitive training.

[0072] The processor can evaluate whether the user remembers the training words provided in the cognitive training performed.

[0073] For example, the processor can acquire user input in which the user inputs a training word provided during cognitive training performed by the user, and can provide real-time feedback to the user based on whether the user's input and the training word are the same.

[0074] Based on the results of the user's cognitive training, the processor can provide the user with information on the improvement of cognitive function, information on cognitive function for the user's age group, etc.

[0075] Meanwhile, the processor may provide the user with a pre-training assessment to evaluate whether the user remembers the previous training words provided in the previous cognitive training before the user performs cognitive training.

[0076] For example, the processor can acquire user input that inputs a previous training word provided in previous cognitive training, and provide real-time feedback to the user based on whether the user's input is the same as the previous training word provided in previous cognitive training.

[0077] In addition, the processor can determine the difficulty of cognitive training based on the user's pre-training evaluation results, as described in step 310 above.

[0078]

[0079] Hereinafter, with reference to FIGS. 4 to 10, a user-customized cognitive function training method using a generative artificial intelligence model will be described in more detail.

[0080] FIG. 4 is a diagram illustrating an example of a method for determining the difficulty level of cognitive training according to one embodiment.

[0081] Hereinafter, with reference to FIG. 4, an example of a method by which a processor determines the difficulty of cognitive training is described.

[0082] Referring to FIG. 4, the processor can determine the difficulty level (430) of cognitive training based on the previous training word (410).

[0083] The processor can determine the difficulty level (430) of the cognitive training based on the results of the user's pre-training evaluation (420).

[0084] The processor can provide the user with a pre-training evaluation (420) based on the previous training words (410) provided in the previous cognitive training.

[0085] For example, the processor may provide the user with a pre-training evaluation (420) to evaluate whether they remember the previous training words (410) provided in the previous cognitive training.

[0086] For example, the processor may provide the user with a pre-training evaluation (420) to assess whether the user remembers the previous training words (410) provided in a previous cognitive training performed at a different point in time at the current time. Here, the different point in time may mean a point in time prior to the current time on the same day as the current time, or a point in time on a different day prior to the current time. Additionally, the current time may mean the time at which the current cognitive training is performed.

[0087] For example, the processor may provide the user with a pre-training assessment (420) that requires the user to input all previous training words provided in the previous cognitive training within a limited time. Additionally, the processor may obtain the user's input for the pre-training assessment (420) and determine whether the user's input is identical to the previous training words (410).

[0088] As an example, the processor may determine that the user's input is the same as the previous training word (410) if the user's input is included in the previous training word (410). That is, the processor may determine the user's input as the correct answer if the user's input is the same as the previous training word (410). As another example, the processor may determine that the user's input is not the same as the previous training word (410) if the user's input is not included in the previous training word (410). Meanwhile, the processor may determine that the user's input is not the same as the previous training word (410) even if there is no user input. That is, the processor may determine the user's input as the incorrect answer if the user's input is not the same as the previous training word (410).

[0089] The processor can provide real-time feedback to the user based on whether the user's input is the same as the previous training word (410) provided in the previous cognitive training.

[0090] For example, if the processor determines that the user's input is not the same as the previous training word (410), it can provide real-time hint feedback to the user.

[0091] More specifically, if the user's input is incorrect, the processor can provide real-time hint feedback to the user based on the content of the detailed training performed in the previous cognitive training. For example, the processor can provide real-time hint feedback by providing the user with at least one of the cognitive training content among the concentration training, association training, and association training provided to the user in the previous cognitive training.

[0092] The processor can determine the difficulty level (430) of the cognitive training based on the results of the user's pre-training evaluation (420).

[0093] For example, the processor can determine the difficulty level (430) of cognitive training based on the number of training words based on the results of the user's pre-training evaluation (420).

[0094] For example, the processor can determine the difficulty level (430) of cognitive training by the number of training words. In other words, the difficulty level (430) of cognitive training can be determined to be more difficult as the number of training words increases, and easier as the number of training words decreases. More specifically, the processor can determine the number of training words to be 2, 4, 6, or 8, and the difficulty level (430) of cognitive training can be determined accordingly, but the number of training words is merely an example for convenience of explanation and is not limited thereto.

[0095] For example, when a user performs cognitive training for the first time, the processor may determine the difficulty level (430) of the cognitive training to be the easiest, that is, the number of training words to be the smallest. Additionally, the processor may determine the difficulty level (430) of the cognitive training to be more or easier than the difficulty level of the previous cognitive training based on the results of the user's pre-training evaluation (420). In other words, the processor may determine the number of training words to be provided in the cognitive training to be more or less than the number of training words provided in the previous cognitive training based on the results of the user's pre-training evaluation (420).

[0096] For example, the processor can calculate the number of user inputs determined to be correct, i.e., the correct rate. Here, the processor can calculate only the number of user inputs determined to be correct without receiving real-time hint feedback in the pre-training evaluation (420).

[0097] Additionally, the processor can determine the difficulty (430) of the cognitive training based on the number of previous training words and the number of user inputs determined as correct answers.

[0098] For example, when the number of previous training words is less than 5, the processor may determine the difficulty of cognitive training (430) to be more difficult than the difficulty of previous cognitive training, i.e., more than the number of previous training words, when the user's correct answer rate for the pre-training evaluation (420) is 100%, the difficulty of cognitive training (430) to be easier than the difficulty of previous cognitive training, i.e., less than the number of previous training words, when the user's correct answer rate for the pre-training evaluation (420) is less than 50%, the difficulty of cognitive training (430) to be the same as the difficulty of previous cognitive training, i.e., less than the number of previous training words, and when the user's correct answer rate for the pre-training evaluation (420) is 50 to 99%, the difficulty of cognitive training (430) to be the same as the difficulty of previous cognitive training, i.e., more than the number of previous training words.

[0099] As another example, when the number of previous training words is 5 or more, the processor may determine the difficulty of cognitive training (430) to be more difficult than the difficulty of previous cognitive training, i.e., the number of training words to be greater than the number of previous training words, when the user's correct answer rate for the pre-training evaluation (420) is 80% or more, and when the user's correct answer rate for the pre-training evaluation (420) is less than 50%, the difficulty of cognitive training (430) to be easier than the difficulty of previous cognitive training, i.e., the number of training words to be less than the number of previous training words, and when the user's correct answer rate for the pre-training evaluation (420) is 50 to 79%, the difficulty of cognitive training (430) to be the same as the difficulty of previous cognitive training, i.e., the number of training words to be the same as the number of previous training words.

[0100] Therefore, the processor can determine the difficulty level (430) of the cognitive training in the manner described above.

[0101] Meanwhile, the processor can determine a training word from a word category based on the difficulty (430) of the cognitive training.

[0102]

[0103] FIG. 5 is a diagram illustrating an example of a method for determining training words according to one embodiment.

[0104] Hereinafter, with reference to FIG. 5, an example of a method by which a processor determines a training word is described.

[0105] Referring to FIG. 5, the processor can determine training words from word categories (510).

[0106] The processor can determine training words from word categories (520) based on the difficulty of cognitive training.

[0107] For example, the processor can obtain user input for selecting a word category (520) by providing the user with at least one candidate word category (510).

[0108] Here, the candidate word category (510) is provided to the user to determine a word in the field the user wants as a training word, and may include categories for various fields such as novels, poetry, plays, essays, sports, travel, etc.

[0109] Additionally, the candidate word category (510) may include a wider variety of categories, not limited to the fields of novel, poetry, play, essay, sports, and travel categories shown in FIG. 5, and may include further subcategories. That is, even if a user selects the novel category among the candidate word categories (510), the user may select various subcategories such as by country, genre, author, or era, and furthermore, may select a specific novel. In other words, the candidate word category (510) and / or word category (520) in the present disclosure may be used as words that encompass everything from broad categories such as novels and poetry to narrow categories such as specific novels and specific poems.

[0110] Therefore, the processor can determine the training words among the words of the word category (520) desired by the user.

[0111] For example, the processor can determine the training words among the words included in the word category (520) based on the difficulty of cognitive training.

[0112] For example, the processor can use a generative artificial intelligence model to extract key words from word categories (520) and determine training words based on the extracted key words.

[0113] For example, the processor can obtain input from a user who selects a word category (520) for poetry among candidate word categories (510), and among them selects a specific poem called the preface by poet Yun Dong-ju.

[0114] For example, the processor can extract key words of a specific poem selected by the user using a pre-trained generative artificial intelligence model. Here, the pre-trained generative artificial intelligence model may refer to an artificial intelligence model trained to extract key words that symbolize the word category (520) from among the words included in the word category (520), based on text, sentences, words, meanings, moods, etc. included in the word category (520).

[0115] More specifically, the processor can input a specific poem selected by the user as input data for a generative artificial intelligence model and extract key words as output data. For example, the processor can identify from the word category (520) selected by the user that the phrase contains a symbolic meaning that one must love all beings, including death and pain, and can extract "star," which symbolizes hope and beauty, and "heart," which is the fundamental emotion of love and fulfillment, as key words.

[0116] Therefore, the processor can determine a number of words among the extracted core words as training words corresponding to the difficulty of cognitive training. In this case, if the number of extracted core words is greater than the number of training words corresponding to the difficulty of cognitive training, the processor may determine the training words by considering the difficulty of the core words themselves, or it may determine the training words randomly. Here, the difficulty of the core words themselves can be determined based on the National Institute of Korean Language's word API (Application Programming Interface).

[0117] Meanwhile, the processor can provide cognitive training to the user using determined training words.

[0118] For example, the processor can use a generative artificial intelligence model to provide cognitive training to the user based on determined training words.

[0119] First, the processor can introduce the training words by presenting the determined training words to the user in the form of introductory sentences or introductory pictures.

[0120]

[0121] FIG. 6 is a diagram illustrating an example of a method for providing an introductory sentence or an introductory picture to a user according to one embodiment.

[0122] Hereinafter, with reference to FIG. 6, an example of a method in which a processor provides an introductory sentence or an introductory picture to a user is described.

[0123] Referring to FIG. 6, the processor can generate an introductory sentence (620) or an introductory picture (630) using the training word (610).

[0124] For example, the processor can use a generative artificial intelligence model to generate at least one of an introductory sentence (620) or an introductory picture (630) based on training words.

[0125] For example, the processor may input training words as input data for a generative artificial intelligence model and generate at least one of an introductory sentence (620) or an introductory picture (630) as output data.

[0126] Additionally, the processor may generate an introductory sentence (620) or an introductory picture (630) by using a prompt for generating an introductory sentence (620) or a prompt for generating an introductory picture (630) together. Here, the prompt may refer to generation guidelines such as the purpose, context, constraints, and output format for generating output data when the artificial intelligence model generates output data.

[0127] For example, the processor may generate an introduction sentence (620) as output data of the generative artificial intelligence model by applying prompts to the generative artificial intelligence model, such as “generate a sentence that is as concise as possible,” “generate one sentence if there are fewer than 5 training words, and generate two sentences if there are 5 or more training words,” “include all training words,” “use polite language,” and “do not include violent, criminal, or sexual content,” in order to generate an introduction sentence (620).

[0128] For example, to generate an introduction picture (630), the processor can generate an introduction picture (630) as output data of the generative AI model by applying prompts to the generative AI model such as “use a soft drawing style,” “use a drawing style unrelated to copyright,” “do not include violent, criminal, or sexual scenes,” “include all training words,” “do not include words other than training,” and “include only pictures and not text.”

[0129] Meanwhile, the processor may generate an introductory picture (630) using only an introductory sentence (620) as input data for a generative artificial intelligence model, in addition to training words. Thus, the processor can generate an introductory picture (630) that explains the situation of the introductory sentence (620) using the introductory sentence (620).

[0130] For example, the processor may provide the user with at least one of an introductory sentence (620) or an introductory picture (630).

[0131] In other words, the processor can present a training word to the user by providing at least one of an introductory sentence (620) or an introductory picture (630), and the user can recognize the training word through the introductory sentence (620) and / or the introductory picture (630).

[0132] Meanwhile, the processor may provide the user with an introductory sentence (620) and / or an introductory picture (630) so that after the user recognizes the training word, it may provide the user with multiple detailed trainings including intensive training, association training, and association training.

[0133] First, the processor can provide the user with intensive training among multiple detailed trainings.

[0134]

[0135] FIG. 7 is a drawing for illustrating an example of a method for providing cognitive training to a user according to one embodiment. More specifically, FIG. 7 is a drawing for illustrating an example of a method for providing concentration training to a user.

[0136] Hereinafter, with reference to FIG. 7, an example of a method in which a processor (220) provides final feedback to a user is described.

[0137] Referring to FIG. 7, the processor may provide the user with training words (710) or non-training words (720). Here, non-training words (720) may mean words that are not included in the word category selected by the user.

[0138] For example, the processor can obtain user input for classifying the training word (710) and the non-training word (720) by providing either the training word (710) or the non-training word (720) to the user.

[0139] For example, the processor may randomly provide the user with either a training word (710) or a non-training word (720). That is, the training word (710) and the non-training word (720) may be provided to the user alternately, the training word (710) may be provided to the user consecutively, and the non-training word (720) may be provided to the user consecutively.

[0140] Additionally, the processor may obtain user input to select whether the provided word is a training word (710) or a non-training word (720). That is, the user may select whether the provided word is a training word (710) or a non-training word (720).

[0141] For example, the processor can provide real-time feedback to the user based on acquired user input.

[0142] First, the processor can determine whether the user's input is correct or incorrect.

[0143] As a first example, when the provided word is a training word (710), if the user selects the provided word as the training word (710), the processor may determine the user's input as the correct answer. As a second example, when the provided word is a training word (710), if the user selects the provided word as a non-training word (720), the processor may determine the user's input as the incorrect answer. As a third example, when the provided word is a non-training word (720), if the user selects the provided word as the training word (710), the processor may determine the user's input as the incorrect answer. As a fourth example, when the provided word is a non-training word (720), if the user selects the provided word as the non-training word (720), the processor may determine the user's input as the correct answer.

[0144] Therefore, the processor can provide real-time feedback to the user based on whether the user's input is correct or incorrect.

[0145] As one example, the processor can provide positive feedback to the user if the user's input is correct. In other words, the processor can provide commendation feedback to the user. As another example, the processor can provide retry feedback to the user if the user's input is incorrect. That is, the processor can indicate that the user's input is incorrect and provide the user with an opportunity to choose again.

[0146] Meanwhile, the processor can provide additional recall training to the user.

[0147] For example, the processor can provide the user with a recall training in which all training words (710) provided in the intensive training performed by the user can be entered. Additionally, the processor can obtain input from the user in which the user enters the training words.

[0148] For example, the processor can determine whether the user's input is correct or incorrect based on whether the word entered by the user is the same as the training word (710).

[0149] As one example, the processor may determine the user's input as the correct answer if the word entered by the user is the same as the training word (710). As another example, the processor may determine the user's input as the incorrect answer if the word entered by the user is not the same as the training word (710) or if there is no input from the user within a limited time.

[0150] The processor may provide a reward to the user based on the accuracy rate of the user who performed the recall training. Here, the reward may be one of the entertainment elements of cognitive training and may serve as a motivating factor to encourage the user to perform the cognitive training. Additionally, the reward may include a maximum reward, a medium reward, and a minimum reward, and the processor may provide the user with the highest number of points (e.g., 30 points) as the maximum reward, a medium number of points (e.g., 20 points) as the medium reward, and the lowest number of points (e.g., 10 points) as the minimum reward, but the method of providing the reward is not limited thereto.

[0151] For example, the processor may provide the highest reward to the user when the user's accuracy rate for recall training is 100%, may provide the middle reward to the user when the user's accuracy rate is 50% to 99% or higher, and may provide the lowest reward to the user when the user's accuracy rate is less than 50%.

[0152] Meanwhile, the processor can provide association training to the user among multiple detailed trainings.

[0153] For example, the processor may provide the user with at least one of unit training using pictures and unit training using quizzes as association training.

[0154] In addition, the processor can provide the user with unit training using pictures and unit training using quizzes based on the number of training words.

[0155] For example, if the number of training words is 2, the processor may provide unit training using pictures using one of the 2 training words and provide unit training using quizzes to the user using the other one.

[0156] As another example, when the number of training words is 4, 6, or 8, the processor may randomly select half of the training words (i.e., 2 when the number of training words is 4, 3 when the number of training words is 6, and 4 when the number of training words is 8) and provide cognitive training to the user in the order of unit training using pictures, unit training using quizzes, and unit training using pictures using the selected half of the training words. Additionally, in the next cognitive training, the processor may provide cognitive training to the user in the order of unit training using quizzes, unit training using pictures, and unit training using quizzes using the selected half of the training words.

[0157]

[0158] FIGS. 8 and 9 are drawings illustrating different examples of a method for providing cognitive training to a user according to one embodiment. More specifically, FIG. 8 is a drawing illustrating an example of a method for providing unit training using pictures as association training, and FIG. 9 is a drawing illustrating an example of a method for providing unit training using quizzes as association training.

[0159] Hereinafter, with reference to FIGS. 8 and 9, another example of a method in which a processor provides cognitive training to a user is described.

[0160] First, referring to FIG. 8, the processor can obtain input from a user drawing a first picture (820) based on a training word (810).

[0161] For example, the processor may present a training word (810) to the user and obtain a first drawing (820) of the training word (810) drawn by the user from the user.

[0162] For example, the processor can use a generative artificial intelligence model to calculate a score based on the degree of correspondence between the training word (810) and the first picture (820).

[0163] For example, the processor can use a generative artificial intelligence model to calculate the degree of correspondence between the actual image of the training word (810) and the first image (820) obtained from the user. More specifically, the processor can input the training word (810) and the first image (820) as input data for the generative artificial intelligence model. Additionally, the processor can derive features of the actual image of the training word (810) and determine whether the derived features of the training word (810) are included in the first image (820).

[0164] Accordingly, the processor can calculate the number of features of the actual image of the training word (810) included in the first picture (820) and calculate a score based on the number of calculated features.

[0165] For example, the processor can calculate 0 points if no features of the actual image of the training word (810) are included in the first drawing (820), 1 point if 1 / 3 (1 / 3) of the features of the actual image are included in the first drawing (820), 2 points if 2 / 3 (2 / 3) of the features of the actual image are included in the first drawing (820), and 3 points if all features of the actual image are included in the first drawing (820). That is, if there are 3 features of the actual image of the training word (810), the processor can calculate 1 point if 1 feature is included in the first drawing (820) drawn by the user, 2 points if 2 features are included, and 3 points if 3 features are included.

[0166] In other words, the processor can calculate 0 points if there are letters, dots, or lines in the first picture (820); 1 point if there are some visual elements related to the training word (810) but the shape is too simple or the color selection is wrong; 2 points if the core features of the training word (810) are expressed and the colors are appropriately selected but the detailed description is lacking; and 3 points if the typical image of the training word (810) is well expressed, the colors are appropriately combined, and it is clearly recognized as the training word (810).

[0167] For example, the processor can provide real-time feedback to the user based on whether the calculated score is greater than or equal to a preset score.

[0168] As an example, the processor can provide positive feedback to the user if the calculated score is 2 points or higher. That is, the processor can provide praise feedback to the user if the calculated score is 2 points or higher.

[0169] As another example, the processor can provide real-time picture feedback to the user when the calculated score is less than 2 points.

[0170] For example, the processor can sequentially provide real-time drawing feedback to the user so that the user draws a second drawing based on the training words (810).

[0171] For example, the processor may provide drawing guidelines as real-time drawing feedback so that the user can draw a second picture based on the training word (810). More specifically, if the training word (810) is "flower," the processor may sequentially provide the user with drawing guidelines for drawing the training word (810), such as first, "draw five petals," second, "draw a stem and leaves," and third, "color the petals red and the stem and leaves green."

[0172] In addition, the processor can provide the user with not only drawing guidelines but also correction feedback that allows the user to erase or correct parts they drew incorrectly.

[0173] Meanwhile, the processor may additionally provide recall training to the user, and the method of providing recall training to the user is the same as described above with reference to FIG. 7, so redundant content is omitted.

[0174] Referring to FIG. 9, the processor can derive the correct word (930) based on a question (910) about a training word and user input (920) regarding the question (910). That is, unit training using a quiz can mean cognitive training in which the processor asks a question (910) about a training word and the user answers the question (910) so that the processor guesses the training word.

[0175] For example, the processor can obtain user input (920) for the question (910) about the training word generated using a generative artificial intelligence model by providing the user with the question (910).

[0176] For example, the processor may present training words to the user and provide the user with questions (910) about the training words.

[0177] For example, the processor can input training words as input data for a generative artificial intelligence model and generate a question (910) as output data that allows the user to think about information about the training words and imagine an image of the training words.

[0178] Accordingly, the processor can provide a question (910) about a training word to the user and obtain an answer to the question (910) from the user. Here, the user's input to the question (910) may be any one of "yes," "no," or "I don't know," but is not limited thereto, and may include an answer consisting of a specific sentence.

[0179] For example, the processor can exchange multiple questions and answers with the user in the form of twenty questions. Additionally, as the question-and-answer session continues, the processor can ask questions (910) that become closer to the training word in the form of elimination, and at this time, can ask sensory questions such as visual, auditory, tactile, olfactory, and gustatory questions. At this time, the processor can ask questions (910) that become semantically closer to the training word in the form of elimination as the question-and-answer session progresses by using a generative artificial intelligence model. That is, the processor can use the user's input (920) as input data for the generative artificial intelligence model to derive questions (910) that become semantically closer to the training word as output data, and can provide the derived questions to the user.

[0180] For example, the processor can derive a correct word (930) based on the user's input (920) for the question (910). In other words, the processor can determine the word derived from the user's input (920) for the question (910) as the correct word (930).

[0181] Additionally, the processor can derive the correct word (930) using a generative artificial intelligence model. For example, the processor can derive the correct word (930) as output data by using the user's input (920), i.e., the answer, for the question (910) provided to the user as input data for the generative artificial intelligence model.

[0182] For example, the processor can provide real-time feedback to the user based on whether the correct answer word (930) is the same as the training word.

[0183] For example, the processor can provide positive feedback to the user using a generative artificial intelligence model. In other words, the processor uses the correct word (930) as input data for the generative artificial intelligence model to determine whether the correct word (930) and the training word are the same, and depending on whether the correct word (930) and the training word are the same, it can derive praise feedback or incorrect feedback as output data.

[0184] As an example, the processor can provide positive feedback to the user when the correct word (930) is the same as the training word. That is, the processor can provide praise feedback to the user. As another example, the processor can provide incorrect feedback to the user when the correct word (930) is not the same as the training word. For example, the processor can provide the user with the fact that the reason the incorrect correct word (930) was derived is due to a specific input among the inputs (920) of multiple users, and can provide feedback that the user's specific input was an incorrect input because it differed from the training word.

[0185] Meanwhile, the processor may additionally provide recall training to the user, and the method of providing recall training to the user is the same as described above with reference to FIG. 7, so redundant content is omitted.

[0186]

[0187] FIG. 10 is a drawing for illustrating another example of a method for providing cognitive training to a user according to one embodiment. More specifically, FIG. 10 is a drawing for illustrating one example of a method for providing associated training to a user.

[0188] Hereinafter, with reference to FIG. 10, another example of a method in which a processor provides cognitive training to a user is described.

[0189] Referring to Fig. 10, the processor can provide the user with a story picture (1010).

[0190] For example, the processor can use a generative artificial intelligence model to generate a story picture (1010) based on training words and a predetermined emotion theme (1020).

[0191] For example, the processor can randomly determine one of multiple emotional themes (1020), such as "fear," "anger," "joy," "sadness," "disgust," and "surprise."

[0192] Additionally, the processor may input training words and a determined emotion theme (1020) as input data for a generative artificial intelligence model, and generate a story picture (1010) containing both the training words and the emotion theme (1020) as output data. That is, the story picture (1010) may include all training words and may include expressions for the emotion theme (1020).

[0193] For example, the processor can provide a story picture (1010) to the user and obtain user input that inputs a sentence based on the story picture (1010).

[0194] For example, the processor may provide the user with a story picture (1010) and requests such as, "Please create a story that clearly shows the emotions in the story picture," or "Please create a story by including all the words trained today." Additionally, the processor may obtain a story about the story picture (1010) from the user.

[0195] For example, the processor can provide real-time feedback to the user based on whether the user's input corresponds to a training word or a predetermined emotion theme (1020).

[0196] For example, the processor can input a user's input of a story picture (1010) as input data for a generative artificial intelligence model, and derive an evaluation of the story input by the user as output data.

[0197] For example, the processor can determine whether the user's input to the story picture (1010) contains all the training words and whether the story contains a specific emotional theme (1020).

[0198] As an example, the processor can provide real-time feedback to the user based on whether the user's input to the story picture (1010) contains all the training words.

[0199] For example, the processor can determine whether there are missing training words in the user's input.

[0200] For example, if there are no missing training words in the user's input, the processor can provide positive feedback to the user. On the other hand, if there are missing training words in the user's input, the processor can provide real-time feedback to the user. That is, the processor can calculate the number of missing training words, provide this number to the user, and have the user recreate the story. Furthermore, if training words are still missing in the recreated story, the processor can directly provide the missing training words to the user to help them recognize them.

[0201] As another example, the processor can provide real-time feedback to the user based on whether the user's input to the story picture (1010) includes a story about a specific emotional theme (1020).

[0202] For example, the processor can determine whether the user's input includes a story about a specific emotional theme (1020) by using the user's input for a story picture (1010) as input data for a generative artificial intelligence model. Additionally, the processor can provide real-time feedback to the user based on the result of determining whether the user's input includes a story about a specific emotional theme (1020). That is, the processor can derive an emotional theme included in the user's input using a generative artificial intelligence model, determine whether the derived emotional theme is the same as a specific emotional theme (1020), and provide real-time feedback to the user based on whether the derived emotional theme is the same as a specific emotional theme (1020).

[0203] Additionally, the processor can derive an emotional theme of the user's input using a generative artificial intelligence model. That is, the processor inputs the user's input regarding the story picture (1010) as input data for the generative artificial intelligence model, and derives the emotional theme included in the user's input as output data. For example, the processor can provide real-time feedback to the user based on whether the emotional theme of the user's input is the same as a predetermined emotional theme (1020).

[0204] For example, if the emotional theme (1020) of the story picture (1010) is "surprise" and the emotional theme of the user's input to the story picture (1010) is "sadness," the processor may determine that the emotional theme of the user's input and the predetermined emotional theme (1020) are not the same. Additionally, the processor may provide the user with the fact that the emotional theme expressed by the user is "sadness," but the emotional theme (1020) of the actual story picture (1010) is "surprise," thereby allowing the user to recreate the story. Furthermore, if the emotional theme of the recreated story is also not the same as the emotional theme (1020) of the story picture (1010), the processor may provide the user with the emotional theme (1020) of the story picture (1010) once again to allow the user to recognize the emotional theme (1020).

[0205] The processor can provide rewards to the user based on the input of the user who performed the federated training.

[0206] For example, the processor may provide the highest reward to the user when all training words are included in the user's input for the story picture (1010) and the emotion theme (1020) is also the same, may provide the middle reward to the user when more than 70% of the training words are included in the user's input but the emotion theme (1020) is not the same, and may provide the lowest reward to the user when less than 70% of the training words are included in the user's input and the emotion theme (1020) is not the same.

[0207] Meanwhile, the processor can provide cognitive training feedback based on the results of the user's cognitive training performance.

[0208] For example, as described above with reference to FIGS. 7 to 10, the processor may provide the user with intensive training, association training, and association training, and provide cognitive training feedback based on the results of the user performing the intensive training, association training, and association training.

[0209] For example, the processor can provide the user with information on the improvement of cognitive function, information on cognitive function for the user's age group, etc., based on the results of the user's cognitive training performance.

[0210] In other words, the processor can provide the user with information regarding the improvement of their cognitive function over a specified period by visually representing it. Additionally, the processor can provide the user with information on the cognitive function of people of a similar age group, enabling the user to recognize their own cognitive state.

[0211] Meanwhile, the processor can provide a cognitive training assessment that enables the recall of all training words provided in the cognitive training.

[0212] For example, the processor may provide a cognitive training assessment to the user, which requires the user to input all training words provided in the cognitive training within a limited time. Additionally, the processor may obtain the user's input for the cognitive training assessment and determine whether the user's input is identical to the training words provided in the cognitive training.

[0213] As one example, the processor can determine that the user's input is identical to a training word provided in cognitive training if the user's input is included among the training words provided in cognitive training. In other words, the processor can determine the user's input as the correct answer if it is identical to a training word provided in cognitive training. As another example, the processor can determine that the user's input is not identical to a training word provided in cognitive training if the user's input is not included among the training words provided in cognitive training. Meanwhile, the processor can determine that the user's input is not identical to a training word provided in cognitive training even in the absence of user input. In other words, the processor can determine the user's input as an incorrect answer if it is not identical to a training word provided in cognitive training.

[0214] The processor can provide cognitive training feedback to the user based on whether the user's input is identical to the training words provided in the cognitive training.

[0215] For example, if the processor determines that the user's input is not identical to the training words provided in cognitive training, it can provide cognitive training feedback to the user.

[0216] More specifically, if the user's input is incorrect, the processor may provide cognitive training feedback to the user based on the content of the detailed training performed in cognitive training. For example, the processor may provide cognitive training feedback by providing the user with at least one of the cognitive training content among the concentration training, association training, and association training provided to the user in cognitive training.

[0217]

[0218] In the method described above in the present disclosure, a device providing cognitive training to a user can communicate in real time and provide cognitive training to the user.

[0219] In addition, the user can receive real-time feedback from the cognitive training device while performing cognitive training, and can also receive final cognitive training feedback after performing the cognitive training.

[0220] In addition, by providing users with cognitive training tailored to their cognitive abilities, their cognitive functions can be effectively improved.

[0221] In addition, by using generative AI models to analyze emotional words and sentence contexts included in the user's input, accurate feedback on the user's input can be provided.

[0222] Meanwhile, the above-described method can be written as a program executable on a computer and can be implemented on a general-purpose digital computer that operates the program using a computer-readable recording medium. In addition, the structure of the data used in the above-described method can be recorded on a computer-readable recording medium through various means. The computer-readable recording medium includes storage media such as magnetic storage media (e.g., ROM, RAM, USB, floppy disk, hard disk, etc.) and optical reading media (e.g., CD-ROM, DVD, etc.).

[0223] A person skilled in the art related to the present embodiment will understand that it may be implemented in modified forms without departing from the essential characteristics of the description above. Therefore, the disclosed methods should be considered in an illustrative rather than a restrictive sense, and the scope of rights is defined in the claims rather than the description above, and should be interpreted to include all differences within the scope of equivalence.

Claims

1. A step of determining the difficulty level of cognitive training based on the user's pre-training evaluation results; A step of determining training words from word categories based on the difficulty level of the above cognitive training; A step of providing the cognitive training based on the determined training words to the user using a generative artificial intelligence model; and A step of providing cognitive training feedback based on the results of the cognitive training performed by the user; comprising User-customized cognitive function training method using generative artificial intelligence models.

2. In Paragraph 1, The step of determining the difficulty level of the above cognitive training is, A step of providing the user with the pre-training evaluation based on the previous training words provided in the previous cognitive training; and A method comprising the step of determining the difficulty of the cognitive training according to the number of training words based on the user's evaluation results prior to training.

3. In Paragraph 1, The step of determining the above training words is, A step of obtaining input from the user selecting at least one of the candidate word categories; and A method comprising the step of determining the training word from the word category based on the difficulty of the cognitive training.

4. In Paragraph 1, The step of providing the above cognitive training to the user is, A step of generating at least one of an introductory sentence or an introductory picture based on the training words using the generative artificial intelligence model; and A method comprising the step of providing at least one of the above introductory sentence or the above introductory picture to the user.

5. In Paragraph 1, The step of providing the above cognitive training to the user is, A step of obtaining input from the user for classifying the training word and the non-training word by providing one of the above training word and the non-training word to the user; and A method comprising the step of providing real-time feedback to the user based on the input of the user.

6. In Paragraph 1, The step of providing the above cognitive training to the user is, A step of obtaining input from the user who draws a first picture based on the above training word; A step of calculating a score based on the degree of correspondence between the training word and the first picture using the generative artificial intelligence model; and A method comprising the step of providing real-time feedback to the user based on whether the score is greater than or equal to a preset score.

7. In Paragraph 6, If the above score is less than the preset score, The step of providing the above real-time feedback is, A method comprising the step of sequentially providing real-time drawing feedback to the user so that the user draws a second drawing based on the training words.

8. In Paragraph 1, The step of providing the above cognitive training to the user is, A step of providing a question regarding the training word generated using the generative artificial intelligence model to the user to obtain the user's input regarding the question; A step of deriving the correct answer word based on the user's input regarding the above question; and A method comprising the step of providing real-time feedback to the user based on whether the correct answer word is the same as the training word.

9. In Paragraph 1, The step of providing the above cognitive training to the user is, A step of generating a story picture based on the training word and a predetermined emotion theme using the above generative artificial intelligence model; A step of obtaining input from the user who inputs a sentence based on the story picture by providing the story picture to the user; and A method comprising the step of providing real-time feedback to the user based on whether the user's input corresponds to the training word or the predetermined emotion theme.

10. In Paragraph 9, The step of providing real-time feedback to the above user is, A step of determining whether there are missing training words in the input of the user; and A method comprising the step of providing real-time feedback to the user based on the above-mentioned missing training words.

11. In Paragraph 9, The step of providing real-time feedback to the above user is, A step of deriving an emotional theme of the user's input using the generative artificial intelligence model; and A method comprising the step of providing real-time feedback to the user based on whether the emotional theme of the user's input is the same as the predetermined emotional theme.

12. A computer-readable recording medium storing a program for executing the method of claim 1 on a computer.

13. Memory in which at least one program is stored; and It includes at least one processor that executes the above at least one program, and The above-mentioned at least one processor is, Determining the difficulty of cognitive training based on the user's pre-training evaluation results, determining training words from word categories based on the difficulty of the cognitive training, providing the cognitive training based on the determined training words to the user using a generative artificial intelligence model, and providing cognitive training feedback based on the user's performance results of the cognitive training. User-customized cognitive function training device utilizing a generative artificial intelligence model.

Images