Artificial intelligence device and method for operating same

Abstract

An artificial intelligence device according to one embodiment of the present disclosure may comprise: a memory for storing a brain-mimicking artificial intelligence model trained through reinforcement learning; a mental health meter for collecting subject data including a numerical value of memory recall confidence, memory recall accuracy, a numerical value of inference confidence, inference accuracy, learning accuracy, and strategic decision-making bias according to the execution of a meta-memory game of a user; and a processor which acquires a plurality of cognitive behavior numerical values from the subject data by using the brain-mimicking artificial intelligence model, acquires a plurality of brain function estimation signals corresponding to the plurality of cognitive behavior numerical values, respectively, and maps each brain function estimation signal to a brain signal corresponding to a specific brain function.

Description

Artificial intelligence device and its operating method

[0001] The present invention relates to an artificial intelligence device, and more particularly, to an artificial intelligence device capable of comprehensively measuring a user's mental health profile.

[0002] Digital healthcare refers to the application of digital technologies in the medical and health fields to manage patient health, prevent disease, diagnose, and treat it. Digital healthcare utilizes a variety of technologies, including information technology, artificial intelligence, sensor technology, and big data.

[0003] As interest in the digital healthcare industry grows, technologies are being proposed to improve the accuracy of mental health monitoring.

[0004] Prior art patent 1 (Korean Patent Publication No. 10-2022-0085863) related to this can superficially confirm the presence or absence of mental health risks from the speech sentence pattern, but has the problem of making it difficult to clearly observe the basis of cognitive decline (e.g., decreased memory recall) that causes the risk.

[0005] In addition, prior art patent 2 (Korean Patent Publication No. 10-2023-0045625) has a problem in that it is difficult to predict the risk of mental illness in which metacognition is the main impairment because it is impossible to measure metacognition (e.g., recall confidence assessment system), which is a potential variable in mental health.

[0006] Prior patent 3 (Korean Patent Publication No. 10-2020-0092457) developed a cognitive explanation model that limited human decision-making to two aspects: learning and control. However, a real high-level decision-making system comprehensively involves learning, memory-inference, and metacognition, which examines each function. Therefore, in situations with higher complexity, Prior patent 3's ability to explain the human cognitive system of prior art is significantly reduced.

[0007] The purpose of the present disclosure may be to estimate a comprehensive mental health profile of learning-memory-inference from a user's game data using a mental health measuring device using a metamemory game and a brain-mimetic artificial intelligence model.

[0008] The purpose of the present disclosure may be to increase the accuracy of monitoring the mental health profile of individual users.

[0009] The purpose of the present disclosure may be to verify the reproduction of a user's brain function pattern using an optimized brain-mimicking artificial intelligence model, and to provide individual brain function values ​​corresponding to each cognitive behavior.

[0010] An artificial intelligence device according to one embodiment of the present disclosure may include a memory that stores a brain-mimicking artificial intelligence model learned through reinforcement learning, a mental health measuring device that collects subject data including a value of memory recall confidence, a value of memory recall accuracy, a value of inference confidence, inference accuracy, learning accuracy, and strategic decision-making bias according to a user's performance of a metamemory game, and a processor that obtains a plurality of cognitive behavioral values ​​from the subject data using the brain-mimicking artificial intelligence model, obtains a plurality of brain function estimation signals corresponding to each of the plurality of cognitive behavioral values, and maps each brain function estimation signal to a brain signal corresponding to a specific brain function.

[0011] According to an embodiment of the present disclosure, the following effects are achieved.

[0012] First, by providing an individual user's mental health profile based on data analysis results from a mental health measurement device using a brain-mimicking artificial intelligence model, it is possible to predict mental health risks such as decreased learning performance, memory recall failure, confidence bias, and decision-making strategy bias.

[0013] Second, by optimizing the parameters of the brain-mimicking artificial intelligence model to maximize the ability to explain users' strategic decision-making patterns, the accuracy of monitoring individual users' mental health profiles can be improved.

[0014] Third, by verifying the reproduction of the user's brain function patterns using an optimized brain-mimicking artificial intelligence model, it is possible to provide individual brain function values ​​corresponding to each cognitive behavior.

[0015] Fourth, by designing a brain-mimetic artificial intelligence model based on cognitive theory, it can be applied to the field of digital healthcare to interpret the impaired functions underlying cognitive impairment, and can help clinical experts establish cognitive treatment strategies for dementia / mild cognitive impairment based on the provided interpretation.

[0016] Fifth, the mental health monitoring technology of the present invention can be installed in a care robot to provide a mental health management service platform in places with insufficient mental health medical personnel and accessibility, such as hospitals, nursing homes, and silver towns.

[0017] FIG. 1 is a block diagram illustrating components of an artificial intelligence device according to an embodiment of the present disclosure.

[0018] FIG. 2 is a diagram for explaining the configuration of an artificial intelligence server according to one embodiment of the present disclosure.

[0019] FIG. 3 is a diagram for explaining the configuration of an artificial intelligence system according to one embodiment of the present disclosure.

[0020] FIG. 4a and FIG. 4b are diagrams illustrating a preliminary learning process according to an embodiment of the present disclosure, and FIG. 4c is a diagram illustrating a process for obtaining recall confidence, recall accuracy, inference confidence, and inference accuracy according to performance of a metamemory game according to an embodiment of the present disclosure.

[0021] FIG. 5 is a diagram illustrating a process of optimizing parameters of a brain-simulating AI model according to an embodiment of the present disclosure.

[0022] FIGS. 6A and 6B are diagrams showing a process of optimizing parameters of a brain-mimetic AI model using subject learning data of a metamemory game according to an embodiment of the present disclosure.

[0023] FIG. 7 is a diagram illustrating a process for verifying the reproducibility of brain function of a brain-simulating AI model according to an embodiment of the present disclosure.

[0024] FIG. 8a is a drawing illustrating improved explanatory power of a brain-mimicking AI model according to an embodiment of the present disclosure, and FIG. 8b is a drawing illustrating reproducibility of a brain-mimicking AI model according to an embodiment of the present disclosure.

[0025] FIG. 9 is a diagram illustrating verification of brain function reproduction and brain function estimation signals of a brain-mimetic artificial intelligence model according to an embodiment of the present disclosure.

[0026] FIG. 10 is a flowchart illustrating an operation method of an artificial intelligence system according to an embodiment of the present disclosure.

[0027] FIG. 11 is a drawing illustrating the configuration of an artificial intelligence device according to another embodiment of the present disclosure.

[0028] Artificial intelligence refers to a field that studies artificial intelligence or the methodologies for creating it, and machine learning refers to a field that defines various problems in the field of artificial intelligence and studies the methodologies for solving them.

[0029] Machine learning is sometimes defined as an algorithm that improves its performance on a task through continuous experience.

[0030] An artificial neural network (ANN) is a model used in machine learning. It can refer to a model with problem-solving capabilities that is composed of artificial neurons (nodes) that form a network through the combination of synapses.

[0031] An artificial neural network can be defined by the connection patterns between neurons in different layers, the learning process that updates model parameters, and the activation function that generates the output values.

[0032] An artificial neural network may include an input layer, an output layer, and optionally one or more hidden layers. Each layer contains one or more neurons, and the artificial neural network may include synapses connecting neurons. In an artificial neural network, each neuron can output a function value of an activation function based on input signals, weights, and biases received through the synapses.

[0033] Model parameters are parameters determined through learning, including synaptic connection weights and neuron biases. Hyperparameters are parameters that must be set before learning in machine learning algorithms, including the learning rate, number of iterations, mini-batch size, and initialization function.

[0034] The goal of artificial neural network training can be seen as determining model parameters that minimize a loss function. The loss function can be used as an indicator for determining optimal model parameters during the artificial neural network training process.

[0035] Machine learning can be classified into supervised learning, unsupervised learning, and reinforcement learning depending on the learning method.

[0036] Supervised learning refers to a method of training an artificial neural network given labels for training data. The labels can refer to the correct answer (or result value) that the artificial neural network must infer when training data is input to the artificial neural network.

[0037] Unsupervised learning can refer to a method of training an artificial neural network without being given labels for the training data.

[0038] Reinforcement learning can refer to a learning method that teaches an agent defined in an environment to select an action or action sequence that maximizes the cumulative reward in each state.

[0039] Among artificial neural networks, machine learning implemented with a deep neural network (DNN) that includes multiple hidden layers is also called deep learning, and deep learning is a part of machine learning.

[0040] Hereinafter, machine learning is used to mean including deep learning.

[0041] FIG. 1 is a block diagram illustrating components of an artificial intelligence device according to an embodiment of the present disclosure.

[0042] The artificial intelligence device (100) can be implemented as a fixed device or a movable device, such as a TV, a projector, a mobile phone, a smart phone, a desktop computer, a laptop, a digital broadcasting terminal, a PDA (personal digital assistant), a PMP (portable multimedia player), a navigation device, a tablet PC, a wearable device, a set-top box (STB), a DMB receiver, a radio, a washing machine, a refrigerator, a desktop computer, digital signage, a robot, a vehicle, etc.

[0043] Referring to FIG. 1, an artificial intelligence device (100) may include a communication interface (110), an input interface (120), a learning processor (130), a sensor (140), an output interface (150), a memory (170), and a processor (180).

[0044] The communication interface (110) can transmit and receive data with external devices such as other artificial intelligence devices or AI servers (200) using wired or wireless communication technology. For example, the communication interface (110) can transmit and receive sensor information, user input, learning models, control signals, etc. with external devices.

[0045] The communication technologies used by the communication interface (110) include GSM (Global System for Mobile communication), CDMA (Code Division Multi Access), LTE (Long Term Evolution), 5G, WLAN (Wireless LAN), Wi-Fi (Wireless-Fidelity), Bluetooth (Bluetooth), RFID (Radio Frequency Identification), Infrared Data Association (IrDA), ZigBee, NFC (Near Field Communication), etc.

[0046] The input interface (120) can obtain various types of data.

[0047] The input interface (120) may include a camera (121) for capturing images, a microphone (122) for receiving audio signals, and a user input interface (123) for receiving information from a user.

[0048] By treating the camera (121) or microphone (122) as a sensor, the signal obtained from the camera (121) or microphone (122) can be called sensing data or sensor information.

[0049] The input interface (120) can acquire input data to be used when obtaining output using learning data and a learning model for model learning. The input interface (120) can also acquire raw input data, in which case the processor (180) or learning processor (130) can extract input features as preprocessing for the input data.

[0050] The camera (121) processes image frames, such as still images or moving images, obtained by the image sensor in video call mode or shooting mode. The processed image frames can be displayed on the display (151) or stored in the memory (170).

[0051] The microphone (122) processes external acoustic signals into electrical voice data. The processed voice data can be utilized in various ways depending on the function (or application program) being performed by the artificial intelligence device (100). Meanwhile, various noise removal algorithms can be applied to the microphone (122) to remove noise generated during the process of receiving external acoustic signals.

[0052] The user input interface (123) is for receiving information from a user. When information is input through the user input interface (123), the processor (180) can control the operation of the artificial intelligence device (100) to correspond to the input information.

[0053] The user input interface (123) may include a mechanical input means (or a mechanical key, for example, a button located on the front / rear or side of the artificial intelligence device (100), a dome switch, a jog wheel, a jog switch, etc.) and a touch input means.

[0054] As an example, the touch input means may be composed of virtual keys, soft keys, or visual keys displayed on a touch screen through software processing, or may be composed of touch keys placed on a part other than the touch screen.

[0055] The learning processor (130) can train a model composed of an artificial neural network using learning data. The trained artificial neural network can be referred to as a learning model. The learning model can be used to infer result values ​​for new input data other than the learning data, and the inferred values ​​can be used as a basis for judgment to perform a certain action.

[0056] The running processor (130) can perform AI processing together with the running processor (240) of the AI ​​server (200).

[0057] The running processor (130) may include a memory integrated or implemented in the artificial intelligence device (100). The running processor (130) may also be implemented using a memory (170), an external memory directly coupled to the artificial intelligence device (100), or a memory maintained in an external device.

[0058] The sensor (140) can obtain at least one of internal information of the artificial intelligence device (100), information about the surrounding environment of the artificial intelligence device (100), and user information by using various sensors.

[0059] The sensor (140) may include one or more of a proximity sensor, a light sensor, an acceleration sensor, a magnetic sensor, a gyro sensor, an inertial sensor, an RGB sensor, an IR sensor, a fingerprint recognition sensor, an ultrasonic sensor, a light sensor, a microphone, a lidar sensor, and a radar sensor.

[0060] The output interface (150) can generate output related to visual, auditory, or tactile sensations.

[0061] The output interface (150) may include a display (151) that outputs images, an audio output interface (152) that outputs audio, a haptic device (153) that outputs tactile information, and a light output interface (154) that outputs light.

[0062] The display (151) displays (outputs) information processed in the artificial intelligence device (100). For example, the display (151) may display execution screen information of an application program running in the artificial intelligence device (100), or UI (User Interface) or GUI (Graphical User Interface) information according to such execution screen information.

[0063] The display (151) can be implemented as a touch screen by forming a mutual layer structure with the touch sensor or forming an integral structure. The touch screen can function as a user input interface (123) that provides an input interface between the artificial intelligence device (100) and the user, and at the same time, provide an output interface between the artificial intelligence device (100) and the user.

[0064] The audio output interface (152) can output audio data received from the communication interface (110) or stored in the memory (170) in a call signal reception mode, call mode, recording mode, voice recognition mode, broadcast reception mode, etc.

[0065] The audio output interface (152) may include at least one of a receiver, a speaker, and a buzzer.

[0066] The haptic device (153) generates various tactile effects that can be felt by the user. A representative example of the tactile effect generated by the haptic device (153) may be vibration.

[0067] The light output interface (154) outputs a signal to notify the occurrence of an event using light from a light source of the artificial intelligence device (100). Examples of events occurring in the artificial intelligence device (100) may include message reception, call signal reception, missed call, alarm, schedule notification, email reception, and information reception through an application.

[0068] The memory (170) can store data that supports various functions of the artificial intelligence device (100). For example, the memory (170) can store input data, learning data, learning models, learning history, etc. obtained from the input interface (120).

[0069] The processor (180) can determine at least one executable operation of the artificial intelligence device (100) based on information determined or generated using a data analysis algorithm or a machine learning algorithm.

[0070] The processor (180) can control components of the artificial intelligence device (100) to perform determined operations.

[0071] To this end, the processor (180) can request, retrieve, receive or utilize data from the running processor (130) or memory (170), and control components of the artificial intelligence device (100) to execute at least one of the executable operations, a predicted operation or an operation determined to be desirable.

[0072] When the processor (180) requires connection to an external device to perform a determined operation, it can generate a control signal for controlling the external device and transmit the generated control signal to the external device.

[0073] The processor (180) can obtain intent information for user input and determine the user's requirements based on the obtained intent information.

[0074] The processor (180) can obtain intent information corresponding to the user input by using at least one of a STT (Speech To Text) engine for converting voice input into a string or a natural language processing (NLP) engine for obtaining intent information of natural language.

[0075] At least one of the STT engine or the NLP engine may be configured with an artificial neural network, at least in part, trained according to a machine learning algorithm. Furthermore, at least one of the STT engine or the NLP engine may be trained by the learning processor (130), the learning processor (240) of the AI ​​server (200), or through distributed processing thereof.

[0076] The processor (180) can collect history information including the operation details of the artificial intelligence device (100) or the user's feedback on the operation, and store the information in the memory (170) or the learning processor (130), or transmit the information to an external device such as an AI server (200). The collected history information can be used to update the learning model.

[0077] The processor (180) can control at least some of the components of the artificial intelligence device (100) to run an application program stored in the memory (170).

[0078] The processor (180) can operate two or more of the components included in the artificial intelligence device (100) in combination to drive the application program.

[0079] FIG. 2 is a diagram for explaining the configuration of an artificial intelligence server according to one embodiment of the present disclosure.

[0080] Referring to FIG. 2, the AI ​​server (200) may refer to a device that trains an artificial neural network using a machine learning algorithm or uses a trained artificial neural network.

[0081] The AI ​​server (200) may be composed of multiple servers to perform distributed processing, and may be defined as a 5G network. The AI ​​server (200) may be included as part of the artificial intelligence device (100) and may perform at least a portion of the AI ​​processing.

[0082] The AI ​​server (200) may include a communication interface (210), memory (230), a learning processor (240), and a processor (260).

[0083] The communication interface (210) can transmit and receive data with an external device such as an artificial intelligence device (100).

[0084] The memory (230) may include a model memory (231). The model memory (231) may store a model (or artificial neural network, 231a) being learned or learned through the learning processor (240).

[0085] A learning processor (240) can train an artificial neural network (231a) using learning data. The learning model can be used while mounted on the AI ​​server (200) of the artificial neural network, or can be mounted on an external device such as an artificial intelligence device (100).

[0086] The learning model may be implemented in hardware, software, or a combination of hardware and software. If part or all of the learning model is implemented in software, one or more instructions constituting the learning model may be stored in memory (230).

[0087] The processor (260) can use a learning model to infer a result value for new input data and generate a response or control command based on the inferred result value.

[0088] FIG. 3 is a diagram for explaining the configuration of an artificial intelligence system according to one embodiment of the present disclosure.

[0089] Referring to FIG. 3, an artificial intelligence system (30) according to one embodiment of the present disclosure may include an AI device (100), a database (300), and an AI server (200).

[0090] The AI ​​device (100) may include a preliminary learning unit (181) and a meta-memory data collector (183).

[0091] The pre-learning unit (181) and the meta-memory data collector (183) may be included in the processor (180) of FIG. 1, or may be separate components from the processor (180).

[0092] The pre-learning device (181), the meta-memory data collector (183) and the database (300) may be included in the named device of the mental health measuring device.

[0093] The pre-learning unit (181) can learn the prior knowledge required for causal inference and memory association. The data collected through the pre-learning unit (181) can serve as background knowledge required for the user (or subject) to perform the metamemory game.

[0094] The preliminary learning unit (181) may include a causal inference learning module (181a) and a memory association learning module (181b).

[0095] The causal inference learning module (181a) allows users to move to the next stage by selecting left or right and check the results according to their selection. The causal inference learning module (181a) can pre-learn the state-action-result rules required for causal inference.

[0096] The memory association learning module (181b) can strengthen the user's recall associations with specific cues. For example, the memory association learning module (181b) can learn the relationship between each animal (cue) and a food acquisition location.

[0097] The metamemory data collector (183) can induce users to make metamemory-based strategic decisions under three environmental conditions after pre-learning through the pre-learning unit (181). The three environmental conditions may include goal specificity, cue clarity, and environmental uncertainty.

[0098] The metamemory data collector (183) can measure memory recall confidence by receiving user input in the first stage of the metamemory game.

[0099] The metamemory data collector (183) can measure inference confidence by receiving user input just before confirming the result in the final stage of the metamemory game.

[0100] The metamemory data collector (183) can measure inference accuracy, memory recall accuracy, and learning accuracy based on feedback presented as a result of the user's decision-making behavior.

[0101] The metamemory data collector (183) can measure strategic decision-making bias from the difference in user behavior patterns between the current trial and the previous trial.

[0102] The metamemory data collector (183) can transmit memory recall confidence, memory recall accuracy, inference confidence, inference accuracy, learning accuracy, and strategic decision-making bias to a database (300) or an AI server (200).

[0103] The database (300) can receive and store user data collected from the meta-memory data collector (183).

[0104] The database (300) may be the memory (230) of the AI ​​server (200) or a separate component from the AI ​​server (200).

[0105] The AI ​​server (200) may include a brain-mimetic AI model (400) and a mental health profile output device (500).

[0106] The brain-mimetic AI model (400) may include a reinforcement learning module (410), a confidence bias detection module (420), a confidence-based memory recall module (430), a strategy bias detection module (440), and a model fitting device (450).

[0107] The reinforcement learning module (410) can produce learning and behavior pattern indices that evaluate the subject's learning and behavior patterns by using the subject's learning data and parameters through reinforcement learning.

[0108] The confidence bias detection module (420) can produce an inference confidence pattern index that evaluates the subject's learning and behavioral patterns using the subject's learning data and parameters.

[0109] The confidence-based memory recall module (430) can produce a memory confidence pattern index using subject learning data and parameters.

[0110] The strategy bias detection module (440) can produce a strategy modification pattern index using subject learning data and parameters.

[0111] The model fitting device (450) can optimize the parameters of each of the reinforcement learning module (410), the confidence bias detection module (420), the confidence-based memory recall module (430), and the strategy bias detection module (440).

[0112] The brain-mimetic AI model (400) can output multiple cognitive behavioral values ​​from subject data received from the AI ​​device (100) after the parameters are optimized.

[0113] The mental health profile output device (500) may include a cognitive behavioral value output module (510) and a brain function value output module (520).

[0114] The cognitive behavioral numerical output module (510) can transmit multiple cognitive behavioral numerical values ​​output from the brain-mimetic AI model (400) to the brain function numerical output module (520).

[0115] The brain function numerical output module (520) can modulate each cognitive behavioral numerical value through parametric modulation.

[0116] The brain function value output module (520) can output a brain function estimation signal (521) by inputting the modified cognitive behavior value into a hemodynamic function.

[0117] FIG. 4a and FIG. 4b are diagrams illustrating a preliminary learning process according to an embodiment of the present disclosure, and FIG. 4c is a diagram illustrating a process for obtaining recall confidence, recall accuracy, inference confidence, and inference accuracy according to performance of a metamemory game according to an embodiment of the present disclosure.

[0118] Metamemory games are games that prompt users to make metamemory-based strategic decisions under three environmental conditions: goal specificity, cue clarity, and uncertainty.

[0119] Referring to FIG. 4a, the causal inference learning module (181a) can display an execution screen according to the execution of the causal inference learning game.

[0120] The causal inference learning module (181a) can provide the following process according to the execution of the causal inference learning game.

[0121] The causal inference learning module (181a) can display a first image (401) and a first box (401a) on the display (151) in stage 1. The first image (401) may be an image randomly displayed among a plurality of images.

[0122] The boxes described below are related to goal specificity and can be colored blue, red, yellow, or white. Blue can be assigned 10 coins, red 20 coins, yellow 40 coins, and white 0 coins.

[0123] In the first box (401a), one of blue, red, yellow, and white colors can be displayed randomly.

[0124] The causal inference learning module (181a) can transition from stage 1 to stage 2 after a certain period of time has elapsed, depending on user input of selecting the left or right button of the mouse.

[0125] The causal inference learning module (181a) may display a second image (402) and a second box (401b) or a third image (403) and a third box (401c) when the user selects the left button in Stage 1. To increase uncertainty, the probability of displaying the second image (402) may be 0.7, and the probability of displaying the third image (403) may be 0.3. The probabilities are not disclosed to the user.

[0126] The causal inference learning module (181a) may display a fourth image (404) and a fourth box (401d) or a fifth image (405) and a fifth box (401e) when the user selects the right button in stage 1. The probability that the fourth image (404) will be displayed may be 0.7, and the probability that the fifth image (405) will be displayed may be 0.3.

[0127] The causal inference learning module (181a) can transition the state from stage 2 to stage 3 when the left button or right button is selected in stage 2.

[0128] For example, the causal inference learning module (181a) may display the sixth image (406) or the seventh image (407) when the left button is selected in stage 2 while the fourth image (404) is being displayed. The probability that the sixth image (406) will be displayed may be 0.7, and the probability that the seventh image (407) will be displayed may be 0.3.

[0129] The causal inference learning module (181a) can output a notification indicating that 20 coins are given as a reward when the 6th image (406) is displayed in stage 3.

[0130] The causal inference learning module (181a) can output a notification indicating that 40 coins are given as a reward when the 7th image (407) is displayed in stage 3.

[0131] Users repeat the state transition process from Stage 1 to Stage 3 to receive as many coins as possible.

[0132] The above process can be performed multiple times for one user, and can be performed multiple times for each of multiple users.

[0133] The causal inference learning module (181a) can provide coin reward results according to stage movement and selection.

[0134] The causal inference learning module (181a) allows the user to move to the next stage by selecting left or right and check the results according to the selection, and can learn the state-action-result rules required for causal inference in advance.

[0135] In particular, the causal inference learning module (181a) can induce users to have reasoning confidence in making optimal judgments to receive a large reward.

[0136] Referring to FIG. 4b, the AI ​​device (100) can display an execution screen according to the execution of the memory association learning game.

[0137] The memory association learning module (181b) may provide the following process according to the execution of the memory association learning game. In the memory association learning game, unlike the causal inference learning game, at least one of the first clue (or cue, 411) or the second clue (412) may be additionally displayed in Stage 1.

[0138] The first clue (411) may be an animal image representing an animal, and the second clue (412) may be a place image representing a place where the animal obtains food.

[0139] The memory association learning module (181b) can display the first image (401), the first box (401a), and either the first clue (411) or the second clue (412) on the display (151) in stage 1.

[0140] The memory association learning module (181b) can transition the state from stage 1 to stage 2 according to the selection of the right button and display the fourth image (404) and the fourth box (401d).

[0141] The memory association learning module (181b) can display the final image by transitioning the state from stage 2 to stage 3 when the left button is selected. The memory association learning module (181b) can output a coin acquisition notification indicating that 40 coins and a bonus coin (50 coins) are awarded when the color of the displayed image matches the color of the first box (401a) and the displayed image matches the first clue (411) or the second clue (412).

[0142] Users can see how much coins they will be rewarded for following this path.

[0143] The memory association learning module (181b) allows the user to remember a path indicating an animal's food acquisition location, thereby strengthening a specific cue-recall association.

[0144] In particular, the memory association learning module (181b) can induce the user to have confidence in memory recall regarding which choice to make when given a clue to maximize the coin reward.

[0145] Figure 4c is a diagram illustrating the process by which a metamemory data collector (183) acquires memory recall confidence, memory recall accuracy, inference confidence, and inference accuracy according to the execution of a metamemory game.

[0146] The metamemory data collector (183) can display a first input window (460) for inputting memory recall confidence before entering stage 1 of the metamemory game, a first clue (411) indicating an animal, and a second clue (412) indicating a location for obtaining food for the animal.

[0147] The first input window (460) can be used to input memory recall confidence, which indicates how well the path to the target (food acquisition location) can be found. The metamemory data collector (183) can obtain a memory recall confidence value through the first input window (460). The memory recall confidence value can be any one of 0 to 10, but is merely an example.

[0148] After that, the meta-memory data collector (183) can display a first image (401) randomly selected from among a plurality of images and a first box (401a) having a random color.

[0149] After that, the meta-memory data collector (183) can transition the state from stage 1 to stage 2 when either the left button or the right button is selected according to the user action.

[0150] After that, the meta-memory data collector (183) can display the fourth image (404) and the fourth box (401d).

[0151] When either the left button or the right button is selected according to the user action, the metamemory data collector (183) can display a second input window (470) for inputting inference confidence before confirming the result of stage 3 of the metamemory game.

[0152] The second input window (470) may be a window for inputting inference confidence, which indicates how well rewarded one will be in this round. For example, inference confidence may be confidence in predicting whether food will be obtained.

[0153] The metamemory data collector (183) can obtain a numerical value of inference confidence through the second input window (470). The numerical value of inference confidence can be any one of 0 to 10, but is merely an example. The second input window (470) may further include a reward prediction input item for inputting the expected amount of reward.

[0154] The metamemory data collector (183) can obtain memory recall accuracy, learning accuracy, and inference accuracy based on the first clue (411), the second clue (412) presented during execution of the metagame, and the subject's selection collected when transitioning from stage 1 to stage 2.

[0155] The metamemory data collector (183) can measure learning accuracy based on feedback presented as a result of the user's decision-making behavior.

[0156] The metamemory data collector (183) can measure strategic decision-making bias based on the difference in the user's behavioral pattern between the current trial and the previous trial.

[0157] The metamemory data collector (183) can obtain subject learning data including the values ​​of memory recall confidence, memory recall accuracy, inference confidence, inference accuracy, learning accuracy, and strategic decision-making bias.

[0158] After optimizing the parameters of the brain-mimetic AI model, the metamemory data collector (183) can obtain subject data including the values ​​of memory recall confidence, memory recall accuracy, inference confidence, inference accuracy, learning accuracy, and strategic decision-making bias.

[0159] FIG. 5 is a diagram illustrating a process of optimizing parameters of a brain-simulating AI model according to an embodiment of the present disclosure.

[0160] The brain-mimetic AI model (400) may include a reinforcement learning module (410), a confidence bias detection module (420), a confidence-based memory recall module (430), a strategy bias detection module (440), and a model fitting device (450).

[0161] The brain-mimetic AI model (400) may be included in any one of the processor (260), the learning processor (240), or the model memory (231a) of the AI ​​server (200).

[0162] The brain-mimetic AI model (400) can determine model parameters using the Maximum Likelihood Estimation (MLE) method. The Maximum Likelihood Estimation method can be a technique for selecting parameters that maximize the likelihood of the occurrence of given data.

[0163] Since the subject learning data of the brain-mimetic AI model (400) may be different for each user, the values ​​of the model parameters may also be optimized differently.

[0164] Each of the reinforcement learning module (410), the confidence bias detection module (420), the confidence-based memory recall module (430), and the strategy bias detection module (440) uses reinforcement learning and may use the maximum likelihood method for parameter optimization.

[0165] The reinforcement learning module (410) can produce learning and behavior pattern indices that evaluate the subject's learning and behavior patterns by using the subject's learning data and parameters through reinforcement learning.

[0166] The parameters of the reinforcement learning module (410) may include a reinforcement learning learning rate, a softmax activation function temperature, a transition function slope, and a continuation action degree.

[0167] Learning and behavior pattern metrics may be indicators of how well the reinforcement learning module (410) describes actual subject data. Learning and behavior pattern metrics may include learned behavior values, recalled behavior values, and learning-memory integrated behavior values.

[0168] A smaller numerical value of the learning and behavior pattern indicator may indicate that the reinforcement learning module (410) better explains the subject data.

[0169] The learning and behavior pattern indicators can be output to a model fitting device (450). The model fitting device (450) can change the values ​​of the parameters in a direction that minimizes the values ​​of the learning and behavior pattern indicators, and can transmit the changed values ​​of the parameters to the reinforcement learning module (410).

[0170] The model fitting device (450) can instruct the reinforcement learning model (410) to produce learning and behavior pattern indicators using the values ​​of the changed parameters. When the values ​​of the new learning and behavior pattern indicators converge to a minimum value, the model fitting device (450) can transmit the output of the values ​​of the parameters corresponding to the minimum value to the mental health profile measuring device (500).

[0171] The model fitting device (450) can change the values ​​of the parameters until the values ​​of the new learning and behavior pattern indicators converge to the minimum values ​​if the values ​​of the new learning and behavior pattern indicators do not converge to the minimum values.

[0172] The confidence bias detection module (420) can use subject learning data and parameters to generate an inference confidence pattern index that evaluates the subject's learning and behavioral patterns. The inference confidence pattern index can include an inference confidence bias.

[0173] The parameters for learning the confidence bias detection module (420) may include an action value recall-induced inference confidence bias threshold.

[0174] The inference confidence pattern metric may be an indicator of how well the confidence bias detection module (420) explains the subject data.

[0175] A smaller numerical value of the inference confidence pattern index may indicate that the confidence bias detection module (420) better explains the subject data.

[0176] The inference confidence pattern index can be output to the model fitting device (450). The model fitting device (450) can change the values ​​of the parameters in a direction in which the values ​​of the inference confidence pattern index are minimized, and can transmit the values ​​of the changed parameters to the confidence bias detection module (420).

[0177] The model fitting device (450) can instruct the confidence bias detection module (420) to produce an inference confidence pattern index using the changed parameter values. When the value of the new inference confidence pattern index converges to a minimum value, the model fitting device (450) can transmit the output of the parameter value corresponding to the minimum value to the mental health profile measuring device (500).

[0178] If the value of the new inference confidence pattern index does not converge to the minimum value, the model fitting device (450) can change the value of the parameter until the value of the new inference confidence pattern index converges to the minimum value.

[0179] The confidence-based memory recall module (430) can calculate a memory confidence pattern index using subject learning data and parameters. The memory confidence pattern index can include a recall confidence bias.

[0180] Parameters for learning of the confidence-based memory recall module (430) may include an action value recall-induced recall confidence threshold.

[0181] The memory confidence pattern index may be an index indicating how well the confidence-based memory recall module (430) explains the subject data.

[0182] A smaller value of the memory confidence pattern index may indicate that the confidence-based memory recall module (430) better explains the subject data.

[0183] The memory confidence pattern index can be output to a model fitting device (450). The model fitting device (450) can change the values ​​of the parameters in a direction that minimizes the value of the memory confidence pattern index, and can transmit the changed parameter values ​​to the confidence-based memory recall module (430).

[0184] The model fitting device (450) can instruct the confidence-based memory recall module (430) to produce a memory confidence pattern index using the changed parameter values. When the value of the new memory confidence pattern index converges to a minimum value, the model fitting device (450) can transmit the output of the parameter value corresponding to the minimum value to the mental health profile measuring device (500).

[0185] The model fitting device (450) can change the values ​​of the parameters until the values ​​of the new memory confidence pattern indicator converge to the minimum value if the values ​​of the new memory confidence pattern indicator do not converge to the minimum value.

[0186] The strategy bias detection module (440) can derive a strategy modification pattern index using subject learning data and parameters. The strategy modification pattern index may include reward prediction error, state prediction error, prediction reliability, and behavioral strategy correction power.

[0187] Parameters for learning the strategy bias detection module (440) may include a state prediction error threshold and a reward prediction error threshold.

[0188] The strategy modification pattern indicator may be an indicator of how well the strategy bias detection module (440) explains the subject data.

[0189] A smaller numerical value of the strategy modification pattern indicator may indicate that the strategy bias detection module (440) better explains the subject data.

[0190] The strategy modification pattern indicator can be output to the model fitting device (450). The model fitting device (450) can change the values ​​of the parameters in a direction in which the values ​​of the strategy modification pattern indicator are minimized, and can transmit the values ​​of the changed parameters to the strategy bias detection module (440).

[0191] The model fitting device (450) can instruct the strategy bias detection module (440) to produce a strategy modification pattern index using the changed parameter values. When the value of the new strategy modification pattern index converges to a minimum value, the model fitting device (450) can transmit the output of the parameter value corresponding to the minimum value to the mental health profile measuring device (500).

[0192] If the value of the new strategy modification pattern indicator does not converge to the minimum value, the model fitting device (450) can change the value of the parameter until the value of the new strategy modification pattern indicator converges to the minimum value.

[0193] In this way, the model fitting device (450) can optimize the values ​​of parameters of each module included in the brain-simulating AI model (400).

[0194] FIGS. 6A and 6B are diagrams showing a process of optimizing parameters of a brain-mimetic AI model using subject learning data of a metamemory game according to an embodiment of the present disclosure.

[0195] Referring to FIG. 6a, the first sampling result (610) obtained according to the execution of the metagame of user number 20 and the second sampling result (630) obtained according to the execution of the metagame of user number 40 are illustrated.

[0196] The first sampling result (610) may include a first subject learning data set (611) and a first brain signal estimation data set (613).

[0197] The first subject learning data set (611) may be a set of subject learning data obtained from each of 20 user trials. The first brain signal estimation data set (613) may be data on brain signals estimated from each of 20 trials.

[0198] Data on brain signals may include values ​​of indicators output by each module of Fig. 5.

[0199] The second sampling result (630) may include a second subject learning data set (631) and a second brain signal estimation data set (633).

[0200] The P1 model parameter is the state prediction error threshold, the P2 model parameter is the reward prediction error threshold, the P3 model parameter is the action-value recall-induced inference confidence bias threshold, the P4 model parameter is the action-value recall-induced recall confidence threshold, the P5 model parameter is the softmax activation function temperature, the P6 model parameter is the reinforcement learning learning rate, the P7 model parameter is the thought transition function slope, and the P8 model parameter is the degree of continuation behavior.

[0201] Referring to Figure 6a, the parameter values ​​of the brain-mimicking AI model corresponding to user 20 may be different from the parameter values ​​of the brain-mimicking AI model corresponding to user 40. In other words, the values ​​of model parameters optimized for different users may vary.

[0202] Referring to Figure 6b, subject learning data measured using the metamemory game is illustrated. Subject learning data may include memory recall confidence, memory recall accuracy, inference confidence, inference ability, learning accuracy, and strategic decision-making bias.

[0203] Figure 6b shows that changes in mental health can be sensitively measured depending on environmental conditions of cue clarity and goal specificity.

[0204] Cue clarity can be categorized as clear or unclear. Goal specificity can be categorized as specific or flexible.

[0205] The model fitting device (450) can search for an optimal combination of multiple model parameters through the process of FIG. 5.

[0206] A brain-mimetic AI model (400) with optimized model parameters can output a set of cognitive behavioral values ​​including reward prediction error, state prediction error, prediction reliability, learned behavioral value, recalled behavioral value, learning-memory integrated behavioral value, inference confidence bias, recall confidence bias, and behavioral strategy correction power through a cognitive behavioral value output module (510).

[0207] FIG. 7 is a diagram illustrating a process for verifying the reproducibility of brain function of a brain-simulating AI model according to an embodiment of the present disclosure.

[0208] Referring to FIG. 7, the brain-mimetic AI model (400) can transmit a cognitive behavioral score set (710) including cognitive behavioral scores to a cognitive behavioral score output module (510). The cognitive behavioral score output module (510) can output the cognitive behavioral score set (710) to a brain function score output module (520).

[0209] The brain function numerical output module (520) can modulate each cognitive behavioral numerical value through parametric modulation.

[0210] The brain function value output module (520) can output a brain function estimation signal (521) by inputting the modified cognitive behavior value into a hemodynamic function.

[0211] The brain function estimation signal (521) can be input into a brain function mapping GLM (General Linear Model) regression function and mapped with a brain signal (523) measured through an actual brain electrode.

[0212] The processor (260) of the AI ​​server (200) can input the brain function estimation signal (521) into the brain function mapping GLM regression function and map it with the previously measured brain function signal (523).

[0213] The memory (230) may store data on a plurality of previously measured brain function signals.

[0214] FIG. 8a is a drawing illustrating improved explanatory power of a brain-simulating AI model according to an embodiment of the present disclosure, and FIG. 8b is a drawing illustrating reproducibility of a brain-simulating AI model according to an embodiment of the present disclosure.

[0215] Referring to FIG. 8a, the brain-simulating AI model (model 9) according to an embodiment of the present disclosure is structured to recall strategic action values ​​by considering inference / recall confidence and utilize this for learning strategy coordination (see FIG. 5), and significantly and significantly explains human strategic decision-making compared to other candidate models (using the 'negative logarithm of explainability' as an input value of the Bayesian Inference Criterion [BIC], which is a model comparison value, the lower the value, the better the explanatory power).

[0216] In particular, the prediction accuracy of the brain-mimetic artificial intelligence model of the present disclosure is 72%, which is an 18% improvement in prediction accuracy compared to the prior art model (prior patent 3, Republic of Korea Patent Publication No. 10-2020-0092457).

[0217] Figure 8b is an example of a reproducibility verification for diagnosing an overfitting problem of a brain-mimetic artificial intelligence model.

[0218] When the model was retrained with behavioral data reproduced by the brain-inspired artificial intelligence model, a significant correlation was confirmed between the initial and reproduced parameters for all eight model parameters (parameters 1 to 8) (p < 0.001).

[0219] FIG. 9 is a diagram illustrating verification of brain function reproduction and brain function estimation signals of a brain-mimetic artificial intelligence model according to an embodiment of the present disclosure.

[0220] While the user is playing the metamemory game, the measured functional MRI (Magnetic Resonance Imaging) brain function activity (brain function data) and the brain function estimation signal output by the brain function numerical output module (520) can be mapped to each cognitive behavioral variable in the nodes of the brain's frontal lobe-basal ganglia-hippocampus network.

[0221] Each cognitive behavioral measure can be mapped to a brain signal with predictive activity for a specific brain function.

[0222] The brain function numerical output module (520) can output a brain function estimation signal representing the brain function mapped to each cognitive behavioral variable (e.g., memory recall confidence, reasoning confidence).

[0223] The first brain function estimation data (911) for the hippocampus or the peri-hippocampal region (915) of the brain related to memory recall confidence can be mapped to the measured first brain function data (913).

[0224] The second brain function estimated data (921) for the ventromedial prefrontal cortex or inferior frontal gyrus (925) of the brain associated with reasoning confidence can be mapped to the measured second brain function data (923).

[0225] FIG. 10 is a flowchart illustrating an operation method of an artificial intelligence system according to an embodiment of the present disclosure.

[0226] The processor (260) of the AI ​​server (200) can obtain subject learning data (S1001).

[0227] The processor (260) can receive subject learning data from the metamemory data collector (183) of the AI ​​device (100). The subject learning data can include data obtained through repeated performance of the preliminary learner (181) and data obtained through repeated performance of the metamemory game.

[0228] Subject learning data may include recall confidence, recall accuracy, inference confidence, inference accuracy, learning accuracy, and strategic decision bias.

[0229] The processor (260) of the AI ​​server (200) can obtain optimal values ​​of a plurality of parameters for the brain-simulating AI model (400) based on subject learning data (S1003).

[0230] The brain-mimicking AI model (400) may be a model based on an artificial neural network trained through deep learning or machine learning. The brain-mimicking AI model (400) may be a model that outputs multiple cognitive behavioral values ​​from subject data through an artificial neural network.

[0231] Multiple cognitive behavioral measures may include reward prediction error, state prediction error, prediction reliability, learned action value, recalled action value, learning-memory integrated action value, inference confidence bias, recall confidence bias, and behavioral strategy correction power.

[0232] Each of the plurality of model parameters constituting the brain-mimetic AI model (400) may have a value optimized by the embodiment of FIG. 3.

[0233] The processor (260) of the AI ​​server (200) can obtain subject data (S1005) and output multiple cognitive behavioral values ​​from the subject data using a brain-simulating AI model (400) (S1007).

[0234] The processor (260) can receive subject data according to the performance of the metamemory game from the metamemory data collector (183) of the AI ​​device (100). The subject data can include data obtained through repeated performance of the metamemory game.

[0235] The processor (260) can input subject data into a brain-simulating AI model (400) with optimized parameter values ​​to obtain multiple cognitive behavioral values.

[0236] Each of the multiple cognitive behavioral measures can be an estimated data describing a specific brain function.

[0237] The processor (260) of the AI ​​server (200) can obtain a brain function estimation signal corresponding to each of a plurality of cognitive behavioral values ​​(S1009).

[0238] The processor (260) can obtain a brain function estimation signal from each cognitive behavioral value through the brain function value output module (520).

[0239] The processor (260) of the AI ​​server (200) can map each brain function estimation signal to a measured brain signal (S1011).

[0240] The memory (230) may store brain signals measured from each of a plurality of brain functions (or brain regions). The stored brain signals may be signals measured using functional branch resonance imaging (FMRI) technology. Functional branch resonance imaging (FMRI) technology is a non-invasive neuroimaging technology for measuring brain activity. It may be a technology that measures blood flow and metabolic activity in the brain and outputs signals indicating how a specific region of the brain responds when performing a specific task.

[0241] The processor (260) can map the acquired brain function estimation signal to any one of the brain signals stored in the memory (230).

[0242] As brain function estimation signals are mapped to measured brain signals, it can be determined whether a specific brain function of the subject is functioning normally.

[0243] Additionally, by providing an individual user's mental health profile based on the analysis of brain function estimation signals, which are the output of a brain-mimetic AI model, it is possible to predict mental health risks such as decreased learning performance, memory recall failure, confidence bias, and decision-making strategy bias.

[0244] FIG. 11 is a drawing illustrating the configuration of an artificial intelligence device according to another embodiment of the present disclosure.

[0245] Each step described in the embodiment of FIG. 10 can be performed by the AI ​​device (100-1) of FIG. 10.

[0246] The AI ​​device (100-1) may include a mental health measuring device (1010), memory (170), mental health profile output device (500), and processor (180).

[0247] The mental health measuring device (1010) may include a pre-learning device (181) and a metamemory data collector (183).

[0248] The functions of the preliminary learning device (181) and the meta-memory data collector (183) are replaced with the embodiments of FIG. 3 and FIG. 4a to FIG. 4c.

[0249] The mental health measuring device (1010) can transmit subject learning data including measured memory recall confidence, memory recall accuracy, reasoning confidence, reasoning accuracy, learning accuracy, and strategic decision-making bias to the brain-mimetic AI model (400).

[0250] The processor (180) may include a brain-mimetic AI model (400).

[0251] The brain-mimetic AI model (400) may include a reinforcement learning module (410), a confidence bias detection module (420), a confidence-based memory recall module (430), a strategy bias detection module (440), and a model fitting device (450).

[0252] The brain-mimetic AI model (400) can optimize the values ​​of multiple parameters from subject learning data. The process of optimizing the values ​​of multiple parameters is replaced with the example of FIG. 5.

[0253] It may include a cognitive behavioral numerical output module (510) and a brain function numerical output module (520).

[0254] The cognitive behavioral numerical output module (510) can transmit multiple cognitive behavioral numerical values ​​output from the brain-mimetic AI model (400) to the brain function numerical output module (520).

[0255] The brain function numerical output module (520) can modulate each cognitive behavioral numerical value through parametric modulation.

[0256] The brain function value output module (520) can output a brain function estimation signal (521) by inputting the modified cognitive behavior value into a hemodynamic function.

[0257] The cognitive behavioral numerical output module (510) may be included in the processor (180).

[0258] The processor (180) can control the overall operation of the AI ​​device (100-1).

[0259] The processor (180) of the AI ​​device (100-1) can obtain subject data and output multiple cognitive behavioral values ​​from the subject data using a brain-mimetic AI model (400).

[0260] The processor (180) of the AI ​​device (100-1) can receive subject data according to the performance of the metamemory game from the metamemory data collector (183). The subject data can include data obtained through repeated performance of the metamemory game.

[0261] The processor (180) of the AI ​​device (100-1) can input subject data into a brain-simulating AI model (400) with optimized parameter values ​​to obtain multiple cognitive behavioral values.

[0262] Each of the multiple cognitive behavioral measures can be an estimated data describing a specific brain function.

[0263] The processor (180) of the AI ​​device (100-1) can obtain a brain function estimation signal corresponding to each of a plurality of cognitive behavioral values.

[0264] The processor (180) of the AI ​​device (100-1) can obtain a brain function estimation signal from each cognitive behavioral value through the brain function value output module (520).

[0265] The processor (180) of the AI ​​device (100-1) can map each brain function estimation signal to a measured brain signal.

[0266] The memory (170) may store brain signals measured from each of a plurality of brain functions (or brain regions). The stored brain signals may be signals measured using functional branch resonance imaging (FMRI) technology.

[0267] The processor (180) of the AI ​​device (100-1) can map the acquired brain function estimation signal to any one of the brain signals stored in the memory (170).

[0268] As brain function estimation signals are mapped to measured brain signals, it can be determined whether a specific brain function of the subject is functioning normally.

[0269] The above-described present disclosure can be implemented as computer-readable code on a program-recorded medium. The computer-readable medium includes all types of recording devices that store data that can be read by a computer system. Examples of computer-readable media include hard disk drives (HDDs), solid-state disk drives (SSDs), silicon disk drives (SDDs), read-only memory (ROM), random-access memory (RAM), CD-ROMs, magnetic tapes, floppy disks, and optical data storage devices. In addition, the computer may include a processor (180) of an artificial intelligence device.

Claims

1. In artificial intelligence devices, Memory that stores brain-inspired artificial intelligence models learned through reinforcement learning; A mental health measure that collects subject data including memory recall confidence, memory recall accuracy, reasoning confidence, reasoning accuracy, learning accuracy, and strategic decision-making bias based on the user's performance in the metamemory game; and A processor that obtains a plurality of cognitive behavioral values ​​from subject data using the above brain-simulating artificial intelligence model, obtains a plurality of brain function estimation signals corresponding to each of the plurality of cognitive behavioral values, and maps each brain function estimation signal to a brain signal corresponding to a specific brain function. Artificial intelligence devices.

2. In paragraph 1, The above processor Optimize the parameters of the brain-inspired artificial intelligence model using the Maximum Likelihood Estimation (MLE) method, The above model parameters are Including reinforcement learning learning rate, softmax activation function temperature, thought transition function slope and continuation behavior degree, action value recall-induced inference confidence bias threshold, action value recall-induced recall confidence threshold, state prediction error threshold and reward prediction error threshold. Artificial intelligence devices.

3. In paragraph 2, The above brain-mimetic artificial intelligence model A reinforcement learning module having the above reinforcement learning learning rate, the softmax activation function temperature, the thought transition function slope, and the degree of continuous behavior as model parameters, and outputting learning and behavior pattern indicators from subject learning data. A confidence bias detection module having the above-mentioned behavioral value recall-induced inference confidence bias threshold as a model parameter and outputting an inference confidence pattern index from the subject learning data, A confidence-based memory recall module that has the above-mentioned behavioral value recall-induced recall confidence threshold as a model parameter and outputs a memory confidence pattern index from the subject learning data. A strategy bias detection module having the above-mentioned state prediction error limit and the above-mentioned compensation prediction error limit as model parameters and outputting a strategy modification pattern indicator; and A model fitting device comprising a model fitting device that optimizes each of the above model parameters. Artificial intelligence devices.

4. In paragraph 3, The above model suitable device is The values ​​of each model parameter are changed in a direction in which the values ​​of the above learning and action pattern indicator, the above inference confidence pattern indicator, the above memory confidence pattern indicator, and the above strategy modification pattern indicator are minimized. Artificial intelligence devices.

5. In paragraph 4, The above processor Modulating each of the above multiple cognitive behavioral values ​​through parametric modulation, Modulated cognitive behavioral scores are used to generate brain function estimation signals using hemodynamic functions, Mapping the generated brain function estimation signal to the brain signal having predicted activity. Artificial intelligence devices.

6. In paragraph 1, The above mental health measuring instrument A pre-learning stage that learns the prior knowledge required for causal inference and memory association. After pre-learning through the above-mentioned preliminary learning device, a meta-memory data collector is included that collects the subject data according to the performance of the meta-memory game. Artificial intelligence devices.

7. In paragraph 2, The above processor Optimize the values ​​of the above model parameters differently for each user. Artificial intelligence devices.

8. A method for operating an artificial intelligence device including a memory that stores a brain-mimetic artificial intelligence model learned through reinforcement learning, A step of collecting subject data including the level of memory recall confidence, memory recall accuracy, level of inference confidence, inference accuracy, learning accuracy and strategic decision-making bias according to the user's performance of the metamemory game; A step of obtaining multiple cognitive behavioral values ​​from subject data using the above brain-mimetic artificial intelligence model; A step of acquiring a plurality of brain function estimation signals corresponding to each of the plurality of cognitive behavioral figures; and A step of mapping each brain function estimation signal to a brain signal corresponding to a specific brain function. How artificial intelligence devices work.

9. In paragraph 8, Further comprising a step of optimizing model parameters of the brain-mimetic artificial intelligence model using Maximum Likelihood Estimation (MLE), The above model parameters are Including reinforcement learning learning rate, softmax activation function temperature, thought transition function slope and continuation behavior degree, action value recall-induced inference confidence bias threshold, action value recall-induced recall confidence threshold, state prediction error threshold and reward prediction error threshold. How artificial intelligence devices work.

10. In paragraph 9, The above brain-mimetic artificial intelligence model A reinforcement learning module having the above reinforcement learning learning rate, the softmax activation function temperature, the thought transition function slope, and the degree of continuous behavior as model parameters, and outputting learning and behavior pattern indicators from subject learning data. A confidence bias detection module having the above-mentioned behavioral value recall-induced inference confidence bias threshold as a model parameter and outputting an inference confidence pattern index from the subject learning data, A confidence-based memory recall module that has the above-mentioned behavioral value recall-induced recall confidence threshold as a model parameter and outputs a memory confidence pattern index from the subject learning data. A strategy bias detection module having the above-mentioned state prediction error limit and the above-mentioned compensation prediction error limit as model parameters and outputting a strategy modification pattern indicator; and A model fitting device comprising a model fitting device that optimizes each of the above model parameters. How artificial intelligence devices work.

11. In paragraph 10, The above optimization steps are The above model fitting device includes a step of changing the value of each model parameter in a direction in which the value of each of the learning and behavior pattern indicator, the inference confidence pattern indicator, the memory confidence pattern indicator, and the strategy modification pattern indicator is minimized. How artificial intelligence devices work.

12. In paragraph 11, The step of acquiring the above multiple brain function estimation signals is A step of modulating each of the above multiple cognitive behavioral values ​​through parametric modulation; and A step of generating a brain function estimation signal using a hemodynamic function of a modified cognitive behavioral score. How artificial intelligence devices work.

13. In paragraph 8, The steps for collecting the above subject data are Steps to learn prior knowledge required for causal inference and memory association; After learning the above prior knowledge, a step of collecting the subject data according to the performance of the metamemory game is included. How artificial intelligence devices work.

14. In paragraph 9, The values ​​of the above model parameters are optimized differently for each user. How artificial intelligence devices work.

Images