A cognitive potential assessment method and system based on reinforcement learning computational model

Abstract

The present invention discloses a cognitive potential assessment method and system based on a reinforcement learning computing model. The method comprises the following steps: obtaining a cognitive assessment task combination based on a preset task library; obtaining the user's cognitive assessment data for each cognitive assessment task, wherein the cognitive assessment data at least includes the user's specific options for completing the cognitive assessment task and the score result; based on the reinforcement learning computing model, respectively calculating the learning rate of each cognitive assessment task; converting the learning rate of each cognitive assessment task into a norm score; weighted averaging the norm scores of each cognitive assessment task to obtain the user's comprehensive assessment score, wherein different comprehensive assessment scores correspond to different cognitive potentials. The present invention can conduct learning modeling on the user's performance in a certain cognitive field through the reinforcement learning computing model, and effectively evaluate the user's cognitive potential in a certain cognitive field.

Description

Technical Field

[0001] The present invention relates to a cognitive potential assessment method based on a reinforcement learning calculation model, and also relates to a corresponding cognitive potential assessment system, belonging to the technical field of cognitive assessment. Background Art

[0002] Cognition refers to the ability of the human brain to convert external information or things into knowledge through processing and handling by internal activities after exposure. This ability involves memory, comprehension, calculation, language ability, visual-spatial ability, and judgment and comprehension. The normal functioning of cognitive functions is essential for an individual's daily life and social activities. However, when these functions are impaired, cognitive dysfunction will occur, affecting the individual's life and social ability. Without prevention and intervention, cognitive dysfunction may gradually worsen, from mild cognitive impairment to more severe cognitive impairment. Therefore, early cognitive assessment is of great significance for the timely detection and treatment of cognitive impairment.

[0003] Cognitive potential refers to the ability of an individual that has not been fully developed or utilized in a specific cognitive field. This potential can be stimulated and improved through appropriate assessment and training, thereby improving the individual's performance in related cognitive tasks. Traditional assessment methods are mainly based on the accuracy of patients' answers to questions. Although this method can reflect the patient's understanding of current knowledge or information, it is difficult to fully reflect the patient's cognitive potential in a certain cognitive field. The limitations of traditional assessment methods are mainly: 1) the patient's understanding and knowledge of the questions; 2) the patient's familiarity with the task. These factors limit the application value of the assessment results in real life.

[0004] In order to more comprehensively evaluate and understand an individual's cognitive ability, it is necessary to consider the important dimension of learning ability. Learning ability is not only an important aspect of cognitive function, but also a key way for cognitive function to play a role in real life. By describing an individual's learning ability in a certain cognitive dimension, their cognitive ability can be more objectively characterized. Therefore, how to evaluate the user's cognitive potential, especially the cognitive potential in practical applications, is still a problem that needs to be solved urgently. Summary of the invention

[0005] The primary technical problem to be solved by the present invention is to provide a cognitive potential assessment method based on a reinforcement learning computational model.

[0006] Another technical problem to be solved by the present invention is to provide a cognitive potential assessment system based on a reinforcement learning computational model.

[0007] In order to achieve the above technical objectives, the present invention adopts the following technical solutions:

[0008] According to a first aspect of an embodiment of the present invention, a method for evaluating cognitive potential based on a reinforcement learning computing model is provided, comprising the following steps:

[0009] Acquire a cognitive assessment task combination based on a preset task library; wherein the preset task library includes a plurality of subtask libraries, each of the subtask libraries corresponds to a cognitive domain, each of the subtask libraries includes a plurality of cognitive assessment tasks of the same type, each of the cognitive assessment tasks includes a plurality of options, and the scoring probabilities of the options in the same cognitive assessment task are different;

[0010] Based on the cognitive assessment task combination, obtaining the user's cognitive assessment data for each cognitive assessment task; wherein the cognitive assessment data at least includes the user's specific options for completing the cognitive assessment task and the score result;

[0011] Based on the pre-trained reinforcement learning computational model, the learning rate α of each cognitive assessment task is calculated separately;

[0012] Based on the preset norms of each cognitive assessment task, the learning rate α of each cognitive assessment task is converted into a norm score; wherein the reinforcement learning calculation model has multiple preset norms built in, and each preset norm corresponds to a cognitive assessment task in the preset task library one by one;

[0013] The norm scores of each cognitive assessment task are weighted averaged to obtain a comprehensive assessment score of the user, wherein different comprehensive assessment scores correspond to different cognitive potentials.

[0014] Preferably, the learning rate α of each cognitive assessment task is calculated as follows:

[0015] Establish the score fluctuation function of the kth option of the current cognitive assessment task to obtain the expected value of the implicit potential of the kth option in the tth task trial ;

[0016] Wherein, the score fluctuation function is expressed as: , represents the actual score of the kth option in the current cognitive assessment task in the tth task trial, represents the expected value of the implicit potential of the kth option in the tth task trial, It represents the deviation caused by the natural fluctuation of option probability, which is determined by the standard deviation of the trial of this option in the whole population;

[0017] Based on the actual score of the kth option of the user's current cognitive assessment task in the t+1th task trial , calculate the user's learning rate α for the current cognitive assessment task;

[0018]

[0019]

[0020] in, represents the actual score of the kth option in the t+1th task trial, represents the actual score of the kth option in the tth task trial, represents the expected error that the user will make for the kth option in the tth task trial.

[0021] Preferably, the cognitive potential assessment method further comprises:

[0022] Based on the softmax function, the probability that the user chooses option k in the tth task trial It is expressed as:

[0023]

[0024] in, Determined by the expected difference of all options; is the strategy parameter of the current cognitive assessment task, which determines the random volatility of the user's choice. Indicates that the user makes a completely random selection. The larger the value, the more the user follows the value to learn; i represents the specific option, i∈(1,k);

[0025] The strategy parameters of the user's current cognitive assessment task Compared with the preset threshold, if the policy parameter is not less than the preset threshold, the learning rate α of the user for the current cognitive assessment task is calculated normally; if the strategy parameter If the score is less than the preset threshold, the cognitive assessment will be repeated.

[0026] Preferably, the preset task library includes a first subtask library, a second subtask library and a third subtask library;

[0027] In the first subtask library, each cognitive assessment task requires the user to find a specific spatial position, so as to evaluate the user's spatial orientation learning ability; wherein, in each cognitive assessment task of the first subtask library, different spatial positions have different scoring probabilities;

[0028] In the second subtask library, each cognitive assessment task requires the user to select a specific object to evaluate the user's objective regularity learning ability; wherein, in each cognitive assessment task of the second subtask library, different objects have different scoring probabilities;

[0029] In the third subtask library, each cognitive assessment task requires the selection of a specific virtual character to evaluate the user's character attribute recognition ability; wherein, in each cognitive assessment task of the third subtask library, different characters have different friendliness attributes.

[0030] Preferably, the cognitive potential assessment method further comprises:

[0031] Based on the norm scores of each cognitive assessment task, averaging the norm scores of the cognitive assessment tasks belonging to the first subtask library to obtain a first assessment score corresponding to the spatial orientation learning ability;

[0032] Based on the norm scores of each cognitive assessment task, averaging the norm scores of the cognitive assessment tasks belonging to the second subtask library to obtain a second assessment score corresponding to the objective regularity learning ability;

[0033] Based on the norm scores of each cognitive assessment task, the norm scores of the cognitive assessment tasks belonging to the third subtask library are averaged to obtain a third assessment score corresponding to the character attribute recognition ability.

[0034] Preferably, the cognitive assessment task combination is obtained in the following manner:

[0035] Acquire the user's symptom information, select different cognitive domains based on the user's symptom information, and then select a corresponding number of cognitive assessment tasks from a corresponding subtask library to form a cognitive assessment task combination;

[0036] Alternatively, a number of tasks is determined in advance, and then the number of cognitive assessment tasks is extracted from each subtask library to form a cognitive assessment task combination;

[0037] Alternatively, the user's assessment requirements are obtained, and based on the assessment requirements, the cognitive domains that need to be assessed are determined, and then a corresponding number of cognitive assessment tasks are selected from a corresponding subtask library to form a cognitive assessment task combination.

[0038] Preferably, the weighted weight of the normative scores of each cognitive assessment task is proportional to the number of tasks of the same type.

[0039] Preferably, the preset norms corresponding to the cognitive assessment tasks include standard scores corresponding to people of different age groups and educational levels.

[0040] According to a second aspect of an embodiment of the present invention, a cognitive potential assessment system based on a reinforcement learning computational model is provided, comprising a processor and a memory, wherein the processor reads a computer program in the memory to execute the above-mentioned cognitive potential assessment method.

[0041] Compared with the prior art, the present invention has the following technical effects:

[0042] 1. Through the reinforcement learning computational model, the user's performance in a certain cognitive field is modeled to evaluate the user's cognitive potential in a certain cognitive field.

[0043] 2. It covers cognitive assessment tasks in space, objects, and society, and can evaluate the user's cognitive potential in different cognitive areas of basic visual space, perception, and complex social cognition, and can obtain the user's comprehensive cognitive potential.

[0044] 3. Considering the differences in learning situations among different users, the learning rates are converted into normative scores to enable more accurate cognitive assessment and comparison of users.

[0045] 4. By establishing a reinforcement learning model for cognitive learning tasks, the user's strategy parameters in the task can be obtained, so as to objectively judge whether the user answers seriously, and then decide whether the user needs to be re-evaluated to improve the accuracy of cognitive assessment. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 An overall flow chart of a cognitive potential assessment method based on a reinforcement learning computing model provided by an embodiment of the present invention;

[0047] Figure 2 A detailed flow chart of a cognitive potential assessment method based on a reinforcement learning computing model provided by an embodiment of the present invention;

[0048] Figure 3 A structural diagram of a cognitive potential assessment system based on a reinforcement learning computational model provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0049] The technical content of the present invention is described in detail below with reference to the accompanying drawings and specific embodiments.

[0050] The embodiment of the present invention provides a cognitive potential assessment method and system based on a reinforcement learning computing model. A subtask library is constructed based on multiple cognitive fields to form a preset task library. Based on the cognitive assessment tasks in the preset task library, a preset reinforcement learning computing model (Rescorla-Wagner Reinforcement Learning Model) is used to calculate the user's learning rate under different task conditions, thereby being able to assess the user's learning ability in various cognitive fields, and then comprehensively assess the user's cognitive potential based on multiple dimensions.

[0051] Specifically, Figure 1As shown, a cognitive potential assessment method based on a reinforcement learning computing model provided by an embodiment of the present invention specifically includes steps S1 to S7:

[0052] S1: Obtain cognitive assessment task combinations based on the preset task library.

[0053] Specifically, the preset task library is composed of multiple subtask libraries, each subtask library corresponds to a cognitive field, each subtask library includes multiple cognitive assessment tasks of the same type, each cognitive assessment task includes multiple options, and the scoring probability of each option in the same cognitive assessment task is different. In one embodiment of the present invention, the preset task library includes a first subtask library (corresponding to spatial orientation learning ability), a second subtask library (corresponding to objective law learning ability) and a third subtask library (corresponding to character attribute recognition ability).

[0054] In the first subtask library, each cognitive assessment task requires the user to find a specific spatial location, and different spatial locations have different scoring probabilities. For example, in a cognitive assessment task, the user is told to find a parking location for an animated car, and needs to click on different locations to understand whether the location can be parked and the degree of accuracy; different locations in space correspond to different scoring probabilities, and the user needs to find the spatial location with the highest scoring probability.

[0055] In the second subtask library, each cognitive assessment task requires the user to select a specific object, and different objects have different scoring probabilities. For example, in a cognitive assessment task, the user is told to find the object that is most likely to get a reward from a series of objects. The user needs to select different objects to understand the score of the object; each object corresponds to a different probability of scoring, and the user needs to find the object with the highest probability of scoring.

[0056] In the third subtask library, each cognitive assessment task requires the selection of a specific virtual character, and different characters have different friendliness attributes. For example: in a cognitive assessment task, the user is told to choose one from a series of virtual characters to interact in a game. The other party may choose to cooperate or betray. The user needs to choose different characters to understand the friendliness attribute of the object; the probability of the friendliness degree corresponding to each character is different, and the user needs to find the most friendly and easiest to cooperate character.

[0057] When you need to evaluate the user's cognitive potential, you need to first select the corresponding cognitive evaluation task based on the preset task library. The specific selection method is as follows:

[0058] The first method: Select by symptoms

[0059] Obtain the user's symptom information, select different cognitive fields based on the symptom information, and then select a corresponding number of cognitive assessment tasks from the corresponding subtask library to form a cognitive assessment task combination.

[0060] The second method: average selection

[0061] Determine a number of tasks in advance, and then extract the number of cognitive assessment tasks from each subtask library to form a cognitive assessment task combination;

[0062] The third method: selection based on evaluation needs

[0063] Alternatively, the user's assessment needs are obtained, and based on the assessment needs, the cognitive domains that need to be assessed are determined, and then a corresponding number of cognitive assessment tasks are selected from the corresponding subtask library to form a cognitive assessment task combination.

[0064] S2: Based on the cognitive assessment task combination, obtain the user's cognitive assessment data for each cognitive assessment task.

[0065] Specifically, the cognitive assessment data at least includes the specific options of the user for completing the cognitive assessment task and the score results. In one embodiment of the present invention, a cognitive assessment task includes multiple task trials (for example, 30 task trials). When completing a cognitive assessment task, the user needs to select one from k options in each task trial, where k is the total number of options for the cognitive assessment task. After the user completes the selection, it is determined whether the user's selection is scored based on the score probability of the option until the user completes 30 selections. It can be understood that the specific options of these 30 times, whether each option is scored, and the total score of the cognitive assessment task together constitute the cognitive assessment data of the cognitive assessment task.

[0066] It can be seen that the cognitive assessment data of each cognitive assessment task can be obtained based on the same method for subsequent data analysis (see step S3 for details).

[0067] S3: Based on the pre-trained reinforcement learning computational model, calculate the learning rate α of each cognitive assessment task separately.

[0068] Specifically, it includes steps S31 to S33:

[0069] S31: Establish the score fluctuation function of the kth option of the current cognitive assessment task to obtain the expected value of the implicit potential of the kth option in the tth task trial .

[0070] Among them, the score fluctuation function is expressed as: , represents the actual score of the kth option in the current cognitive assessment task in the tth task trial, represents the expected value of the implicit potential of the kth option in the tth task trial, It represents the deviation caused by the natural fluctuation of option probability, which is determined by the standard deviation of the trials of this option in the entire population.

[0071] S32: The actual score of the kth option in the t+1th task trial based on the user's current cognitive assessment task , calculate the user's learning rate α for the current cognitive assessment task.

[0072]

[0073]

[0074] in, represents the actual score of the kth option in the t+1th task trial, represents the actual score of the kth option in the tth task trial, It represents the expected error of the user for the kth option in the tth task trial. The value range of α is between 0 and 1. The closer it is to 1, the faster the user can master the rules of this type of cognitive assessment task.

[0075] S33: Repeat the above steps S31 to S32 until the learning rate α of each cognitive assessment task is calculated.

[0076] S4: Based on the pre-trained reinforcement learning computational model, calculate the strategy parameters for each cognitive assessment task separately .

[0077] Specifically, it includes steps S41 to S42:

[0078] S41: Based on the softmax function, the probability of the user selecting option k in the tth task trial It is expressed as:

[0079]

[0080] in, Determined by the expected difference of all options; is the strategy parameter of the current cognitive assessment task, which determines the random volatility of the user's choice. Indicates that the user makes a completely random selection. The larger the value, the more the user follows the value to learn; i represents the specific option, i∈(1,k);

[0081] S42: Set the strategy parameters of the user's current cognitive assessment task Compared with the preset threshold, if the strategy parameter is not less than the preset threshold, the user's learning rate α for the current cognitive assessment task is calculated normally; if the strategy parameter If the score is less than the preset threshold, the cognitive assessment will be repeated.

[0082] Specifically, in one embodiment of the present invention, the preset threshold is three standard deviations. is defined as an extreme value (i.e., less than three standard deviations), indicating that the user made a random choice in the cognitive assessment task and did not answer seriously, so a new cognitive assessment is required. If it is not less than three standard deviations, it means that the user follows the values ​​to learn in the cognitive assessment task and can proceed to the next step of norm transformation (see step S5 for details).

[0083] S5: Based on the preset norms of each cognitive assessment task, convert the learning rate α of each cognitive assessment task into a norm score.

[0084] In one embodiment of the present invention, considering the differences in learning situations of different users, the learning rate α is converted into a norm score to achieve a more accurate evaluation and comparison of different users. Specifically, the reinforcement learning calculation model has multiple preset norms built in, and the multiple preset norms correspond to the cognitive assessment tasks in the preset task library one by one, and the preset norms corresponding to each cognitive assessment task include standard scores corresponding to people of various age groups and educational levels.

[0085] Therefore, when the learning rate α of each cognitive assessment task of the user is determined based on step S3, and when it is judged through step S4 that the user answers the questions seriously, the learning rate α of each cognitive assessment task is compared with the corresponding preset norm, so as to determine the norm score corresponding to the learning rate α of each cognitive assessment task.

[0086] S6: Take a weighted average of the normative scores of each cognitive assessment task to obtain a comprehensive assessment score for the user.

[0087] Specifically, it includes steps S61 to S62:

[0088] S61: Determine the weighted weights of the normative scores of each cognitive assessment task.

[0089] Among them, the weighted weight of the normative score of each cognitive assessment task is proportional to the number of tasks of the same type.

[0090] S62: Taking a weighted average of the normative scores corresponding to each cognitive assessment task, thereby obtaining a comprehensive assessment score for the user.

[0091] It is understandable that different comprehensive evaluation scores correspond to different cognitive potentials. In one embodiment of the present invention, the comprehensive cognitive potential of the user can be obtained by evaluating the cognitive fields of the user in three dimensions.

[0092] S7: Obtain the cognitive potential of each cognitive domain separately.

[0093] Specifically, it includes steps S71 to S73:

[0094] S71: Based on the norm scores of the cognitive assessment tasks, the norm scores of the cognitive assessment tasks belonging to the first subtask library are averaged to obtain a first assessment score corresponding to the spatial orientation learning ability.

[0095] S72: Based on the norm scores of the cognitive assessment tasks, the norm scores of the cognitive assessment tasks belonging to the second subtask library are averaged to obtain a second assessment score corresponding to the objective rule learning ability.

[0096] S73: Based on the norm scores of each cognitive assessment task, the norm scores of the cognitive assessment tasks belonging to the third subtask library are averaged to obtain a third assessment score corresponding to the character attribute recognition ability.

[0097] It should be noted that there is no fixed order for steps S71 to S73 in the embodiment of the present invention, and there is no fixed order for steps S6 and S7.

[0098] like Figure 3 As shown, based on the above embodiment, an embodiment of the present invention further provides a cognitive potential assessment system based on a reinforcement learning computing model. The cognitive potential assessment system includes one or more processors 21 and a memory 22. The memory 22 is coupled to the processor 21 and is used to store one or more programs. When the one or more programs are executed by the one or more processors 21, the one or more processors 21 implement the cognitive potential assessment method based on the reinforcement learning computing model in the above embodiment.

[0099] Among them, the processor 21 is used to control the overall operation of the cognitive potential assessment system to complete all or part of the steps of the cognitive potential assessment method based on the reinforcement learning calculation model. The processor 21 can be a central processing unit (CPU), a graphics processing unit (GPU), a field programmable gate array (FPGA), an application-specific integrated circuit (ASIC), a digital signal processing (DSP) chip, etc. The memory 22 is used to store various types of data to support the operation of the cognitive potential assessment system, and these data may include, for example, instructions for any application or method for operating on the cognitive potential assessment system, and application-related data. The memory 22 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as a static random access memory (SRAM), an electrically erasable programmable read-only memory (EEPROM), an erasable programmable read-only memory (EPROM), a programmable read-only memory (PROM), a read-only memory (ROM), a magnetic memory, a flash memory, etc.

[0100] In an exemplary embodiment, the cognitive potential assessment system can be implemented by a computer chip or entity, or by a product with a certain function, for executing the above-mentioned cognitive potential assessment method based on the reinforcement learning computational model, and achieving the same technical effect as the above-mentioned method. A typical embodiment is a computer. Specifically, the computer can be, for example, a personal computer, a laptop computer, a vehicle-mounted human-computer interaction device, a cellular phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

[0101] In another exemplary embodiment, the present invention further provides a computer-readable storage medium including program instructions, which, when executed by a processor, implements the steps of the cognitive potential assessment method based on the reinforcement learning computational model in any of the above embodiments. For example, the computer-readable storage medium may be the above-mentioned memory including program instructions, and the above-mentioned program instructions may be executed by a processor of a cognitive potential assessment system to complete the above-mentioned cognitive potential assessment method based on the reinforcement learning computational model, and achieve the technical effect consistent with the above-mentioned method.

[0102] In summary, the cognitive potential assessment method and system based on the reinforcement learning computing model provided by the embodiment of the present invention have the following beneficial effects:

[0103] 1. Through the reinforcement learning computational model, the user's performance in a certain cognitive field is modeled to evaluate the user's cognitive potential in a certain cognitive field.

[0104] 2. It covers cognitive assessment tasks in space, objects, and society, and can evaluate the user's cognitive potential in different cognitive areas of basic visual space, perception, and complex social cognition, and can obtain the user's comprehensive cognitive potential.

[0105] 3. Considering the differences in learning situations among different users, the learning rates are converted into normative scores to enable more accurate cognitive assessment and comparison of users.

[0106] 4. By establishing a reinforcement learning model for cognitive learning tasks, the user's strategy parameters in the task can be obtained, so as to objectively judge whether the user answers seriously, and then decide whether the user needs to be re-evaluated to improve the accuracy of cognitive assessment.

[0107] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0108] The above is a detailed description of the cognitive potential assessment method and system based on the reinforcement learning computational model provided by the present invention. For those skilled in the art, any obvious changes made to it without departing from the essence of the present invention will constitute an infringement of the patent right of the present invention and will bear the corresponding legal liability.

Claims

1. A cognitive potential assessment method based on a reinforcement learning computational model, characterized in that The steps include: Acquire a cognitive assessment task combination based on a preset task library; wherein the preset task library includes multiple subtask libraries, each of which corresponds to a cognitive field, each of which includes multiple cognitive assessment tasks of the same type, and each of which includes multiple task trials. When completing a cognitive assessment task, the user needs to select one from k options in each task trial, k is the total number of options in each task trial, and the scoring probability of each option is different; when the user completes the selection, determine whether the user's selection is scored according to the scoring probability of the option until the user completes all task trials; Based on the cognitive assessment task combination, obtaining the user's cognitive assessment data for each cognitive assessment task; wherein the cognitive assessment data at least includes the specific options of the user completing all task trials, whether a single task trial is scored, and the total score of the cognitive assessment task; Based on the softmax function, the probability that the user chooses option k in the tth task trial It is expressed as: in, Determined by the expected difference of all options; is the strategy parameter of the current cognitive assessment task, which determines the random volatility of the user's choice. Indicates that the user makes a completely random selection. The larger the value, the more the user follows the value to learn; i represents the specific option, i∈(1,k); The strategy parameters of the user's current cognitive assessment task Compared with the preset threshold, if the policy parameter is not less than the preset threshold, the learning rate α of the user for the current cognitive assessment task is calculated normally; if the strategy parameter If it is less than the preset threshold, the cognitive assessment will be repeated; Based on the pre-trained reinforcement learning computational model, the learning rate α of each cognitive assessment task is calculated respectively; wherein, the learning rate α of each cognitive assessment task is calculated by: establishing the score fluctuation function of the kth option of the current cognitive assessment task to obtain the expected value of the implicit potential of the kth option in the tth task trial ; The score fluctuation function is expressed as: , represents the actual score of the kth option in the current cognitive assessment task in the tth task trial, represents the expected value of the implicit potential of the kth option in the tth task trial, represents the deviation caused by the natural fluctuation of option probability, which is determined by the standard deviation of the option and task trials in the entire population; based on the actual score of the kth option in the t+1th task trial of the user's current cognitive assessment task , calculate the user's learning rate α for the current cognitive assessment task; , , represents the actual score of the kth option in the t+1th task trial, represents the actual score of the kth option in the tth task trial, represents the expected error that the user will make for the kth option in the tth task trial; Based on the preset norms of each cognitive assessment task, the learning rate α of each cognitive assessment task is converted into a norm score; wherein the reinforcement learning calculation model has multiple preset norms built in, and each preset norm corresponds to a cognitive assessment task in the preset task library one by one; The norm scores of each cognitive assessment task are weighted averaged to obtain a comprehensive assessment score of the user, wherein different comprehensive assessment scores correspond to different cognitive potentials.

2. The cognitive potential assessment method according to claim 1, characterized in that: The preset task library includes a first subtask library, a second subtask library and a third subtask library; In the first subtask library, each cognitive assessment task requires the user to find a specific spatial position, so as to evaluate the user's spatial orientation learning ability; wherein, in each cognitive assessment task of the first subtask library, different spatial positions have different scoring probabilities; In the second subtask library, each cognitive assessment task requires the user to select a specific object to evaluate the user's objective regularity learning ability; wherein, in each cognitive assessment task of the second subtask library, different objects have different scoring probabilities; In the third subtask library, each cognitive assessment task requires the selection of a specific virtual character to evaluate the user's character attribute recognition ability; wherein, in each cognitive assessment task of the third subtask library, different characters have different friendliness attributes.

3. The cognitive potential assessment method according to claim 2, characterized in that Also includes: Based on the norm scores of each cognitive assessment task, averaging the norm scores of the cognitive assessment tasks belonging to the first subtask library to obtain a first assessment score corresponding to the spatial orientation learning ability; Based on the norm scores of each cognitive assessment task, averaging the norm scores of the cognitive assessment tasks belonging to the second subtask library to obtain a second assessment score corresponding to the objective regularity learning ability; Based on the norm scores of each cognitive assessment task, the norm scores of the cognitive assessment tasks belonging to the third subtask library are averaged to obtain a third assessment score corresponding to the character attribute recognition ability.

4. The cognitive potential assessment method according to claim 1, characterized in that The cognitive assessment task combination is obtained in the following way: The user's symptom information is obtained, different cognitive domains are selected based on the user's symptom information, and then a corresponding number of cognitive assessment tasks are selected from a corresponding subtask library to form a cognitive assessment task combination.

5. The cognitive potential assessment method according to claim 1, characterized in that The cognitive assessment task combination is obtained in the following way: A number of tasks is determined in advance, and then cognitive assessment tasks of that number are extracted from each subtask library to form a cognitive assessment task combination.

6. The cognitive potential assessment method according to claim 1, characterized in that The cognitive assessment task combination is obtained in the following way: The user's assessment requirements are obtained, and based on the assessment requirements, the cognitive domains that need to be assessed are determined, and then a corresponding number of cognitive assessment tasks are selected from the corresponding subtask library to form a cognitive assessment task combination.

7. The cognitive potential assessment method according to claim 1, characterized in that: The weighted weight of the normative scores of each of the cognitive assessment tasks is proportional to the number of tasks of the same type.

8. A cognitive potential assessment system based on a reinforcement learning computational model, characterized in that It includes a processor and a memory, wherein the memory is coupled to the processor; wherein, The memory is used to store a computer program, and when the computer program is executed by the processor, it is used to implement the cognitive potential assessment method described in any one of claims 1 to 7.

Images