Cognitive competence assessment method, system and equipment based on large model and medium

Abstract

The embodiment of the invention relates to the field of cognitive power, and discloses a cognitive ability assessment method, system and device based on a large model and a medium. The method comprises the steps that corresponding cognitive tasks are allocated to a user according to roles of the user participating in cognitive ability evaluation of at least one cognitive module, answer information input by the user in the process of processing all the cognitive tasks is acquired, and all the cognitive tasks and the corresponding answer information are input into a large model; performing preset processing on the cognitive task of each cognitive module and the corresponding answer information by using the large model to obtain an output result corresponding to each cognitive module; and analyzing the output result corresponding to each cognitive module to obtain a dynamic assessment result of the cognitive ability corresponding to each cognitive module. According to the method, the cognitive competence state of the user can be evaluated more accurately.

Description

Technical Field

[0001] The present invention relates to the field of cognitive dynamics, and in particular to a cognitive ability assessment method, system, device and medium based on a large model. Background Art

[0002] Cognitive ability refers to the ability of the human brain to process, store and retrieve information, that is, people's ability to grasp the composition of things, the relationship between their performance and other things, the driving force of development, the direction of development and the basic laws. The abilities of understanding, attention, thinking and imagination are all considered cognitive abilities.

[0003] Generally speaking, the learning behaviors, ability levels, and understanding levels displayed by learners during the learning process are closely related to their cognitive abilities. According to their cognitive abilities, learners can be guided to learn or think more efficiently. However, traditional brain imaging techniques for capturing human cognitive dynamics have high requirements for brain imaging tools, low universality, and cannot conduct targeted assessments of the cognitive abilities of learners at different cognitive levels.

[0004] Therefore, how to accurately and specifically evaluate learners' cognitive abilities has become an urgent problem to be solved. Summary of the invention

[0005] The purpose of the embodiments of the present invention is to provide a method, system, device and medium for evaluating cognitive ability based on a large model, and to evaluate different cognitive abilities of users according to the feedback answer information of users performing different cognitive tasks, and to obtain dynamic evaluation results of corresponding cognitive abilities. By obtaining the dynamic evaluation results of users performing cognitive tasks, the various cognitive ability conditions of users can be more accurately understood, and the changes and trends of the corresponding cognitive abilities of users can be analyzed according to the dynamic evaluation results, so as to formulate learning or guidance strategies in a targeted manner.

[0006] In order to solve the above technical problems, the embodiments of the present invention provide a cognitive ability assessment method based on a large model, comprising:

[0007] Assign corresponding cognitive tasks to users according to their roles when participating in cognitive ability assessment in at least one cognitive module, obtain answer information input by users during the processing of each cognitive task, and input each cognitive task and the corresponding answer information into a large model; use the large model to perform preset processing on the cognitive tasks and the corresponding answer information of each cognitive module to obtain output results corresponding to each cognitive module; analyze the output results corresponding to each cognitive module to obtain dynamic assessment results of cognitive abilities corresponding to each cognitive module.

[0008] An embodiment of the present invention also provides a cognitive ability assessment system based on a large model, including: a cognitive module, used to assign corresponding cognitive tasks to users according to the roles of users when participating in cognitive ability assessment, obtain answer information input by users during the processing of each cognitive task, and input each cognitive task and the corresponding answer information into the large model; a data processing module, using the large model to perform preset processing on the cognitive tasks and the corresponding answer information of each cognitive module to obtain output results corresponding to each cognitive module; a cognitive module, used to analyze the output results to obtain corresponding dynamic assessment results of cognitive ability.

[0009] An embodiment of the present invention also provides an electronic device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the large model-based cognitive ability assessment method as described above.

[0010] An embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the steps of the cognitive ability assessment method based on a large model as described above are implemented.

[0011] In addition, before assigning corresponding cognitive tasks to users according to the roles of users when participating in cognitive ability assessment of at least one cognitive module as described above, the method also includes: providing users with basic ability test questions of at least one cognitive module, obtaining corresponding test results, and determining the role of the user in participating in cognitive ability assessment in each cognitive module according to the test results; wherein each cognitive module is used to assess a cognitive ability.

[0012] In addition, each of the cognitive tasks as described above includes multiple cognitive questions; the step of obtaining answer information input by the user during the process of handling each cognitive task, and inputting each cognitive task and the corresponding answer information into the large model includes: obtaining multiple reply information input by the user during the process of handling each cognitive task to sequentially answer the multiple cognitive questions corresponding to the cognitive task, the multiple reply information constituting the answer information corresponding to the cognitive task; constituting each cognitive question and its corresponding reply information into a question-answer pair, and inputting each question-answer pair into the agent driven by the large model in the chronological order in which the cognitive questions in the question-answer pair appear in the cognitive task.

[0013] In addition, the cognitive module as described above includes at least: a comprehension module; the use of the large model to preset the cognitive tasks of each cognitive module and the corresponding answer information, including: using the agent to score the reply information in each question and answer pair in turn according to the input order of the question and answer pairs corresponding to each cognitive task.

[0014] In addition, the cognitive module as described above includes at least: a reasoning power module; the use of the large model to preset the cognitive tasks and the corresponding answer information of each cognitive module, including: using the agent to generate a reasoning process for the reply information of each question and answer pair according to the input order of the question and answer pairs corresponding to each cognitive task, and in turn according to the cognitive questions and the reply information in each question and answer pair.

[0015] In addition, the output results corresponding to each of the cognitive modules are analyzed as described above to obtain dynamic evaluation results of the cognitive abilities corresponding to each of the cognitive modules, including: extracting the score results of the response information of the agent in each of the question-answer pairs corresponding to the comprehension module in the output results; and obtaining dynamic evaluation results of the user's comprehension based on each of the score results.

[0016] In addition, the output results corresponding to each of the cognitive modules are analyzed as described above to obtain dynamic evaluation results of the cognitive abilities corresponding to each of the cognitive modules, including: extracting the reasoning process of each of the question-answer pairs corresponding to the reasoning power module by the agent in the output results; using a preset evaluation model to evaluate the rationality of each of the reasoning processes, outputting the results corresponding to the rationality index of the reasoning power module, and obtaining dynamic evaluation results of the user's reasoning power.

[0017] Compared with the prior art, the embodiments of the present invention assign corresponding cognitive tasks to users according to the roles of users when participating in the cognitive ability assessment of at least one cognitive module, obtain the answer information input by the user in the process of processing each cognitive task, and input each cognitive task and the corresponding answer information into the big model; use the big model to score each answer information, and generate a reasoning process according to each cognitive task and the corresponding answer information, and use each score result and each reasoning process as the output result of the big model; analyze the output results corresponding to each cognitive module, and obtain the dynamic evaluation results of the cognitive ability corresponding to each cognitive module. That is, the present invention determines the role of the user participating in the subsequent cognitive ability assessment by testing the basic level of the user's cognitive ability in different aspects, obtains the corresponding execution results by providing different cognitive tasks for the user, uses the big model to process the execution results to obtain the output results, and analyzes the output results of the big model to obtain the dynamic evaluation results of the corresponding cognitive ability. By obtaining the dynamic evaluation results of the user's execution of cognitive tasks, the cognitive ability status can be more accurately understood, and then the corresponding cognitive ability changes and trends of the user can be analyzed according to the dynamic evaluation results, and then the learning or guidance strategies can be formulated in a targeted manner. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 The specific process of the cognitive ability assessment method based on a large model according to an embodiment of the present invention is Figure 1 ;

[0019] Figure 2 The specific process of the cognitive ability assessment method based on a large model according to an embodiment of the present invention is Figure 2 ;

[0020] Figure 3 The specific process of the cognitive ability assessment method based on a large model according to an embodiment of the present invention is Figure 3 ;

[0021] Figure 4 is a schematic diagram of the structure of a cognitive ability assessment system based on a large model according to an embodiment of the present invention;

[0022] Figure 5 is a schematic structural diagram of an electronic device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0023] In order to make the purpose, technical scheme and advantages of the embodiments of the present invention clearer, the following will be described in detail with reference to the accompanying drawings. However, it will be appreciated by those skilled in the art that in the various embodiments of the present invention, many technical details are provided in order to enable the reader to better understand the present application. However, even without these technical details and various changes and modifications based on the following embodiments, the technical scheme claimed in the present application can be implemented.

[0024] One embodiment of the present invention relates to a method for evaluating cognitive ability based on a large model, such as Figure 1 As shown, this embodiment specifically includes the following steps.

[0025] Step 101: assign corresponding cognitive tasks to users according to the roles of users when participating in cognitive ability assessment of at least one cognitive module, obtain answer information input by users during the processing of each cognitive task, and input each cognitive task and corresponding answer information into the large model.

[0026] Specifically, cognitive ability refers to the ability of the human brain to process, store and extract information, including various basic cognitive abilities such as perception, attention, memory, and understanding, as well as various high-level cognitive abilities such as artistic creation, listening, speaking, reading and writing, problem solving, and creative thinking. Therefore, the cognitive ability in this embodiment is not a kind of ability, but a general term for one or several cognitive abilities to be evaluated. This embodiment uses at least one cognitive module to be set in a cognitive ability evaluation system based on a large model, and each cognitive module is used to evaluate a cognitive ability. Accordingly, the more cognitive modules there are, the more diverse the cognitive abilities that can be evaluated. This embodiment does not limit the number of cognitive modules set, that is, it does not limit the types of cognitive abilities to be evaluated.

[0027] Considering that different users must have differences in different cognitive abilities, in this embodiment, a role corresponding to the cognitive level is assigned to each user who participates in the cognitive ability assessment in the cognitive module, and the role is used to represent the baseline level of the user's current cognitive ability. Since the role of the user in each cognitive module represents the baseline level of the user's cognitive ability assessed in the module, this embodiment assigns cognitive tasks corresponding to the role to the user according to the role when the user participates in the cognitive ability assessment, so as to better assess the user's cognitive ability. This embodiment does not limit the specific content of the cognitive task. The cognitive task is a cognitive task pre-set in each cognitive module according to the type of cognitive ability to be assessed. For example, when the user's reasoning ability is to be assessed, different levels of reasoning tasks are pre-set as cognitive tasks, and reasoning tasks of corresponding difficulty are assigned according to the user's role in the cognitive module for assessing reasoning ability. The answer information input by the user in the process of processing each cognitive task is obtained, and each cognitive task and the corresponding answer information are input into the large model.

[0028] In one example, before step 101 is executed, the method of this embodiment also includes: providing the user with basic ability test questions of at least one cognitive module, obtaining corresponding test results, and determining the role of the user in participating in cognitive ability assessment in each cognitive module according to the test results; wherein each cognitive module is used to assess a cognitive ability.

[0029] Specifically, this embodiment can first set basic ability test questions for each cognitive ability in advance according to the different types of cognitive abilities to be evaluated, and the basic level of the user's cognitive ability can be determined by obtaining the test results of the test questions corresponding to different cognitive abilities. Since there must be differences in cognitive abilities between different users, it is necessary to evaluate the different cognitive abilities of different users in a targeted manner. This embodiment determines the role of the user in participating in the subsequent related cognitive ability evaluation in each cognitive module based on the test results of the user's basic ability test for different types of cognitive abilities. Different roles can represent different basic levels of different users under the same basic cognitive ability test, so as to achieve a more targeted and accurate evaluation.

[0030] For example, the scheme is applied in the context of education and teaching, and the scheme of the present invention is specifically described by taking students in the context of education and teaching as users. Taking students as users of the scheme, different cognitive abilities of students such as comprehension, creativity, and reasoning are evaluated. By obtaining the test results of students on the above basic ability test questions, a basic division is made for the ability level of students in each cognitive ability. For example, students participating in each cognitive ability assessment can be divided into beginners, intermediate learners, and advanced learners according to the test results. Different assessment means or methods are used for students of different levels to obtain corresponding assessment results, so as to provide direction for their subsequent learning and improvement of different subjects according to the different types of cognitive ability assessment results of students.

[0031] Determining the role of the user in participating in cognitive ability assessment in each cognitive module based on the test results may include: determining the baseline level of the user's cognitive ability corresponding to each cognitive module based on the test results corresponding to each cognitive module; determining the role of the user in participating in cognitive ability assessment in each cognitive module based on the preset correspondence between each baseline level and the role.

[0032] Specifically, the test results in this embodiment can be expressed in the form of scores, and the score segments are pre-divided to represent different baseline levels. For example, when the test results are in a percentage system, the corresponding baseline level can be a low level if the score is below 60 points; 60-80 points are a medium level, and 80-100 points are a high level. It should be noted that the division rules for dividing the baseline levels of each cognitive ability in this embodiment can be the same or different, and this embodiment does not make specific restrictions on this. Accordingly, this embodiment does not limit the specific method for obtaining the test results. When the basic ability test questions are objective questions, the test results can be directly reflected in the percentage system. When the basic ability test questions are subjective questions, the keyword scoring system can be used to obtain the test results. This embodiment does not make specific restrictions on this. After obtaining the baseline levels of different cognitive abilities of the user, the role of the user participating in the subsequent cognitive ability assessment in each cognitive module is determined according to the corresponding relationship between the preset baseline level and the role. Specifically, the correspondence between the baseline level and the role corresponding to each cognitive ability may be the same or different. This embodiment does not impose any specific limitation on this. It only requires that the user's test results be divided into roles of different levels for subsequent cognitive ability assessment.

[0033] For example, when there are two cognitive modules, which are used to evaluate the comprehension and creativity of Student A respectively, if it is determined that the comprehension of Student A is at the upper level among the upper, middle and lower levels according to the test results of the students in these two cognitive modules, then a role corresponding to the upper level is assigned to the student to participate in the subsequent dynamic evaluation of comprehension; if it is determined that the creativity of Student A is at the lower level among the upper and lower levels, then a role corresponding to the lower level is assigned to the student to participate in the subsequent dynamic evaluation of creativity.

[0034] Since the role of the user in each cognitive module represents the baseline level of the user's cognitive ability evaluated in the module, this embodiment assigns cognitive tasks corresponding to the role of the user when participating in this cognitive ability assessment, so as to better assess the user's cognitive ability. This embodiment does not limit the specific content of the cognitive task. The cognitive task is a cognitive task pre-set in each cognitive module according to the type of cognitive ability to be assessed. For example, when the user's reasoning ability is to be assessed, reasoning tasks of different levels are pre-set as cognitive tasks, and reasoning tasks of corresponding difficulty are assigned according to the user's role in the cognitive module for assessing reasoning ability. The answer information input by the user in the process of processing each cognitive task is obtained, and each cognitive task and the corresponding answer information are input into the large model.

[0035] In one example, each cognitive task includes multiple cognitive questions. Accordingly, the step 102 of obtaining the answer information input by the user during the process of processing each cognitive task and inputting each cognitive task and the corresponding answer information into the macro model includes:

[0036] Acquire multiple reply information input by the user in the process of processing each cognitive task, in which the user sequentially answers multiple cognitive questions corresponding to the cognitive task, and the multiple reply information constitutes the answer information corresponding to the cognitive task;

[0037] Each cognitive question and its corresponding reply information form a question-answer pair, and each question-answer pair is input into the agent driven by the large model in the chronological order in which the cognitive questions in each question-answer pair appear in the cognitive task.

[0038] Specifically, each cognitive task contains multiple cognitive questions that are continuous and logically connected. Therefore, the multiple reply information generated by the user's answers to the multiple cognitive questions in a cognitive task must be related to each other, and these answers can reflect the user's changes in thinking. Therefore, by obtaining the multiple reply information input by the user in the process of processing each cognitive task and answering the multiple cognitive questions corresponding to the cognitive task in sequence as the answer information corresponding to the cognitive task, information that can reflect the changes in the corresponding cognitive ability can be obtained.

[0039] Specifically, if you want to dynamically analyze the user's cognitive ability, you need to regard the process of the user handling cognitive tasks as a dynamic process, and divide different nodes for analysis. Since each cognitive task in this embodiment contains multiple cognitive problems, each or several cognitive problems can be used as an analysis node, and this embodiment does not limit this. In this implementation, each cognitive problem is analyzed as a node, and each cognitive problem and its corresponding reply information form a question-answer pair. Each question-answer pair is input into the agent driven by the large model according to the time order in which the cognitive problems in each question-answer pair appear in the cognitive task. The agent, also known as an artificial intelligence agent (AI Agent), is an intelligent entity that can perceive the environment, make decisions and perform actions. The intelligent agent is like a human. It has memory, logical analysis ability, task decomposition ability, problem decomposition ability and the ability to integrate and solve problems in the end.

[0040] Step 102: Use the large model to perform preset processing on the cognitive tasks of each cognitive module and the corresponding answer information to obtain output results corresponding to each cognitive module.

[0041] Specifically, the answer information input by the user during the process of processing each cognitive task is obtained, and after each cognitive task and the corresponding answer information are input into the macro model, the macro model is used to perform corresponding preset processing on each input answer question by inputting a specified prompt word into the macro model. That is, in this embodiment, different processing methods can be adopted to extract different information for different cognitive tasks and their corresponding answer information, so that the extracted information can reflect the cognitive abilities corresponding to different cognitive tasks.

[0042] In one example, the cognitive module includes at least: a comprehension module; in this case, the use of the large model in step 103 to preset the cognitive tasks and corresponding answer information of each cognitive module may include: using an agent to score the response information in each question and answer pair in turn according to the input order of the question and answer pairs corresponding to each cognitive task.

[0043] Specifically, when the cognitive module includes a comprehension module, assuming that the user is a student, since comprehension is an important indicator for measuring learning benefits, the comprehension of the student can be evaluated through the comprehension module, so that subsequent learning guidance based on their comprehension ability can help students improve their learning efficiency. The agent is used to score the answer information of the students to each cognitive task in turn. The specific scoring can be summarized by the text summary or keyword classification of the large model, etc., and this embodiment does not make specific restrictions here. The response information in each question and answer pair can be scored by the agent to preliminarily judge the student's answer to each question and preliminarily understand his comprehension ability. The student's comprehension can be reflected in the degree of text comprehension, that is, the comprehension module can evaluate the student's understanding of the text content, such as the ability to extract the main idea and key information of the article; it can also be reflected in the ability to grasp the main idea of ​​the text, evaluate the student's ability to grasp the main idea of ​​the text, such as summarizing the main idea of ​​the article, analyzing the author's point of view, etc.; it can also be reflected in the ability to analyze the text structure, evaluate the student's ability to analyze the text structure, such as analyzing the relationship between paragraphs, identifying the argument structure of the article, etc. By adjusting different cognitive tasks, a specific comprehension can be evaluated, and this application does not make specific restrictions on this.

[0044] In another example, the cognitive module includes at least: a reasoning module; in this case, the use of the large model in step 103 to preset the cognitive tasks and corresponding answer information of each cognitive module may include: using the agent to generate the reasoning process of the reply information of each question and answer pair according to the input order of the question and answer pairs corresponding to each cognitive task, and in turn according to the cognitive questions and reply information in each question and answer pair.

[0045] Specifically, when the cognitive module includes a comprehension module, assuming that the user is a student, since reasoning ability is the core ability of sharp thinking and analysis, quick response, and rapid grasp of problems, the reasoning ability of the student can be evaluated through the reasoning ability module, so that subsequent learning guidance for its reasoning ability can help students improve their learning efficiency. The agent is used to generate the reasoning process of the reply information of each question and answer pair according to the input order of the question and answer pairs corresponding to each cognitive task, and the cognitive questions and reply information in each question and answer pair are used in turn. Through the reasoning process in each question and answer pair, the student's reasoning process for each question can be preliminarily judged, and the reasoning ability of the student can be preliminarily understood. The student's reasoning ability can be reflected in logical reasoning ability, that is, the reasoning ability module evaluates the student's deductive reasoning, inductive reasoning, analogical reasoning and other abilities; it can also be reflected in mathematical reasoning ability, that is, the reasoning ability module evaluates the student's mathematical operation, logical reasoning and other abilities; it can also be reflected in scientific reasoning ability, that is, the reasoning ability module evaluates the student's understanding and application of scientific principles. By adjusting different cognitive tasks, a specific reasoning ability can be evaluated, and this application does not make specific restrictions on this.

[0046] There may be many ways to generate the reasoning process, and this embodiment does not impose any specific limitation on this.

[0047] For example, the reasoning process based on prompt engineering can be used to generate the information, as follows:

[0048] 1. Manually written prompts: Experts write a series of prompts for the agent based on the characteristics of cognitive modules and cognitive tasks to guide its reasoning. For example, in the reasoning module, the prompt can be "Please analyze the cause of the problem based on the known information" or "Please list several possible solutions and analyze their advantages and disadvantages."

[0049] 2. Template-based prompt generation: Using templates and natural language generation technology, prompts are automatically generated based on the information of cognitive modules and cognitive tasks. For example, the template can be "Problem: XXX, Analysis: XXX, Solution: XXX", and the system can fill in the template content according to the specific problem.

[0050] 3. Machine learning-based prompt generation: Train a machine learning model to generate prompts based on information from cognitive modules and cognitive tasks. The model can learn from prompts written by experts and be able to generate new prompts based on new tasks.

[0051] Alternatively, a reasoning process based on Chain of Thought (CoT) can be used, as follows:

[0052] 1. Manually write thought chains: Experts write a series of thought chains for agents based on the characteristics of cognitive modules and cognitive tasks to show the reasoning process. For example, in the reasoning module, the thought chain can be "Knowing A, because B, so C" or "Assuming A, if B, then C".

[0053] 2. Template-based thinking chain generation: Using templates and natural language generation technology, automatically generate thinking chains based on the information of cognitive modules and cognitive tasks. For example, the template can be "Step 1: XXX, Step 2: XXX, Step 3: XXX", and the system can fill in the template content according to the specific task.

[0054] 3. Thought chain generation based on machine learning: Train a machine learning model to generate thought chains based on information from cognitive modules and cognitive tasks. The model can learn thought chains written by experts and be able to generate new thought chains based on new tasks.

[0055] Alternatively, it can be generated using a reasoning process based on symbolic reasoning, as follows:

[0056] 1. Use a logical reasoning engine: Use a logical reasoning engine to perform symbolic reasoning based on information from cognitive modules and cognitive tasks, and generate reasoning processes. For example, use logical rules, predicate logic, etc.

[0057] 2. Use formal languages: Use formal languages, such as first-order logic, predicate logic, etc., to represent information about cognitive modules and cognitive tasks, and to perform reasoning and generate reasoning processes.

[0058] Alternatively, the reasoning process can be generated using case-based reasoning, as follows:

[0059] 1. Utilize the case library: Establish a case library containing cases of problem solving in the past. The agent can match the current task with the cases in the case library and generate reasoning processes.

[0060] 2. Use the case-based reasoning engine: Use the case-based reasoning engine to match the current task with the cases in the case library and generate the reasoning process.

[0061] Step 103: Analyze the output results corresponding to each cognitive module to obtain a dynamic evaluation result of the cognitive ability corresponding to each cognitive module.

[0062] Specifically, due to the different types of cognitive abilities, this embodiment can adopt different evaluation methods to evaluate different cognitive abilities. By analyzing the output results of each cognitive module, the cognitive ability corresponding to each cognitive module can be evaluated more specifically to obtain dynamic evaluation results of the cognitive ability corresponding to each cognitive module.

[0063] In one example, the cognitive module includes at least: a comprehension module; the step 102 of using the large model to preset the cognitive tasks and corresponding answer information of each cognitive module may include: using the agent to score the reply information in each question and answer pair in turn according to the input order of the question and answer pairs corresponding to each cognitive task, such as Figure 2 As shown, step 103 may include the following steps:

[0064] Step 1031: extracting the score results of the response information in each question-answer pair corresponding to the agent comprehension module in the output results;

[0065] Step 1032: Obtain a dynamic evaluation result of the user's comprehension ability based on each scoring result.

[0066] Specifically, since the cognitive questions in the cognitive task appear one by one in sequence, the score results of the reply information in each question-answer pair can reflect the dynamic changes of the student's understanding ability in the process of performing the cognitive task. The score results of the reply information in each question-answer pair corresponding to the agent pair understanding module in the output result are extracted. The score results can reflect the user's cognitive ability when answering the corresponding cognitive questions, that is, the dynamic evaluation results of the understanding module can be obtained according to each score result. For example, when a user needs to answer five cognitive questions to handle a cognitive task, when the large model uses a 10-point scoring system, the score results of the reply information in each question-answer pair are 6, 7, 7, 8, and 8 respectively, which can reflect that the student's cognitive ability is constantly improving in the performance of the cognitive task. In this case, the teacher can use the question logic similar to the cognitive task to guide the student to improve his understanding.

[0067] In another example, the cognitive module includes at least: a reasoning module; the step of using the large model to preset the cognitive tasks and corresponding answer information of each cognitive module in step 103 includes: using the agent to generate the reasoning process of the answer information of each question and answer pair according to the input order of the question and answer pairs corresponding to each cognitive task, and in turn according to the cognitive questions and answer information in each question and answer pair, such as Figure 3 As shown, step 103 may include the following steps:

[0068] Step 1033: extract the reasoning process of each question-answer pair corresponding to the agent reasoning module in the output result;

[0069] Step 1034: Use a preset evaluation model to evaluate the rationality of each reasoning process, output the result corresponding to the rationality index of the reasoning power module, and obtain a dynamic evaluation result of the user's reasoning power.

[0070] Specifically, an evaluation model can be pre-trained as a measure of the rationality index to evaluate the rationality of each reasoning process. The specific evaluation process is not specifically limited in this application. For example, the logic and consistency of the reasoning process can be evaluated by using logical rules, or the student's reasoning process can be compared with the reasoning process in the case library to evaluate its rationality. The evaluation model can adopt a five-level scoring system, as follows:

[0071] 5 points: The reasoning process is completely consistent with human expectations, highly consistent with personal profile or known information, and without any errors.

[0072] 4 points: The reasoning is clear and relevant, and the available information is accurately inferred, although there are some minor flaws.

[0073] 3 points: The reasoning is relevant but lacks clarity, e.g. only vague explanations are provided when a clear emotional inclination is required.

[0074] 2 points: The reasoning lacks clarity or shows weak cause-effect logic, characterized by forced analogies or simple repetition of the question.

[0075] 1 point: The reasoning is irrelevant and lacks logic.

[0076] Accordingly, by taking the output score as the result corresponding to the rationality index, a dynamic evaluation result of the user's reasoning ability can be obtained.

[0077] Compared with the related art, the embodiment of the present invention assigns corresponding cognitive tasks to users according to the roles of users when participating in the cognitive ability assessment of at least one cognitive module, obtains the answer information input by the user during the processing of each cognitive task, and inputs each cognitive task and the corresponding answer information into the big model; uses the big model to score each answer information, and generates a reasoning process according to each cognitive task and the corresponding answer information, and uses each score result and each reasoning process as the output result of the big model; analyzes the output results corresponding to each cognitive module, and obtains the dynamic evaluation results of the cognitive ability corresponding to each cognitive module. That is, the present invention determines the role of the user participating in the subsequent cognitive ability assessment by testing the basic level of the user's cognitive ability in different aspects, obtains the corresponding execution results by providing different cognitive tasks for the user, obtains the output results by processing the execution results using the big model, and analyzes the output results of the big model to obtain the dynamic evaluation results of the corresponding cognitive ability. By obtaining the dynamic evaluation results of the user's execution of cognitive tasks, the general situation of the user's cognitive ability can be obtained, and then the corresponding cognitive ability changes and trends of the user can be analyzed according to the dynamic evaluation results, and then the learning or guidance strategies can be formulated in a targeted manner.

[0078] Another embodiment of the present invention relates to a cognitive ability assessment system based on a large model, such as Figure 4 As shown, including:

[0079] The cognitive module 201 is used to assign corresponding cognitive tasks to users according to the roles of users when participating in cognitive ability assessment, obtain the answer information input by users during the process of processing cognitive tasks, and input the cognitive tasks and corresponding answer information into the large model;

[0080] The data processing module 202 is used to perform preset processing on the cognitive tasks of each cognitive module and the corresponding answer information to obtain output results corresponding to each cognitive module;

[0081] The cognitive module 201 is used to analyze the output results and obtain the dynamic evaluation results of the corresponding cognitive ability.

[0082] Specifically, the number of cognitive modules in this embodiment is at least one, and each cognitive module is used to evaluate a cognitive ability, such as comprehension, reasoning, creativity, etc., which is not limited in this embodiment. There is a corresponding role for participating in the cognitive ability evaluation of each cognitive module. Each cognitive module is used to obtain the answer information input by the user when performing its cognitive task, and input its cognitive task and corresponding answer information into the large model. Similarly, each cognitive module is used to obtain its corresponding output result processed by the large model from the large model and analyze it to obtain the dynamic evaluation result of the cognitive ability corresponding to the cognitive module.

[0083] It is not difficult to find that this embodiment is a system embodiment corresponding to the aforementioned method embodiments. The implementation details of the aforementioned embodiments are still applicable in this embodiment and will not be described in detail in this embodiment.

[0084] Another embodiment of the present invention relates to an electronic device, such as Figure 5 As shown, it includes at least one processor 302; and a memory 301 that is communicatively connected to the at least one processor 302; wherein the memory 301 stores instructions that can be executed by the at least one processor 302, and the instructions are executed by the at least one processor 302 so that the at least one processor 302 can execute any of the above method embodiments.

[0085] The memory 301 and the processor 302 are connected in a bus manner, and the bus may include any number of interconnected buses and bridges, and the bus connects various circuits of one or more processors 302 and the memory 301 together. The bus can also connect various other circuits such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and are therefore not further described herein. The bus interface provides an interface between the bus and the transceiver. The transceiver can be one element or multiple elements, such as multiple receivers and transmitters, providing a unit for communicating with various other devices on a transmission medium. The data processed by the processor 302 is transmitted on a wireless medium via an antenna, and further, the antenna also receives data and transmits the data to the processor 302.

[0086] The processor 302 is responsible for managing the bus and general processing, and can also provide various functions, including timing, peripheral interfaces, voltage regulation, power management and other control functions. The memory 301 can be used to store data used by the processor 302 when performing operations.

[0087] Another embodiment of the present invention relates to a computer-readable storage medium storing a computer program, which implements any of the above method embodiments when executed by a processor.

[0088] That is, those skilled in the art can understand that all or part of the steps in the above-mentioned embodiment method can be completed by instructing the relevant hardware through a program, and the program is stored in a storage medium, including several instructions to enable a device (which can be a single-chip microcomputer, chip, etc.) or a processor to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0089] Those skilled in the art will appreciate that the above-mentioned embodiments are specific examples for implementing the present invention, and in actual applications, various changes may be made thereto in form and detail without departing from the spirit and scope of the present invention.

Claims

1. A cognitive ability assessment method based on a large model, characterized in that: include: Assign corresponding cognitive tasks to users according to their roles when participating in cognitive ability assessment of at least one cognitive module, obtain answer information input by users during processing of each cognitive task, and input each cognitive task and the corresponding answer information into a large model; Using the large model, the cognitive tasks of each cognitive module and the corresponding answer information are processed in advance to obtain output results corresponding to each cognitive module; The output results corresponding to each of the cognitive modules are analyzed to obtain dynamic evaluation results of the cognitive abilities corresponding to each of the cognitive modules.

2. The method according to claim 1, characterized in that Before assigning corresponding cognitive tasks to users according to the roles of users when participating in cognitive ability assessment of at least one cognitive module, the method further includes: Providing the user with at least one basic ability test question of the cognitive module, obtaining the corresponding test result, and determining the role of the user in participating in the cognitive ability assessment in each cognitive module according to the test result; Each of the cognitive modules is used to evaluate a cognitive ability.

3. The method according to claim 1, characterized in that Each of the cognitive tasks includes a plurality of cognitive questions; The step of obtaining the answer information input by the user during the process of processing each cognitive task, and inputting each cognitive task and the corresponding answer information into the macro model includes: Acquire multiple reply information input by the user in the process of processing each of the cognitive tasks by sequentially answering the multiple cognitive questions corresponding to the cognitive task, wherein the multiple reply information constitute the answer information corresponding to the cognitive task; Each cognitive question and its corresponding reply information are combined into a question-answer pair, and each question-answer pair is input into the large model-driven agent in the time sequence in which the cognitive questions in the question-answer pairs appear in the cognitive task.

4. The method according to claim 3, characterized in that: The cognitive module includes at least: a comprehension module; The using of the large model to preset the cognitive tasks of each cognitive module and the corresponding answer information includes: The agent is used to score the response information in each question-answer pair in sequence according to the input order of the question-answer pairs corresponding to each cognitive task.

5. The method according to claim 3, characterized in that: The cognitive module at least includes: a reasoning module; The using of the large model to preset the cognitive tasks of each cognitive module and the corresponding answer information includes: The agent is used to generate a reasoning process for the response information of each question and answer pair according to the input order of the question and answer pairs corresponding to each cognitive task and in turn according to the cognitive questions and the response information in each question and answer pair.

6. The method according to claim 4, characterized in that The analyzing the output results corresponding to each of the cognitive modules to obtain the dynamic evaluation results of the cognitive ability corresponding to each of the cognitive modules includes: Extracting the score result of the response information of the agent in each question-answer pair corresponding to the comprehension module in the output result; A dynamic evaluation result of the user's comprehension ability is obtained according to each of the scoring results.

7. The method according to claim 5, characterized in that The analyzing the output results corresponding to each of the cognitive modules to obtain the dynamic evaluation results of the cognitive ability corresponding to each of the cognitive modules includes: Extracting the reasoning process of each of the question-answer pairs corresponding to the reasoning power module by the agent in the output result; A preset evaluation model is used to evaluate the rationality of each of the reasoning processes, and the result corresponding to the rationality index of the reasoning power module is output to obtain a dynamic evaluation result of the user's reasoning power.

8. A cognitive ability assessment system based on a large model, characterized in that: include: A cognitive module, used to assign corresponding cognitive tasks to users according to the roles of users when participating in cognitive ability assessment, obtain answer information input by users during the process of processing the cognitive tasks, and input the cognitive tasks and the corresponding answer information into the large model; A data processing module, used to perform preset processing using the cognitive tasks of each cognitive module and the corresponding answer information to obtain output results corresponding to each cognitive module; The cognitive module is used to analyze the output results to obtain dynamic evaluation results of corresponding cognitive abilities.

9. An electronic device, characterized in that: include: at least one processor; as well as, a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the large model-based cognitive ability assessment method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the large-model-based cognitive ability assessment method according to any one of claims 1 to 7.

Images