Information recognition method and apparatus

Abstract

The application discloses an information identification method and device, electronic equipment and a computer readable storage medium. Embodiments of the application obtain operation records of a target user, wherein the operation records include at least one adjustment intention obtained by the target user adjusting task content, and a reply result made by the target user for the task content before and after the adjustment; and a weak item result of the target user is located based on the adjustment intention and the reply result. Embodiments of the application can meet individual learning needs of different users.

Description

Technical Field

[0001] This application relates to the field of information recognition technology, specifically to an information recognition method and apparatus. Background Technology

[0002] With the continuous development of artificial intelligence technology, the education field has gradually introduced intelligent education systems to intelligently assist users in learning and teaching activities. Intelligent education systems are usually based on the relationship between the knowledge points that users learn to help users improve their learning outcomes. Since the relationship between knowledge points is an objective knowledge structure, although it can be used to help improve the learning outcomes of different users, it cannot meet the personalized learning needs of different users. Summary of the Invention

[0003] This application provides an information identification method, apparatus, electronic device, and computer-readable storage medium that can meet the personalized learning needs of different users.

[0004] In a first aspect, embodiments of this application provide an information identification method, the method comprising: Obtain the target user's operation records, wherein the operation records include at least one adjustment intention obtained by the target user in adjusting the task content, and the response results made by the target user regarding the task content before and after the content adjustment; Based on the adjustment intent and response results, the weaknesses of the target users were identified.

[0005] In some embodiments, identifying the target user's weaknesses based on the adjustment intent and response results includes: Based on the response results and adjustment intentions, calculate at least one moderating effect value for the target user on the task content, so as to obtain the weakness results according to each moderating effect value.

[0006] In some embodiments, the adjustment intent includes at least one content expression dimension, which indicates the direction of content adjustment for the task content; calculating at least one moderating effect value for the target user on the task content based on the response results and the adjustment intent, to obtain the weakness result based on the moderating effect value, including: For each content expression dimension, the moderating effect value of the target user on the task content is calculated based on the response results; The weaknesses of the target users are determined based on the moderating effect values ​​corresponding to each content expression dimension.

[0007] In some embodiments, the adjustment intent also includes a difficulty adjustment trend corresponding to the content expression dimension, and the response results include a first response to the task content before adjustment and a second response to the task content after adjustment; the adjustment effect value of the target user on the task content is calculated based on the response results, including: Determine the response conclusions for the first and second responses, whereby the response conclusions are used to indicate the scores of the target user's responses to the task content before and after the adjustment. Based on the content expression dimensions, difficulty adjustment trends, and response conclusions, the moderating effect values ​​of target users on task content under each content expression dimension are calculated.

[0008] In some embodiments, the response result further includes the response behavior of the first response content and the second response content, and the moderating effect value of the target user on the task content is calculated based on the response result, including: Based on each content expression dimension, difficulty adjustment trend, response conclusion, and response behavior, the moderating effect value of the target user on the task content under each content expression dimension is calculated.

[0009] In some embodiments, the method further includes: Based on the moderating effect value, determine the target user's mastery of the target expression dimensions in the task content; In response to the content adjustment operation, the task content is adjusted based on at least one content expression dimension other than the target expression dimension to obtain the adjusted task content.

[0010] In some embodiments, the response result includes a first response and a second response, and obtaining the target user's operation records includes: Obtain at least one first response from the target user regarding the task content before the adjustment; In response to a content adjustment event related to task content, determine at least one adjustment intent for the task content; Adjust the task content based on the adjustment intention to obtain the adjusted task content, and obtain at least one second response from the target user regarding the adjusted task content.

[0011] In some embodiments, the adjustment intent includes a content expression dimension and a difficulty adjustment trend corresponding to the content expression dimension. The task content is adjusted based on the adjustment intent to obtain the adjusted task content, including: Based on the content expression dimensions and difficulty adjustment trends, the task content is adjusted to obtain the adjusted task content, wherein the content expression of the task content before and after the adjustment remains consistent.

[0012] In some embodiments, identifying the target user's weaknesses based on the adjustment intent and response results includes: The adjustment intention and response results are input into the neural network model; By using a neural network model to process the adjustment intentions and response results, the results of the weaknesses can be obtained.

[0013] In some embodiments, after identifying the target user's weaknesses based on the adjustment intent and response results, the method further includes: Based on the results of weaknesses, learning recommendations for the target user are displayed on the graphical user interface.

[0014] Secondly, embodiments of this application also provide an information identification device, the device comprising: The record acquisition module is used to acquire the operation records of the target user. The operation records include at least one adjustment intention obtained by the target user in adjusting the task content, and the response results made by the target user to the task content before and after the content adjustment. The user identification module is used to locate the target user's weaknesses based on the adjustment intent and response results.

[0015] Thirdly, embodiments of this application also provide an electronic device, including a memory storing a computer program, which, when executed by a processor, causes the processor to execute any of the information identification methods provided in embodiments of this application.

[0016] Fourthly, embodiments of this application also provide a computer-readable storage medium including a computer program. When the computer program is run on an electronic device, the computer program is used to cause the electronic device to perform any of the information identification methods provided in the embodiments of this application.

[0017] In this embodiment, by acquiring the target user's operation records, which include at least one adjustment intention obtained by the target user in adjusting the task content, and the target user's response results to the task content before and after the content adjustment; based on the adjustment intention and response results, the target user's weakness results are located, thereby identifying the user's weakness results by adjusting the task content during the user's response to the task content, so as to meet the personalized learning needs of different users based on the user's weakness results. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a schematic flowchart of one embodiment of the information identification method provided in this application. Figure 2 This is a schematic diagram of the structure of the information recognition device provided in the embodiments of this application; Figure 3 This is a schematic diagram of the structure of the electronic device provided in the embodiments of this application. Detailed Implementation

[0020] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0021] Before providing a detailed explanation of the embodiments of this application, some terms involved in the embodiments of this application will be explained.

[0022] In the description of the embodiments of this application, the terms "first," "second," etc., may be used herein to describe various concepts, but unless specifically stated otherwise, these concepts are not limited by these terms. These terms are used only to distinguish one concept from another. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to these processes, methods, products, or devices.

[0023] This application provides an information identification method, apparatus, electronic device, and computer-readable storage medium. Specifically, the information identification method of this application can be executed by an electronic device, which can be a terminal or a server. The terminal can be a learning machine, smartphone, tablet computer, laptop computer, touch screen, game console, personal computer (PC), personal digital assistant (PDA), VR glasses, AR head-mounted display device, etc. The terminal can also include a client, which can be a game application client, a browser client carrying a game program, or an instant messaging client, etc. The server can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDN), and big data and artificial intelligence platforms.

[0024] For example, taking a terminal as an example, this electronic device can acquire the target user's operation records, which include at least one adjustment intention obtained by the target user in adjusting the task content, and the target user's response results to the task content before and after the content adjustment; based on the adjustment intention and response results, the target user's weakness results can be located, thereby identifying the user's weakness results by adjusting the task content during the user's response to the task content, so as to meet the personalized learning needs of different users based on the user's weakness results.

[0025] To address the aforementioned issues, this application provides a method, apparatus, electronic device, and computer-readable storage medium for error localization, which can meet the personalized learning needs of different users.

[0026] The following is a detailed description in conjunction with the accompanying drawings. It should be noted that the order of description of the following embodiments is not intended to limit the preferred order of the embodiments. Although a logical order is shown in the flowcharts, in some cases, the steps shown or described may be performed in a different order than that shown in the drawings.

[0027] In this embodiment, a terminal is used as an example for illustration. This embodiment provides an information identification method, such as... Figure 1 As shown, the specific process of this information recognition method can be as follows: Step 101: Obtain the target user's operation records, wherein the operation records include at least one adjustment intention obtained by the target user in adjusting the task content, and the response results made by the target user regarding the task content before and after the content adjustment.

[0028] Among them, the operation record is the relevant record of the target user's participation in the task content response processing in a historical period. For example, the task content can be test questions, and the corresponding operation record is the target user's intention to modify the test questions, the adjustment direction, and the response results before and after the adjustment of the test questions.

[0029] Step 102: Identify the target user's weaknesses based on the adjustment intent and response results.

[0030] In this embodiment, by analyzing the adjustments made to the task content by the target user's adjustment intentions and the target user's response results before and after the task content adjustment, the changes in the target user's response performance before and after the task content adjustment are clarified. This allows for tracking and comparison, in order to accurately identify the target user's comprehension obstacles and cognitive biases regarding different knowledge points during the response process, thereby obtaining the reasons for errors in the response process, i.e., the target user's weaknesses.

[0031] It is understandable that this embodiment, by identifying the target user's weaknesses, can provide information assistance in scenarios such as cognitive assessment, graded training, and personalized recommendations for the target user. This breaks through the limitations of traditional coarse-grained information in assessing whether the target user has learned a certain knowledge point, so as to facilitate personalized assessment and intervention for the target user.

[0032] It is understandable that in some embodiments, the target user's weaknesses can be located based on the target user's operation records. Therefore, the task content in the operation records obtained by the terminal can be the content that the target user has already participated in answering.

[0033] Optionally, when responding to the task content, the target user may adjust, modify, or regenerate the task content at least once, so that the operation record contains at least one intention to adjust, modify, or regenerate the task content, as well as the target user's response to the task content before and after the adjustment.

[0034] Understandably, adjustment intent can be used to indicate the direction of adjustment when modifying task content.

[0035] Optionally, the intention to adjust may include at least one content expression dimension, as well as the difficulty adjustment trend corresponding to the content expression dimension.

[0036] Among them, the content expression dimension refers to the information dimension included in the task content. Taking English short sentences as an example, the content expression dimension can include word dimension, grammar dimension, information density dimension, etc.; taking objective questions as an example, the content expression dimension can include various aspects such as interference dimension, computational volume dimension, and information density.

[0037] Optionally, the adjustment intent may also include a sentence or paragraph input by the user regarding the direction of adjustment to the task content. The system can analyze the adjustment to determine the content expression dimensions included in the adjustment intent.

[0038] For example, a user inputs a prompt into the large model: This text / This question is too difficult / Too hard to understand, make it simpler; After obtaining the adjusted task content, by comparing the task content before and after the adjustment, we can determine which content expression dimensions have changed in the adjusted task content compared to the task content before the adjustment.

[0039] For example, if the system adjusts a sentence from difficult to easier to understand and more readable from the perspective of words, it can be determined that the user's adjustment intention is to adjust or modify the task content from the perspective of content expression (word dimension).

[0040] It is understandable that the adjustment intention can be explicit information actively input by the user, or adjustment information obtained by the system based on the fuzzy instructions input by the user.

[0041] It is understood that in some embodiments, when adjusting the task content based on the adjustment intention, the content fragments in the task content can be adjusted. The content fragment can be the smallest content unit when adjusting the task content, such as a sentence, paragraph or title description in the task content; or the entire task content can be adjusted.

[0042] It is understandable that the response results may include the target user's first response to the task content before the adjustment and the second response to the task content after the adjustment. Alternatively, it may further include the corresponding response behaviors when responding to the first and second responses, etc. The specific settings can be configured according to the needs and are not limited here.

[0043] It should be noted that each round of test question adjustments includes a first response and a second response. The first response is the response generated by the target user when responding to the task content before the adjustment, while the second response is the response generated by the target user when responding to the task content after the adjustment.

[0044] For example, if the task involves multiple-choice questions, then the corresponding response will be the option selected by the target user.

[0045] It should be noted that when responding to the task content before and after the adjustment, the target user will generate certain response behaviors. When needed, the terminal can also use these response behaviors to locate the target user's weaknesses. These response behaviors are user behaviors when the target user responds to the task content. These response behaviors include, but are not limited to, response duration, marking information during the response, viewing of prompts for the task content (such as the task content answer), exit behavior, skipping behavior, attention behavior, etc.

[0046] Optionally, the operation log may further include content attribute information of the task content before and after the adjustment. This content attribute information includes, but is not limited to, the knowledge background information to which the task content belongs (e.g., knowledge point chapters, test papers), the standard answer of the task content, the organizational structure of the task content (e.g., question types), and the difficulty coefficient of the task content (the higher the difficulty coefficient, the more difficult it is to answer the task content).

[0047] For example, in a scenario where the task content is test questions, the operation record can record relevant information about all the test questions in a test paper answered by the target user, and the task content in the operation record can be at least one test question in the test paper. Alternatively, the operation record can record relevant information about all the test questions answered by the target user for a certain knowledge point, and the task content in the operation record can be at least one test question under that knowledge point. Or, the operation record can be customized and generated based on the target user's selection of test questions, etc.

[0048] It is understandable that during at least one round of content adjustment for the task content, the unadjusted task content and the task content after each round of adjustment match in terms of content expression. That is, the two are consistent or basically consistent in terms of semantics, main idea, knowledge core, or intention orientation, so as to promote a recognizable connection between the two. In other words, although the task content will be adjusted, the essential information or meaning of the task content will not be changed.

[0049] For example, during the process of adjusting the task content, although the specific content format changes, both test the same knowledge point, such as the basic properties of trigonometric functions (maximum value, period, etc.). Therefore, keeping the knowledge points consistent can be considered a case of content expression matching.

[0050] It is understandable that the task content may include, but is not limited to, text content, image content, or video content.

[0051] The tasks can be of different types, such as objective questions (multiple choice, true / false, fill-in-the-blank, etc.) and subjective questions (essay questions, calculation questions, etc.). The presentation of different types of task content varies, and the corresponding responses from the target users also differ.

[0052] It is understandable that the content expression dimension can be interpreted as the classification perspective that needs to be adjusted when adjusting the task content. The task content can contain multiple different classification perspectives, from which at least some classification perspectives are selected as the classification perspectives that need to be adjusted. The difficulty adjustment trend corresponding to this content expression dimension can be understood as the intended difficulty level when adjusting a specific classification perspective of the task content. It indicates the degree of difficulty that the task content needs to be adjusted in the content expression dimension, and the content difficulty of the task content should be increased or decreased based on the adjustment direction indicated by this difficulty adjustment trend.

[0053] It should be noted that by adjusting specific dimensions of task content expression, the learning adaptability of target users can be improved, reducing their cognitive load or increasing the challenge. Especially in student learning, maintaining consistency in content expression helps to accurately convey teaching objectives, ensuring that the core knowledge points or assessment targets remain unchanged, and preventing students from deviating from the learning focus due to changes in format.

[0054] In other words, even if target users use task content of different difficulty from other target users, they can still maintain consistency with other target users on a specific learning path. That is, the learning order of knowledge points on the learning path does not diverge, which makes it easier for teachers to grade or for the system to manage synchronously. This is conducive to meeting the corresponding personalized adjustments for different students in a large-scale teaching environment, so as to achieve individualized teaching.

[0055] Furthermore, it can also help the target user better meet their needs by making the task content they are currently handling more suitable. For example, if the task content is too difficult and the target user has difficulty understanding it, the difficulty of the content to be adjusted can be reduced by adjusting the difficulty trend of at least one dimension of the content expression. This will help the target user understand the meaning of the task content without changing the meaning of the text, such as understanding the corresponding difficult words or sentences, so as to respond and process the response.

[0056] Moreover, the adjusted cognitive load is controllable, avoiding the sunk cost of students with weak foundations giving up on more difficult teaching content. For example, applying the technical solution of this application to an intelligent educational learning machine allows for different levels of difficulty and different types of tasks to be provided for the same knowledge point, which is particularly beneficial for target users with poor learning foundations. It allows them to gradually transition from simple tasks to tasks with a certain degree of difficulty, thereby improving the quality of teaching.

[0057] In addition, after identifying the user's difficulties in the content expression dimension, the system can also recommend knowledge points and test questions that primarily or secondarily test that content expression dimension, so that the user can learn and consolidate their knowledge.

[0058] It is understandable that the content expression dimension refers to the various dimensions that constitute the language expression content, such as the vocabulary dimension, sentence structure dimension, structural hierarchy dimension, semantic reasoning dimension, etc. Different content expression dimensions can be set for the task content according to different types of task content.

[0059] Optionally, when the task content is an objective question, the content expression dimension may include at least one of the following dimensions: numerical dimension, distractor dimension, thinking guidance dimension, question stem scenario dimension, expression method dimension, number of steps dimension, use of graphs / tables dimension, language trap dimension, and reliance on teaching tools dimension.

[0060] Optionally, when the task content is a subjective question, the content expression dimension may include at least one of the following dimensions: style dimension, cultural adaptation dimension, vocabulary dimension, sentence structure dimension, allusion usage dimension, and expression subjectivity dimension.

[0061] It is understandable that, in the case of only one round of adjustments to the task content, the task content before the adjustment is the initial task content presented without any adjustments, and the task content after the adjustment is the task content obtained after adjusting the task content based on at least one adjustment intention.

[0062] Understandably, the task content can be adjusted multiple times to obtain the task content after each adjustment.

[0063] For example, when the task content is adjusted at least twice, the task content before adjustment can be the initial task content presented without any adjustment, or the task content obtained by adjustment in the previous adjustment rounds. The task content after adjustment is the task content obtained after adjusting the task content based on at least one adjustment intention of the current round.

[0064] It is understood that, in some embodiments, the terminal can calculate the target user's weakness results based on a specific algorithm, that is, locate the target user's weakness results based on adjustment intention and response results, which may include: Based on the response results and adjustment intentions, calculate at least one moderating effect value for the target user on the task content, so as to obtain the weakness results according to each moderating effect value.

[0065] In this embodiment, by analyzing the changes in the target user's responses before and after the task content adjustment, the target user's adjustment intention can be assessed using a specific algorithm. The impact of the adjustment intention on which the task content adjustment is based can be evaluated, i.e., the moderating effect value. Based on this moderating effect value, the target user's weakness results can be obtained.

[0066] Specifically, in the process of obtaining the results of the weak items based on each moderating effect value, the terminal can pre-set an evaluation condition to evaluate whether the moderating effect value meets the preset evaluation condition. If the moderating effect value meets the preset evaluation condition, the result of the weak item corresponding to the adjustment intention is determined.

[0067] Understandably, the preset evaluation conditions can be preset evaluation thresholds. These preset evaluation thresholds serve as the minimum requirement for determining whether the moderating effect value constitutes a weakness of the target user. This prevents the system from misjudging slight fluctuations as cognitive deficiencies. In other words, the moderating effect value must be greater than the preset evaluation threshold before the target user's weakness can be determined based on the adjustment intention associated with the moderating effect value.

[0068] The preset evaluation threshold can be set according to needs, and there are no restrictions here. For example, it can be set to 0.4 or 1.5 based on engineering experience, or the corresponding threshold can be set based on the fluctuation of historical performance.

[0069] Optionally, when the adjustment intention includes at least one content expression dimension, each moderating effect value can be the moderating effect value corresponding to each content expression dimension. The step "if the moderating effect value meets the preset evaluation criteria, then determine the weakness result corresponding to the adjustment intention" can, for example, include: If the moderating effect value corresponding to a certain content expression dimension is determined to meet the preset evaluation conditions, then the content expression dimension is determined to be a dimension that the target user has not mastered. Then, based on the content expression dimension, the corresponding weakness results are determined.

[0070] Furthermore, identifying the weaknesses corresponding to this content expression dimension can include: Based on the preset dimensional information mapping relationship, the target capability indication information corresponding to the content expression dimension is determined; the target capability indication information is used as the target user's weakness result.

[0071] It is understandable that, since different content expression dimensions indicate different capabilities required by the target users, certain rules or algorithms can be used to determine the weaknesses of the target users.

[0072] For example, a one-dimensional information mapping relationship can be established in advance so that when determining the content expression dimension that the target user has not mastered, the target user's required capabilities can be determined based on this one-dimensional information mapping relationship, i.e., target capability indication information, thereby obtaining the target user's weakness results.

[0073] For example, if a task involves the knowledge point of "function monotonicity," the system identifies or sets the content expression dimension corresponding to that task to include at least one of the following: "graphics usage dimension," "distractor dimension," "expression method dimension," and "teaching tool dependence dimension," as shown in Table 1 below:

[0074] Table 1 Among them, based on the mapping relationship between the content expression dimensions and the task content description in Table 1, the corresponding weaknesses of the target user are determined. This description can be used to explain the weaknesses of the identified user in a certain content expression dimension. For example, the description corresponding to the "expression mode dimension" is used to indicate whether the target user understands the application of abstract terms.

[0075] For example, the task content includes the knowledge point of function monotonicity, including dimensions such as graphical dimension, abstract language dimension, and interference complexity.

[0076] If the question stem is a low-interference, high-graphic question and the user answers correctly, it seems that the user has basic graphic recognition ability, but the corresponding suspected cognitive weakness is: weak understanding of terminology and weak language abstraction ability. If a user makes frequent mistakes in questions with complex and abstract expressions, it indicates a weakness in understanding abstract language. The corresponding suspected cognitive weakness is insufficient ability to distinguish easily confused functional properties.

[0077] Specifically, if a user answers correctly in a question with a low-interference, high-difficulty diagram, it indicates that the user has basic image recognition ability, but weak understanding of terminology and language abstraction ability; while if the user answers incorrectly in a question with a complex stem and abstract expression, it indicates that the user has a weak understanding of abstract language and insufficient ability to distinguish the properties of functions.

[0078] Specifically, the adjustment intention includes at least one content expression dimension, which is used to indicate the direction of content adjustment for the task content; based on the response results and the adjustment intention, at least one moderating effect value of the target user on the task content is calculated, so as to obtain the weakness results according to the moderating effect value, including: for each content expression dimension, calculating the moderating effect value of the target user on the task content based on the response results; and determining the weakness results of the target user according to the moderating effect value corresponding to each content expression dimension.

[0079] It is understood that in this embodiment, the terminal can determine the changes in the response results before and after the adjustment based on the response results, and calculate the moderating effect value of the target user on the task content based on the changes in the response results.

[0080] For example, if adjusting the task content from a certain content expression dimension results in a user's incorrect answer before the adjustment and a correct answer after the adjustment, it can be determined that adjusting the content expression dimension has a certain impact on the user's accuracy rate in answering questions.

[0081] For example, regarding the response time in the response results, if the response time before the adjustment is higher or lower than the response time after the adjustment, it can be seen that the change in the response results is a change in the response time. Therefore, it can be determined that adjusting this content expression dimension has a certain impact on the time it takes for users to answer questions.

[0082] In some embodiments, in a scenario where the difficulty adjustment trend of the content expression dimension in the adjusted intent changes from difficult to easy, the response results include a first response to the task content before adjustment and a second response to the task content after adjustment; the moderating effect value of the target user on the task content is calculated based on the response results, including: Determine the response conclusions for the first and second responses, whereby the response conclusions indicate the target user's scores for the responses to the task content before and after adjustment; and calculate the moderating effect value of the target user on the task content under each content expression dimension based on each content expression dimension and the response conclusions.

[0083] Specifically, based on each content expression dimension and the response conclusion, the moderating effect value of the target user on the task content under each content expression dimension can be calculated. This may include: determining the response change between the response conclusion of the first response content and the response conclusion of the second response content, and calculating the moderating effect value of the target user on the task content under each content expression dimension based on each content expression dimension and the response change.

[0084] The response conclusion can be obtained based on the degree of matching between the response content and the standard answer to the task content. The higher the degree of matching, the higher the score indicated by the response conclusion.

[0085] For example, in scenarios where the standard answer to a task is not fixed (such as an essay question), the degree of matching between the response and the standard answer can be determined by calculating similarity, hit rate, etc. The response conclusion can then be indicated by the score corresponding to the degree of matching; the higher the score, the higher the corresponding score.

[0086] For example, in scenarios where the standard answer to a task is fixed (such as multiple choice or true / false questions), if the answer is consistent with the standard answer, the answer conclusion can be indicated by text (correct) or numerical value (1) to indicate the highest score. If the answer is inconsistent with the standard answer, the answer conclusion can be indicated by text (incorrect) or numerical value (0) to indicate the lowest score.

[0087] Among them, the change in response refers to the change in the response results from before the task content was adjusted to after the task content was adjusted.

[0088] It should be noted that in scenarios where the standard answers to the task content are fixed (such as multiple-choice questions and true / false questions), the variations in responses can occur in the following ways: The conclusions of both the first and second responses are incorrect; the conclusions of both the first and second responses are correct; or one of the conclusions of the first and second responses is correct while the other is incorrect.

[0089] Furthermore, the response results also include the response behavior of the first response content and the second response content. Based on the response results, the moderating effect value of the target user on the task content is calculated, including: Based on each content expression dimension, response conclusion, and response behavior, the moderating effect value of the target user on the task content under each content expression dimension is calculated.

[0090] Specifically, based on each content expression dimension, response conclusion, and response behavior, the moderating effect value of the target user on the task content under each content expression dimension is calculated, which may include: Determine the change in response between the conclusion of the first response and the conclusion of the second response; determine the change in behavior between the response behavior of the first response and the response behavior of the second response; and calculate the moderating effect value of the target user on the task content under each content expression dimension based on each content expression dimension, the change in response, and the change in behavior.

[0091] Among them, the behavior change information is used to indicate the changes in the target user's behavior before and after the task content is adjusted, such as changes in the level of attention, changes in response time, and changes in the target user's confidence score through response behavior. The specific settings can be set according to the needs and are not limited here.

[0092] For example, for the target user Alice, the content expression dimension d of a task is adjusted. For instance, if the content expression dimension d is the vocabulary dimension (denoted as vocab), the following formula can be obtained for calculating the moderating effect value:

[0093] in, This is used to indicate the moderating effect value of the target user Alice in the content expression dimension d.

[0094] The Δ response status includes changes in response and changes in behavior.

[0095] in, This is a comprehensive function, which can be a weighted summation function, a Bayesian update function, an attention mechanism network function, etc. The specific function can be set according to the requirements, and there are no restrictions here.

[0096] For example, if Since the summation function is weighted, formula (1) for calculating the moderating effect value can be: (1) in, The conclusion of the second response is defined as follows: if the target user Alice answers the task content correctly after the difficulty of the vocabulary dimension is reduced, it is defined as 1.

[0097] in, The conclusion of the first response is defined as 0, for example, if the target user Alice answers the task incorrectly with high-difficulty vocabulary.

[0098] in, The response time for the first response, such as 75 seconds.

[0099] in, The response time for the second response content. For example, if the vocabulary in the task content is simplified, the response time will be changed to 42 seconds.

[0100] in, The weights corresponding to the changes are given, such as 0.7.

[0101] in, The weight corresponding to the change in behavior, such as 0.3. and The sum is 1.

[0102] Based on the examples corresponding to each parameter, the AES value can be obtained as follows:

[0103] To address this, a preset evaluation condition can be set as a preset evaluation threshold. For example, if the threshold τ is 2, and 10.6 is greater than 2, meaning the AES value is high, it indicates that Alice answered correctly immediately after simplifying the vocabulary, and the response time was also significantly shortened. This suggests that the target user Alice has an obstacle in the vocabulary dimension, and the weak points of the target user Alice can be determined based on the vocabulary dimension.

[0104] Based on formula (1), the AES value (modification effect value) can be set under different circumstances, and the AES value can be used to determine whether error localization can be performed, that is, whether the target user's weakness can be located, as shown in Table 2 below:

[0105] Table 2 In Table 2, “Dimensional Trend” refers to the difficulty adjustment trend. “Becoming more difficult” means that the task content is made more difficult based on the content expression dimension, thus increasing the content difficulty of the task content in the content expression dimension. “Becoming easier” means that the task content is made easier based on the content expression dimension, thus reducing the content difficulty of the task content in the content expression dimension.

[0106] In Table 2, "Changes in Responses" refers to the changes in the response conclusions of the first and second responses, i.e., the response changes. "1 to 0" means that the response conclusion of the first response is "1" and the response conclusion of the second response is "0". "1 to 1" means that the response conclusions of both the first and second responses are "1". "0 to 0" means that the response conclusions of both the first and second responses are "0". "0 to 1" means that the response conclusion of the first response is "0" and the response conclusion of the second response is "1". Here, "1" indicates that the response conclusion is correct and "0" indicates that the response conclusion is incorrect.

[0107] In Table 2, “duration change” refers to the change in the response time of the first response and the second response, i.e., the change in behavior. “Increase” means that the response time of the first response is shorter than that of the second response. “No change” means that the response time of the first response and the second response is the same. “Decrease” means that the response time of the first response is longer than that of the second response.

[0108] In Table 2: "AES value" refers to the moderating effect value; “ "" refers to the change in the response expressed in numerical value; “ "" refers to changes in behavior expressed in numerical values; “ =-0.2 means the duration increases by 0.2. =0 means the duration remains unchanged. =0.2” means the duration is reduced by 0.2; It should be noted that, ; in, The original data can be understood as the absolute change in answering speed, that is, the difference between the time a user takes to answer the adjusted questions and the time taken to answer the original questions before the adjustment. The scale is a normalized reference time, a standard for measuring time changes, for example, it can be set to 60 seconds as needed; Standardized data refers to the normalized value of time variation, a dimensionless numerical value used to correlate time variation with accuracy variation. Integrating them into the same order of magnitude facilitates planning and can be used to indicate the relative rate of change in answering speed; For example, when calculating AES values, you can set... It is 0.7. It is 0.3. It lasts for 60 seconds; "Error location" refers to whether the target user's weakness can be located. If so, the target user's weakness can be located; if not, the target user's weakness cannot be accurately located. "Explanation" is an explanation of the relevant situation.

[0109] It should be noted that the user's response conclusion can also be defined as the accuracy rate, rather than an absolute definition of 1 or 0.

[0110] For example, in the process of answering subjective questions, the change in the accuracy of users' answers before and after adjustment can also be used in this scheme to indicate the calculation of the adjustment effect value.

[0111] It is understandable that, in some embodiments, in order to accurately pinpoint the target user's weaknesses in scenarios where the difficulty adjustment trend changes, a variable representing the difficulty adjustment trend can be introduced into the formula for calculating the adjustment effect value.

[0112] Specifically, the moderating effect value of the target user on the task content is calculated based on the response results, including: Determine the response conclusions for the first and second responses, whereby the response conclusions indicate the scores of the target users' responses to the task content before and after adjustment; and calculate the moderating effect values ​​of the target users on the task content under each content expression dimension based on each content expression dimension, the difficulty adjustment trend, and the response conclusions.

[0113] Specifically, based on each content expression dimension, difficulty adjustment trend, and response conclusion, the moderating effect value of the target user on the task content under each content expression dimension is calculated, which may include: Determine the change in response between the conclusion of the first response and the conclusion of the second response. Based on each content expression dimension, difficulty adjustment trend, and response change, calculate the moderating effect value of the target user on the task content under each content expression dimension.

[0114] Furthermore, the response results also include the response behavior of the first response content and the second response content. Based on the response results, the moderating effect value of the target user on the task content is calculated, including: according to each content expression dimension, difficulty modulation trend, response conclusion and response behavior, the moderating effect value of the target user on the task content is calculated for each content expression dimension.

[0115] Specifically, based on each content expression dimension, difficulty adjustment trend, response conclusion, and response behavior, the moderating effect value of the target user on the task content under each content expression dimension is calculated. This may include: determining the response change between the response conclusion of the first response content and the response conclusion of the second response content; determining the behavioral change between the response behavior of the first response content and the response behavior of the second response content; and calculating the moderating effect value of the target user on the task content under each content expression dimension based on each content expression dimension, difficulty adjustment trend, response change, and behavioral change.

[0116] For example, for the target user Alice, the content expression dimension d of a task is adjusted, and the adjustment direction factor Sd is used to indicate the difficulty adjustment trend. For example, the content expression dimension d is the vocabulary dimension (denoted as vocab), and the following formula (2) can be obtained for calculating the moderating effect value: (2) in, This is used to indicate the moderating effect value of the target user Alice in the content expression dimension d.

[0117] If the difficulty adjustment trend is from difficult to easy, then Sd = +1; if the difficulty adjustment trend is from easy to difficult, then Sd = -1.

[0118] in, , The value ranges from -1 to 1 and is used to indicate changes in the response.

[0119] in, , , After normalization, the value ranges from approximately -1 to 1, and is used to indicate changes in behavior.

[0120] Among them, compared with formula (1), formula (2) ensures that when the difficulty adjustment trend changes from difficult to easy, the data of improved user performance can be summarized (i.e., the score of the response changes increases and / or the response behavior becomes better, such as the response behavior time decreases), and the AES value is positive, so that the cause of the error can be located.

[0121] Among them, compared with formula (1), formula (2) ensures that when the difficulty adjustment trend changes from easy to difficult, it is possible to summarize the data of deteriorating user performance (i.e., the score in the response change situation decreases and / or the response behavior situation worsens, such as the response behavior duration increases), and obtain a positive AES value, which can locate the cause of the error.

[0122] In contrast to formula (1), formula (2) ensures that when the difficulty adjustment trend is opposite to the summarized user performance change data (i.e., response changes and / or response behavior), such as when the difficulty adjustment trend changes from easy to difficult, data on improved user performance can be summarized, and when the difficulty adjustment trend changes from difficult to easy, data on deteriorated user performance can be summarized, resulting in a negative AES value, and thus no error can be located.

[0123] Based on formula (2), the AES value (modification effect value) can be set under different circumstances, and the AES value can be used to determine whether error localization can be performed, that is, whether the target user's weakness can be located, as shown in Table 3 below:

[0124] Table 3 In Table 3, “Dimensional Trend” refers to the difficulty adjustment trend. “Becoming more difficult” means that the task content is made more difficult based on the content expression dimension, thus increasing the content difficulty of the task content under the content expression dimension. “Becoming easier” means that the task content is made easier based on the content expression dimension, thus reducing the content difficulty of the task content under the content expression dimension.

[0125] In Table 3, "Changes in Responses" refers to the changes in the response conclusions of the first and second responses, i.e., the response changes. "1 to 0" means that the response conclusion of the first response is "1" and the response conclusion of the second response is "0". "1 to 1" means that the response conclusions of both the first and second responses are "1". "0 to 0" means that the response conclusions of both the first and second responses are "0". "0 to 1" means that the response conclusion of the first response is "0" and the response conclusion of the second response is "1". Here, "1" indicates that the response conclusion is correct and "0" indicates that the response conclusion is incorrect.

[0126] In Table 3, “duration change” refers to the change in the response time of the first response and the second response, i.e., the change in behavior. “Increase” means that the response time of the first response is shorter than that of the second response. “No change” means that the response time of the first response and the second response are the same. “Decrease” means that the response time of the first response is longer than that of the second response.

[0127] Among them, in Table 3: "AES value" refers to the moderating effect value; “ΔY” refers to the change in the response, expressed in numerical terms; “ "" refers to changes in behavior expressed in numerical values; “ =-0.2 means the duration increases by 0.2. =0 means the duration remains unchanged. =0.2” means the duration is reduced by 0.2; It should be noted that, ; in, The original data can be understood as the absolute change in answering speed, that is, the difference between the time a user takes to answer the adjusted questions and the time taken to answer the original questions before the adjustment. The scale is a normalized reference time, a standard for measuring time changes, for example, it can be set to 60 seconds as needed; Standardized data refers to the normalized value of time variation, a dimensionless numerical value used to correlate time variation with accuracy variation. Integrating them into the same order of magnitude facilitates planning and can be used to indicate the relative rate of change in answering speed; For example, when calculating AES values, you can set... It is 0.7. It is 0.3. It lasts for 60 seconds; “Sd” refers to the difficulty adjustment trend expressed in numerical form; "Error location" refers to whether the target user's weakness can be located. If so, the target user's weakness can be located; if not, the target user's weakness cannot be accurately located. "Explanation" is an explanation of the relevant situation.

[0128] It should be noted that the user's response conclusion can also be defined as the accuracy rate, rather than an absolute definition of 1 or 0.

[0129] For example, in the process of answering subjective questions, the change in the accuracy of users' answers before and after adjustment can also be used in this scheme to indicate the calculation of the adjustment effect value.

[0130] It is understood that, in some embodiments, this solution further includes: Based on the moderating effect value, the target user's mastery of the target expression dimension in the task content is determined; in response to the content adjustment operation, the task content is adjusted based on at least one content expression dimension other than the target expression dimension to obtain the adjusted task content.

[0131] In this embodiment, by recording the target expression dimensions already mastered by the target user, the system avoids repeatedly adjusting the task content based on the content expression dimensions already mastered by the target user. This prevents the target user from repeatedly training and learning the relevant abilities of the target expression dimensions they have already mastered, while failing to train the relevant abilities of the content expression dimensions they have not yet mastered in a timely manner, resulting in a cognitive gap for the target user. By determining and recording the target expression dimensions, the system can adjust the task content learned by the target user based on the content expression dimensions other than the target expression dimensions, thereby achieving a fine-tuned task content based on different dimensions of content expression dimensions with varying difficulty levels, such as pushing content based on the cognitive gap of the target user.

[0132] Specifically, when the target user adjusts the task content, the target user can be provided with content expression dimensions other than the target expression dimension, so that the target user can select a dimension and adjust the task content based on the dimension selected by the target user.

[0133] Specifically, it can respond to a task content adjustment instruction issued by a target user, which carries an adjusted content difficulty trend, and adjust the task content based on the content difficulty adjustment trend and content expression dimensions other than the target expression dimension.

[0134] The task content adjustment instruction can be triggered when the target user's response to the task content does not meet the preset accuracy conditions, or when the target user's response behavior is detected to meet the preset response obstacle conditions.

[0135] Among them, response behaviors that meet the preset response obstacle conditions include, but are not limited to: response time exceeding the preset time, and response behavior conforming to preset events, such as the event of the target user viewing the response prompt information.

[0136] In some embodiments, the identification of a target user's weaknesses can be based on a neural network model, and the location of the target user's weaknesses can be based on the adjustment intention and response results. This can include: inputting the adjustment intention and response results into a neural network model; and processing the adjustment intention and response results through the neural network model to obtain the weakness results.

[0137] Among them, the neural network model is used to identify weaknesses (i.e. error location). This model can infer the weaknesses of the target user's response based on the target user's response before and after adjusting a specific content expression dimension.

[0138] Specifically, the historical adjustment intentions and corresponding response results of target users in historical periods can be used as samples, and the historical weakness results marked by humans or machines can be used as labels to train the neural network model. This enables the neural network model to learn how to identify and locate the weakness results of target users, so that when the model is applied, the weakness results of target users can be identified based on adjustment intentions and response results.

[0139] In some embodiments, after identifying the target user's weaknesses based on the adjustment intent and response results, the method further includes: displaying learning recommendation information for the target user on a graphical user interface based on the weaknesses results.

[0140] In one alternative implementation, recommended keywords can be determined based on the target user's weakness results, and then the recommended keywords can be filled into the corresponding recommendation template to provide the target user with more accurate teaching suggestions.

[0141] For example, when a target user gets a function question wrong, instead of simply displaying the message "You got a function question wrong," a more effective way to recommend learning methods can be provided. For instance, the learning recommendation message could be something like, "You are more likely to make mistakes when dealing with monotonic problems containing abstract terms. We suggest reviewing the meaning of derivatives and the language expression of graph trends."

[0142] In an optional implementation, a personalized learning plan, or learning recommendation information, can be generated for each of the target user's weaknesses based on the target user's results. This personalized learning plan can be displayed to the target user through graphics or text to improve the target user's results in their weaknesses.

[0143] In one optional implementation, a dynamic learning profile of the target user can be constructed based on the target user's weaknesses, operation records, and other information. This allows for the automatic matching of subsequent content with the most suitable expression methods based on the target user's ability level and preferences. Furthermore, learning recommendation information that aligns with the target user's dynamic learning scenario can be recommended to the target user.

[0144] For example, for target users who prefer to reduce the difficulty of vocabulary, materials that are more suitable for their vocabulary level can be recommended first, avoiding repeated manual adjustments.

[0145] It is understandable that a graphical target user interface can refer to the display interface of an electronic device held by the target user, a pop-up window on the display interface, a specified area, or a virtual interface projected by a certain device, etc., without limitation.

[0146] It is understandable that, in order to adapt to different learning scenarios, in some embodiments, if the user does not respond to at least one of the task content before and after the adjustment, the response content corresponding to the unresponded task content is assumed to be empty, the response conclusion corresponding to the response content is wrong, or the score is 0.

[0147] It is understood that, in some embodiments, the response results included in the operation log during at least one round of task content adjustment may also include response results corresponding to the task content before and after the adjustment, so as to comprehensively evaluate the influencing factors affecting the target user's response, i.e. the target user's weaknesses, in order to further improve the accuracy of identifying the target user's weaknesses. For example, it can be determined whether the difficulty is caused by "the structure diagram in the question itself (i.e., the difficulty of the language expression of the task content)" or "the student does not understand the knowledge point to which the task content belongs (the knowledge point itself has not been mastered)".

[0148] Furthermore, the effectiveness of the adjustment strategy can be determined based on the target user's responses to the task content before and after the adjustment. That is, if the target user did not respond correctly to the task content before the adjustment, but responded correctly to the task content after the adjustment, it indicates that the target user's adjustment strategy for the task content is effective. This not only helps to identify the target user's weaknesses, but also clarifies under which expressions the target user is less likely to make mistakes.

[0149] In some embodiments, where the adjustment intent includes at least one content expression dimension, locating the target user's weakness based on the adjustment intent and the response result includes: The task content can be adjusted multiple times, with each adjustment including one or more content expression dimensions. To identify weaknesses, the response results after each adjustment can be used to determine the content expression dimensions that the target user has not mastered; based on these unmastered content expression dimensions, the target user's weaknesses can be identified.

[0150] It is understandable that the response results include the first response and the second response from the target user regarding the task content before and after the adjustment, and each response has a corresponding response result.

[0151] Specifically, each response can be compared with the standard answer to determine the conclusion of each response.

[0152] Optionally, the responses can be graded and reviewed using a large language model to obtain a response that meets the requirements.

[0153] It should be noted that the content of the response can vary depending on the type of task. For example, for objective questions, since the standard answer has only one solution, the response can be either correct or incorrect. For subjective questions, there may not be a single correct answer. Therefore, the accuracy of the response can be assessed. For example, the accuracy and score can be determined by the degree of matching or similarity between the response and the standard answer.

[0154] Understandably, in order to accurately pinpoint the user's weaknesses in specific dimensions, the task content can be adjusted multiple times, the user's answers recorded, and comparative analysis conducted to accurately identify the user's knowledge or skill deficiencies.

[0155] In this context, the knowledge or skills gaps can be understood as the content expression dimensions that the target user lacks, as mentioned in this embodiment. The real difficulties are identified by comparing the user's difficulty adjustment in a certain expression dimension and their performance in answering the task content before and after the adjustment.

[0156] The task content can be understood as a question or task to be tested on the user, such as a math problem or a simple question.

[0157] It should be noted that task content adjustment refers to adjusting the difficulty of one or more content dimensions within the same core task, while keeping the knowledge and skills to be tested unchanged. This can be done by lowering or increasing the difficulty, in order to assess the user's depth of understanding and mastery of the knowledge.

[0158] The response result may include the user's response behavior, the content of the response, and the corresponding response conclusion, i.e., whether it is right or wrong, or the score.

[0159] The "adjustment round" refers to the number of times the task content has been adjusted.

[0160] Understandably, if a user answers incorrectly regarding the original task content, it's important to determine which knowledge point the user hasn't mastered and in which dimension there's an obstacle to understanding or mastering it; for example, whether the problem lies in not knowing multiplication itself, not understanding word problems, or being unable to convert diagrams into mathematical problems.

[0161] In existing technologies, this situation is generally only identified as a user not knowing the knowledge point tested in a certain question, such as not knowing multiplication. However, it may just be that the user is not used to the way the question is expressed, rather than that they do not know the core knowledge point. This leads to an inaccurate technical problem in identifying the cause of the error.

[0162] In this embodiment, we can first record which round of task adjustment the user answered correctly, and then determine whether the "content expression dimension" used or included in that round of task is something the user can adapt to and understand.

[0163] Then, the system can record which rounds the user answered incorrectly. The "content expression dimensions" used in these rounds might have included aspects the user didn't understand, thus identifying the rounds in which the user answered incorrectly. Finally, all content expression dimensions encountered by the user in the incorrect rounds are treated as a set to be analyzed. Then, using the content expression dimensions that the user has demonstrated mastery in the correct rounds (rounds in which the user answered correctly) as a filtering criterion, the incorrect set is filtered out. This reveals the dimensions that appeared in the rounds where the user made mistakes but did not appear in the rounds where the user answered correctly, or were not demonstrated to be mastered. These filtered dimensions can be determined as content expression dimensions that the user has not truly mastered.

[0164] For example, a question might test the core knowledge point of understanding proportional relationships, while also assessing the user's ability to comprehend text in terms of information expression methods.

[0165] The initial question is: if 2 apples cost 10 yuan, how much do 6 apples cost? The user's answer was incorrect.

[0166] At this point, the system cannot accurately determine whether the user has not grasped the proportion of core knowledge points or has difficulty reading the question.

[0167] After the system or the user actively adjusts the task content in a certain dimension, for example, changing the question from a text description to a graphic description, showing a picture of 2 apples priced at 10 yuan, and 6 apples next to it, asking for the total price.

[0168] The user's answer is correct.

[0169] The system can determine that the user has the ability to understand proportions through graphics.

[0170] Finally, the task content is adjusted, and the formula y=kx is used to solve the ratio problem. The adjusted task content is: y=kx, given that y equals 10 when x=2, find the value of y when x=6.

[0171] The task content will be adjusted from the perspective of information expression, focusing on understanding more complex abstract formulas.

[0172] The user's answer was incorrect.

[0173] Based on the above operation records, the system can calculate that if the user answers the task content correctly in the graphical dimension, it means that there is no problem in mastering the core knowledge points. However, the user cannot understand the more abstract textual descriptions, indicating that the understanding of textual descriptions and abstract formulas may be the user's weakness.

[0174] Understandably, this embodiment can accurately pinpoint the core concept of proportion for users (because they can answer correctly with graphics), but users' shortcomings lie in their ability to transform textual questions into mathematical models and their ability to use abstract algebraic symbols.

[0175] Furthermore, based on the conclusions drawn above, targeted learning plans can be generated instead of repeatedly explaining the basic concepts of proportion, which greatly improves the user's learning efficiency.

[0176] It should be noted that if the response to the adjusted task content is correct, it means that the target user has grasped the content expression dimensions of the adjusted task content in the correct round, and therefore the target user can answer the corresponding task content correctly. However, if the response to the adjusted task content is incorrect, it means that the adjusted task content in the incorrect round includes content expression dimensions that the user has not grasped, and therefore the target user cannot answer the corresponding task content correctly.

[0177] Therefore, in this embodiment, the target user's understanding of the content expression dimension can be assessed based on the response results corresponding to the task content after multiple rounds of adjustments.

[0178] For example, if a task A requires adjustments to the following content expression dimensions in a certain round: "graphic usage dimension," "distractor dimension," and "expression method dimension," and the answer result for task A is "correct," then task A does not have any content expression dimensions that the target user is unaware of in the current round. However, if the answer result for task A is "incorrect," then task A has content expression dimensions that the user is unaware of in the "graphic usage dimension," "distractor dimension," and "expression method dimension" in the current round.

[0179] Furthermore, since there may be at least two content expression dimensions when adjusting the task content in each round, if the response result is incorrect, it may be because the target user cannot grasp one of the content expression dimensions, leading to the failure of the response result. Therefore, the content expression dimensions in the incorrect round can be filtered based on the content expression dimensions in the correct round or the content expression dimensions that have been proven to be grasped by the user. That is, the same content expression dimensions are filtered out, and the remaining content expression dimensions after filtering are the content expression dimensions that the user has not grasped.

[0180] It should be noted that if the responses before and after a certain round of adjustment are both correct, it means that the target user has mastered the content expression dimension adjusted in that correct round. Therefore, regardless of the difficulty of the content corresponding to the content expression dimension, the target user can answer the corresponding task content correctly. However, if there are errors in the responses before and / or after the task content adjustment, it means that there is a content expression dimension that the target user has not fully mastered in that incorrect round. Therefore, there may be situations where the content corresponding to the content expression dimension is difficult, or where the target user cannot answer the corresponding task content correctly regardless of the difficulty of the content corresponding to the content expression dimension.

[0181] Therefore, in this embodiment, the response results corresponding to the task content before and after the adjustment can be used to accurately assess whether the target user has grasped the content expression dimension when the task content is adjusted.

[0182] For example, if a task content A requires adjustments to its content expression dimensions in a certain round, namely "graphic usage dimension," "distractor dimension," and "expression method dimension," and the responses to task content A before and after the adjustments are both correct, then task content A does not have any content expression dimensions that the target user is unaware of in the current round. However, if errors occur in the responses to task content A before and / or after the adjustments, then task content A has content expression dimensions that the user is unaware of in the "graphic usage dimension," "distractor dimension," and "expression method dimension" in the current round.

[0183] In some embodiments, if the response results before and after the task content adjustment are different (i.e., one is correct and the other is incorrect), it indicates that the target user has mastered the knowledge points to which the task content belongs, but the target user has not mastered the content expression dimension corresponding to the adjustment, which leads to inconsistent response results before and after the task content adjustment when the content corresponding to the content expression dimension of the task content is adjusted.

[0184] For example, if the answer before the adjustment is correct, but the answer after the adjustment is incorrect, it means that the target user cannot answer the task content correctly after the difficulty of the content expression dimension is increased. Or, if the answer before the adjustment is incorrect, but the answer after the adjustment is correct, it means that the target user can answer the task content correctly after the difficulty of the content expression dimension is reduced. This shows that the target user has mastered the task content, and the reason why the target user cannot answer the task content correctly is simply because the expression of the task content has been updated. This indicates that the target user has not mastered the content expression dimension corresponding to the adjustment.

[0185] In order to better identify the current weaknesses of the target user, or to clarify the reasons for the target user's incorrect answers to the task content, the results of the target user's weaknesses can include the target user's ability indication information and knowledge mastery indication information corresponding to the content expression dimension.

[0186] Specifically, if the response results before and after the task content adjustment are both correct or different, the knowledge mastery indication information is the first knowledge mastery indication information. This first knowledge mastery indication information is used to indicate the knowledge points that the target user has currently mastered in relation to the task content.

[0187] Specifically, if the response results before and after the task content adjustment are both incorrect, then the knowledge mastery indication information is the second knowledge mastery indication information, which is used to indicate the knowledge points to which the target user currently does not master the task content.

[0188] In some embodiments, in order to meet the personalized needs of the current target user when responding to the task content and to improve the learning efficiency of the current target user, when displaying the task content (which is the initial content without adjustment by the target user) to the target user, the overall difficulty level of the task content can be matched with the target user's current learning level. Thus, by displaying task content that matches the target user's current ability, the current target user can learn the task content or information related to the task content more efficiently.

[0189] In some embodiments, in a scenario where the response result includes a first response and a second response, obtaining the target user's operation record includes: obtaining at least one first response from the target user to the task content before adjustment; in response to a content adjustment event for the task content, determining at least one adjustment intent for the task content; adjusting the task content based on the adjustment intent to obtain the adjusted task content, and obtaining at least one second response from the target user to the adjusted task content.

[0190] In this embodiment, in order to achieve a more granular adjustment capability for task content and a more controllable selection capability for content expression dimensions, thereby improving the adjustability of learning content and the personalized learning experience of target users, a content adjustment event is used to prompt the terminal to understand the target user's intention to adjust the task content. Based on the target user's intention to adjust the task content, the task content is adjusted, thereby facilitating the target user's response to the adjusted task content.

[0191] Specifically, the adjustment intention includes the content expression dimension and the difficulty adjustment trend corresponding to the content expression dimension. The task content is adjusted based on the adjustment intention to obtain the adjusted task content, including: adjusting the task content based on the content expression dimension and the difficulty adjustment trend to obtain the adjusted task content, wherein the content expression of the task content before and after the adjustment remains matched.

[0192] In some embodiments, the step "adjusting the task content based on the content expression dimension and difficulty adjustment trend to obtain adjusted task content" may include: Based on the content type and structure of the task content, the corresponding structure division rules are used to parse the task content in order to generate multiple sequence information. The target sequence information is determined from multiple sequence information, and the task content is adjusted based on the content expression dimension and difficulty adjustment trend to obtain the adjusted task content and adjust the target sequence information. Based on the adjusted target sequence information, the adjusted task content is obtained.

[0193] The methods for determining target sequence information from sequence information may include, but are not limited to: target users selecting target sequence information from it (for example, target users selecting or circling the stem, options, etc. of a multiple-choice question), intelligently identifying target sequence information that matches the content expression dimension, and randomly selecting target sequence information.

[0194] Among them, sequence information refers to the content fragments mentioned above. This sequence information can be information units in the task content that have structural or semantic recognizability and can express a certain knowledge objective independently or in combination. The sequence information of the smallest unit can refer to the indivisible basic expression structure, such as words, numbers, code blocks, image units, etc. Combinatable sequence information can be an expressive structure composed of multiple smallest units, such as a sentence, a question stem, or a complete multiple-choice question; Structural partitioning rules can refer to configurable rules used by the system to parse the structure of task content, determining which segments can be used as sequence information: The structural partitioning rules include, but are not limited to: Linguistic levels: word, phrase, sentence; Structural markers: Question stem, options, and explanation section; Semantic units: knowledge point expression, semantics of distractors; Independently editable marked areas: title annotations, image elements, etc.

[0195] For example, if the task type is a subjective question, the granularity of its sequence information can be refined as follows: Minimum granularity: a single word (verb, noun, adjective, adverb, etc.); Medium granularity: A complete sentence structure (such as a sentence containing a subject-verb-object structure); Larger granularity: a paragraph or question-and-answer structure; The system can identify the hierarchy based on language segmentation rules and syntactic analysis models, and generate multiple sequence information options for the target user.

[0196] For example, if the task type is objective questions, the granularity of its sequence information can be refined as follows: The question stem itself can be further divided into: (1) Background description (which may be simplified or have complex grammar added to increase reading difficulty) (2) Core test points expression section (synonyms can be used and sentence structure can be adjusted) The options area can be further divided into: (1) Correct option content (can be refined or semantically enhanced) (2) Distractors (can change their deceptiveness and enhance their contrast) Understandably, the system can generate corresponding sequence information independently for each part of the content, or it can generate sequence information for the overall task content.

[0197] In some embodiments, the difficulty adjustment trend includes a target difficulty level corresponding to the content expression dimension, wherein the target difficulty level is the intended difficulty level relative to the current difficulty level of the content expression dimension, and can be used to indicate the difficulty level of the task content in at least one content expression dimension.

[0198] It should be noted that the target difficulty level can be different from or the same as the initial difficulty level of the task content.

[0199] It is understandable that different content expression dimensions have at least two difficulty levels. For example, the difficulty level corresponding to the vocabulary dimension can be from vocabulary difficulty Level-1 to Level-5, so that the target difficulty level to which the content expression dimension needs to be adjusted can be determined.

[0200] For example, taking a linear equation in two variables as an example, the original task content is as follows: Solve the system of equations: (1) 2x + 3y = 12 (2) xy=1 The system identifies the content expression dimension corresponding to the question, as well as the current difficulty level and corresponding adjustable range of the content expression dimension, as shown in Table 4 below:

[0201] Table 4 The first adjustment to the difficulty level of the content expression dimension, such as increasing the numerical complexity of the numerical dimension, results in the following adjusted task content: Solve the system of equations: (1) 2x + 3y = 12 (2) 1.5x + 2.25y = 9.75 The second adjustment to the difficulty level of the content expression dimension, such as lowering the difficulty levels of the expression method and number of steps dimension, yields the following adjusted task content: Solve the system of equations: (1) xy=1 (2) y=2 The comparison results after the two adjustments are shown in Table 5 below:

[0202] Table 5 Optionally, when the task content consists of objective questions, the target difficulty level corresponding to different content expression dimensions can be different or the same, for example: The target difficulty level corresponding to the numerical dimension includes the numerical difficulty level; The target difficulty level corresponding to the interference dimension includes the interference difficulty level; The target difficulty level corresponding to the thought-guiding dimension includes the induction difficulty level; The target difficulty level corresponding to the scenario dimension of the question stem includes the scenario difficulty level; The target difficulty level corresponding to the expression style dimension includes the expression difficulty level; The target difficulty level corresponding to the number of steps includes the step difficulty level; The target difficulty level corresponding to the dimension of graph / table usage includes the difficulty level of graph / table usage; The target difficulty level corresponding to the language trap dimension includes the language trap difficulty level.

[0203] Among them, the difficulty level of numerical problems can be assessed based on numerical complexity, the difficulty level of the situation can be assessed based on the presence or absence of a question stem, the difficulty level of the expression can be assessed based on the directness or indirectness of the expression method, the difficulty level of the steps can be assessed based on the degree of simplification or increase in the number of steps, the difficulty level of the distractors can be assessed based on the presence or absence of distractors, the degree of presence of distractors and / or the degree of confusion or ease of elimination of distractors, the difficulty level of using graphs / tables can be assessed based on the degree of addition and / or removal of graphs / tables, and the difficulty level of language traps can be assessed based on the degree of introduction and / or removal of language traps.

[0204] Optionally, when the task content is subjective, the target difficulty level corresponding to different content expression dimensions can be different or the same, for example: The target difficulty level corresponding to the literary style dimension includes the style difficulty level; The target difficulty level corresponding to the cultural adaptation dimension includes the cultural difficulty level; The target difficulty level corresponding to the vocabulary dimension includes the vocabulary difficulty level; The target difficulty level corresponding to the sentence structure dimension includes the sentence structure difficulty level; The target difficulty level corresponding to the dimension of allusion application includes the application difficulty level; The target difficulty level corresponding to the dimension of subjective expression includes the level of expression difficulty.

[0205] Among them, the difficulty level of style can be assessed based on the formality or colloquialism of the writing style; the difficulty level of culture can be assessed based on the presence and / or degree of cultural adaptation; the difficulty level of vocabulary can be assessed based on the vocabulary level; the difficulty level of sentence structure can be assessed based on the complexity or simplicity of the sentence structure; the difficulty level of application can be assessed based on the presence and / or degree of allusion; and the difficulty level of expression can be assessed based on the presence and / or degree of emotion / attitude.

[0206] It should be noted that, since the difficulty adjustment trend between the current difficulty level and the target difficulty level of the content expression dimension can be to increase by at least one difficulty level, decrease by at least one difficulty level, or remain unchanged, the content adjustment method and the degree of adjustment under the content adjustment method will vary when adjusting the task content based on the content expression dimension and the difficulty adjustment trend.

[0207] Compared with traditional generative AI, the output of the adjusted task content in the technical solution of this application is not freely generated, but constrained by the consistency or similarity of the content expression, such as knowledge points, examination dimensions, factual information, etc. The adjusted task content that matches the intention to be adjusted is generated through a pre-trained constrained optimization expression generation model, rather than a simple replacement of synonyms.

[0208] It is understandable that adjusting the difficulty of a specific content expression dimension is not a blind replacement, but rather an adjustment based on the intended purpose. That is, the system needs to ensure that the expression changes but the semantics remain unchanged, or the semantics remain largely unchanged, or the original question's testing purpose remains unchanged. Under the premise of multi-objective constraints, content adjustments are made to achieve the goal of personalized learning and teaching.

[0209] Specifically, the knowledge points, semantics, and other expressive content of the content to be adjusted can be extracted and input into the model as invariable constraints to ensure that the difficulty changes but the goal does not deviate, or in other words, the goal remains consistent, so as to generate the task content corresponding to the adjusted task content.

[0210] It is understood that responding to a content adjustment event related to task content, in order to determine at least one adjustment intent regarding the task content, may include: Display a dimension selection control on a graphical target user interface, wherein the dimension selection control includes at least one candidate content expression dimension based on task content; in response to a difficulty determination instruction for the dimension selection control, determine at least one content expression dimension from the candidate content expression dimensions, and the target difficulty level corresponding to the content expression dimension.

[0211] Among them, the candidate content expression dimension is all the classification perspectives in the task content, or in other words, the adjustable, system-defined, and target user-defined partial classification perspectives. This can be set as needed, and this article does not make specific limitations.

[0212] Understandably, the following steps are included before displaying the dimension selection control on the graphical target user interface: The system analyzes the task content to identify the classification perspectives included in the task content.

[0213] Specifically, large language models (such as GPT series models) can be used to analyze task content to obtain a classification perspective of task content, or the initial expression difficulty of task content in each candidate content expression dimension, that is, the current difficulty level of task content under that candidate content expression dimension.

[0214] Optionally, if the task content is a paper-based math test paper, the target user can select a specific objective question from the test paper as the task content by taking a photo: Title: Let the function be... Then the monotonically increasing interval of the function on the interval [-2, 2] is: A. [-2, -1] B.[-1,1] C.[1,2] D.[-2,0] After receiving the task content, the system identifies the adjustable candidate content expression dimensions and their corresponding current difficulty levels, as shown in Table 6 below:

[0215] Table 6 The core knowledge points tested in this question remain unchanged: the monotonicity of functions, and the determination of the maximum value of a function on an interval (extreme points, endpoints), i.e., using derivatives to find extreme values ​​and comparing function values. If the difficulty is increased to a higher level by selecting the interference dimension, and after constraining through the pre-trained model, the adjusted task content is as follows: Title: Let the function be... Then the monotonically increasing interval of the function on the interval [-2, 2] is: A. (- [,-1] B.[-1,0) C.(1, ) D.[-2, ] The comparison of the questions before and after the adjustment is shown in Table 7 below:

[0216] Table 7 If both the expression structure dimension and the distractor dimension are adjusted to a higher difficulty level, the adjusted task content can be obtained as follows: Problem: Given the function If a function is continuously differentiable on the closed interval [-2, 2], then which of the following descriptions of its monotonicity on this interval is correct? A. In monotonically increasing B. In monotonically decreasing C. In monotonically increasing D. In monotonically increasing The revised task content is more challenging in terms of expression structure, introducing continuous differentiable terms and more academic language. The difficulty of the distractors has also been adjusted to achieve the goal of setting up a confusion range and inducing memory errors.

[0217] Furthermore, to more clearly illustrate the implementation of this application's solution, we can take a high school physics multiple-choice question as an example. If the system recognizes that a target user has circled or selected a physics multiple-choice question on the smart learning machine, and the user believes they can answer the question but wants to increase the difficulty of the question when testing the same knowledge point, it determines that the difficulty should be increased from the perspective of distractors to adjust the task content: Question: An object falls freely from a height. Assuming negligible air resistance, which of the following statements about its falling process is correct? A. The object is always subject to gravity and air resistance. B. The object's velocity remains constant. C. The object's acceleration is always 9.8 m / s². D. The speed of an object is inversely proportional to time. This question tests two main knowledge points: the fundamental characteristics of free fall motion and the constancy of gravitational acceleration (ignoring air resistance). Based on the above intended adjustments, the new options for the adjusted task content could be: A. The object is only subject to gravity, and its direction is always upward. B. The object's initial velocity is zero, and its acceleration gradually decreases. C. The object's velocity continuously increases, its direction is downward, and its acceleration is constant. D. The object is only subject to gravity, and its acceleration is 0. The answer to this question remains C, but the distractors are more semantically complex, incorporating words such as direction, initial velocity, and gradual, making it easy for target users to mistakenly choose A (direction misleading) or B (mistaking velocity for acceleration change) under some subjective factors or other objective factors, thus increasing the test of the target users' understanding of concepts and their ability to use the elimination method.

[0218] If the difficulty is increased in terms of formal structure, the adjusted task content could be: Question: In an experiment, an object falls freely from rest. Ignoring air resistance, its motion data is observed and recorded. Which of the following physical descriptions of the experimental process is the most reasonable? A. The object is only subject to gravity, and its acceleration remains constant. B. Gravity and air resistance are in equilibrium, and the speed tends to be constant. C. Velocity increases with time, but acceleration decreases. D. The falling distance is directly proportional to the time, and the speed is constant. The adjusted task content incorporates experimental context and observer information, and requires users to select the most reasonable option from multiple nearly correct statements. This will simultaneously test and examine the target user's ability to distinguish test terms, understand language precision, and comprehend the physical essence.

[0219] For example, when teachers need to test students' mastery of a certain knowledge point at different learning stages, they can set questions of varying difficulty for that knowledge point at different learning stages as learning progresses. This tests the target user's understanding of the knowledge point. If the task content is a test question, the teacher uploads the task content by taking a picture and selects at least one question as the task content for adjustment. The system identifies multiple dimensions of the question, such as the problem dimension, formal structure dimension, clarity of the question stem, and variable dimension. If the target user increases the difficulty from the formal structure dimension, then: Task details: Problem: A car with a mass of 1000 kg travels at a constant speed of 20 m / s on a horizontal road. It brakes immediately in an emergency and comes to a stop within 50 m. Find: (1) The acceleration of the car; (2) The magnitude of the resistance experienced by the car during braking.

[0220] The core knowledge points for this question are the formula for uniformly decelerated linear motion and Newton's second law.

[0221] Adjusted task details: Question: An electric car weighing 1200kg is undergoing an emergency braking system test. When the emergency braking system is activated while the car is traveling at a constant speed of 18m / s on a level road, the car decelerates smoothly until it comes to a complete stop, with a braking distance of 40m.

[0222] (1) Find the braking acceleration; (2) Find the frictional force exerted by the ground on the car during this process; (3) If there are 3 additional passengers (70kg each) in the vehicle, please recalculate the friction.

[0223] The adjusted task content has more complex data settings and multivariate models. For example, there is additional information (passenger quality) that needs to be processed, and it also involves the impact of changes in quality on force. Although the core knowledge points remain the same, it places higher demands on students' expression dimensions, modeling ability, and variable analysis ability. It is suitable for learning scenarios that assess different abilities of students at different learning stages.

[0224] Understandably, teachers or systems can set the cognitive level and knowledge integration requirements of questions as needed, quickly generate questions of different difficulties, and accurately match teaching objectives. For example, in the same classroom, personalized questions can be provided for students with different ability levels, which can both cater to students with slower learning progress and challenge students with faster progress. In addition, the solution proposed in this application can also identify which dimensions students have cognitive blind spots in by adjusting their answer performance before and after the adjustment, such as problems in non-knowledge points such as weak sense of time and space or insufficient argumentation ability.

[0225] For example, under history questions, the system can generate questions focusing on different dimensions around the same knowledge point based on students' learning progress: If students have already mastered the knowledge points of "Zheng He's Voyages to the Western Ocean" and "Ming Dynasty's Overseas Relations" or related knowledge points, the system will focus on assessing students' ability to integrate time and space based on their learning progress. Therefore, the system can generate adjusted task content by combining the above knowledge points from the perspective of time and space integration. Question: Please compare and contrast Zheng He's voyages to the Western Ocean with Columbus's voyages in terms of purpose, impact, and historical significance.

[0226] If the focus is on examining the cognitive level of the target user, then the adjusted task content can be generated: Question: Do you agree that Zheng He's voyages to the Western Ocean did not have a lasting impact on China's maritime diplomacy? Please provide your argument based on historical facts.

[0227] Understandably, if the task is merely to assess the target user's basic understanding of the above knowledge points, then the adjusted task content can be generated: Question: Who was Zheng He? What did he do? What impact did he have on China? To comprehensively assess the above abilities, the adjusted task content can be generated: Question: Some believe Zheng He's voyages to the Western Ocean represent the pinnacle of Chinese diplomacy, while others see them as a waste of resources. What is your opinion? Please analyze this in the context of the Ming Dynasty's diplomatic history.

[0228] In summary, the generated and adjusted task content not only increases the difficulty from the cognitive level dimension, but also adjusts the difficulty to a higher level from the spatiotemporal integration dimension.

[0229] Understandably, through the above methods, the system can be personalized and intelligently adjusted from dimensions such as cognitive level (memory, analysis, evaluation), spatiotemporal integration (single event, comparison between China and the West), and expression mode. This is conducive to generating questions that are more in line with students' current abilities, so as to advance step by step. The expression goal remains unchanged before and after the adjustment, that is, it is still about the historical impact of Zheng He's voyages to the West.

[0230] The adjustable difficulty level of subjective history questions allows for more granular personalized teaching, breaking through the limitations of the traditional one-size-fits-all approach. It enables different questions for different students and different questions for the same student at different stages, while all questions test the same core knowledge, greatly improving teaching efficiency.

[0231] Based on the above core solution, a teaching experiment was conducted in a holiday experimental class (90 students). Using the controlled variable method, two groups (45 students in each group) of students with equal academic levels answered an equal number of questions. They completed tasks based on the overall difficulty adjustment of existing questions and tasks generated based on content adjustments. Multiple indicators, including improved ability recognition accuracy, improved response success rate, and reduced frustration rate, were obtained and compared. The experimental data are shown in Table 8 below.

[0232] Table 8 It should be noted that the overall difficulty adjustment plan uses questions that test the same knowledge points but can be used at three difficulty levels: junior high, intermediate, and advanced.

[0233] Pre-test refers to a test conducted on students before teaching intervention or system use to understand their initial mastery of a certain knowledge point; post-test refers to a test conducted on students after teaching intervention or system use to assess their improved mastery of the same knowledge point after training.

[0234] Understandably, in this application, the pre-test refers to all students taking a set of objective or subjective questions on the same knowledge point before they come into contact with the system, in order to understand their original level; the system intervention period refers to processing two sets of questions, one currently only available in three difficulty levels (junior, intermediate, and senior), and the other using the technical solution of this application.

[0235] The post-test refers to using another set of equivalent questions to test students' mastery of the same knowledge points. The questions are different from those in the pre-test, but the knowledge points tested are the same, and the formats are more diverse.

[0236] Then, the changes in scores between the pre-test and post-test were compared and analyzed, and the comparative data shown in Table 8 were obtained.

[0237] Understandably, experimental data shows that by flexibly controlling the difficulty of task content or explanations through dimensions such as vocabulary and structure, students can obtain learning content that aligns with their cognitive development while "keeping the knowledge points unchanged." Finally, targeted reinforcement and training of their weaknesses can achieve a more effective success rate in understanding compared to traditional static difficulty adjustments.

[0238] Furthermore, after the dimensions and difficulty are broken down in detail, it is possible to more accurately locate which type of expression students have difficulty understanding (such as causal relationship expression, complex structure), which has a significant effect on the accuracy of ability identification in the teaching system.

[0239] Most importantly, compared to the chaotic reasons for errors, the target users accurately identified their weaknesses after completing the task, which significantly reduced their sense of frustration and increased their willingness and acceptance to continue trying after making mistakes.

[0240] Under the same knowledge point task conditions, by using the adjusted task content generated by this application, compared with the traditional static difficulty division teaching mechanism, it can improve students' post-test accuracy by 14.8%, increase their willingness to try again after failure by 32.7%, and improve the accuracy of identifying students' weak points in learning ability by 23.2%, thus having higher teaching adaptability and cognitive diagnosis efficiency.

[0241] In some embodiments, the step "obtaining at least one first response from the target user regarding the task content before adjustment" may include: In response to a trigger event on the task content, the task content is displayed on the target graphical user interface; in response to a content response operation on the task content, at least one first response from the target user to the task content before adjustment is obtained.

[0242] Among them, the triggering event is used to instruct the target user to process the task content. For example, when the task content is an essay topic, the student writes an essay based on the essay topic.

[0243] In an optional implementation, the triggering event may be an event triggered after completing a learning task for a certain knowledge point, so that the task content corresponding to the knowledge point is used as the task content.

[0244] In an optional implementation, the triggering event may be an event triggered by clicking on a function, such as clicking on a task content page, displaying multiple candidate task contents on the task content page, and thus determining the task content from the candidate task contents. For example, the target user may further select from the candidate task contents, or the task content may be randomly determined from the candidate task contents.

[0245] To better implement the above methods, this application also provides an information identification device, which can be integrated into an electronic device, such as a computer device, which can be a terminal, server, or other device.

[0246] The terminal can be a mobile phone, tablet computer, smart Bluetooth device, laptop computer, personal computer, etc.; the server can be a single server or a server cluster composed of multiple servers.

[0247] For example, in this embodiment, the method of this application embodiment will be described in detail by taking the information identification device specifically integrated into the terminal as an example. This embodiment provides an information identification device, such as... Figure 2 As shown, the information identification device may include: The record acquisition module 201 is used to acquire the operation record of the target user, wherein the operation record includes at least one adjustment intention obtained by the target user in adjusting the task content, and the response results made by the target user regarding the task content before and after the content adjustment; User location module 202 is used to locate the weak points of the target user based on the adjustment intention and response results.

[0248] In some embodiments, the user positioning module 202 is specifically used for: Based on the response results and adjustment intentions, calculate at least one moderating effect value for the target user on the task content, so as to obtain the weakness results according to each moderating effect value.

[0249] In some embodiments, the adjustment intent includes at least one content expression dimension, which is used to indicate the direction of content adjustment for the task content; the user positioning module 202 is specifically used for: For each content expression dimension, the moderating effect value of the target user on the task content is calculated based on the response results; The weaknesses of the target users are determined based on the moderating effect values ​​corresponding to each content expression dimension.

[0250] In some embodiments, the adjustment intent also includes a difficulty adjustment trend corresponding to the content expression dimension, and the response result includes a first response to the task content before adjustment and a second response to the task content after adjustment; the user positioning module 202 is specifically used for: Determine the response conclusions for the first and second responses, whereby the response conclusions are used to indicate the scores of the target user's responses to the task content before and after the adjustment. Based on the content expression dimensions, difficulty adjustment trends, and response conclusions, the moderating effect values ​​of target users on task content under each content expression dimension are calculated.

[0251] Understandably, the response result also includes the response behavior of the first response content and the second response content. The user positioning module 202 is specifically used for: Based on each content expression dimension, difficulty adjustment trend, response conclusion, and response behavior, the moderating effect value of the target user on the task content under each content expression dimension is calculated.

[0252] It is understandable that the information recognition device also includes an adjustment module, which is specifically used for: Based on the moderating effect value, determine the target user's mastery of the target expression dimensions in the task content; In response to the content adjustment operation, the task content is adjusted based on at least one content expression dimension other than the target expression dimension to obtain the adjusted task content.

[0253] It is understandable that the response result includes the content of the first response and the content of the second response. The record acquisition module 201 is specifically used for: Obtain at least one first response from the target user regarding the task content before the adjustment; In response to a content adjustment event related to task content, determine at least one adjustment intent for the task content; The task content is adjusted based on the adjustment intention to obtain the adjusted task content, and at least one second response from the target user to the adjusted task content is obtained.

[0254] Understandably, the adjustment intent includes the content expression dimension and the corresponding difficulty adjustment trend. The record acquisition module 201 is specifically used for: Based on the content expression dimensions and difficulty adjustment trends, the task content is adjusted to obtain the adjusted task content, wherein the content expression of the task content before and after the adjustment remains consistent.

[0255] It is understandable that the user positioning module 202 is specifically used for: The adjustment intention and response results are input into the neural network model; By using a neural network model to process the adjustment intentions and response results, the results of the weaknesses can be obtained.

[0256] It is understandable that the information recognition device also includes a learning and recommendation module, which is specifically used for: Based on the results of weaknesses, learning recommendations for the target user are displayed on the graphical user interface.

[0257] As can be seen from the above, the information recognition device in this embodiment obtains the target user's operation record, wherein the operation record includes at least one adjustment intention obtained by the target user to adjust the task content, and the response result of the target user to the task content before and after the content adjustment; based on the adjustment intention and the response result, the target user's weakness result is located, thereby identifying the user's weakness result by adjusting the task content during the user's response to the task content, so as to meet the personalized learning needs of different users based on the user's weakness result.

[0258] Accordingly, this application also provides an electronic device, which can be a terminal, such as a smartphone, tablet computer, laptop computer, touch screen, game console, personal computer (PC), personal digital assistant (PDA), or other terminal device. Figure 3 As shown, Figure 3This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. The electronic device 300 includes a processor 301 with one or more processing cores, a memory 302 with one or more computer-readable storage media, and a computer program stored in the memory 302 and executable on the processor. The processor 301 and the memory 302 are electrically connected. Those skilled in the art will understand that the electronic device structure shown in the figure does not constitute a limitation on the electronic device, and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0259] The processor 301 is the control center of the electronic device 300. It connects various parts of the electronic device 300 through various interfaces and lines. By running or loading software programs and / or modules stored in the memory 302, and calling data stored in the memory 302, it performs various functions of the electronic device 300 and processes data, thereby monitoring the electronic device 300 as a whole.

[0260] In this embodiment, the processor 301 in the electronic device 300 loads the computer program corresponding to the process of one or more application programs into the memory 302 according to the following steps, and the processor 301 runs the application programs stored in the memory 302 to realize various functions: Obtain the target user's operation records, wherein the operation records include at least one adjustment intention obtained by the target user in adjusting the task content, and the response results made by the target user regarding the task content before and after the content adjustment; Based on the adjustment intent and response results, the weaknesses of the target users were identified.

[0261] Therefore, the electronic device 300 provided in this embodiment can bring the following technical effects: meet the personalized learning needs of different users.

[0262] For details on the implementation of each of the above operations, please refer to the previous examples, which will not be repeated here.

[0263] Optional, such as Figure 3 As shown, the electronic device 300 also includes: a touch display screen 303, a radio frequency circuit 304, an audio circuit 305, an input unit 306, and a power supply 307. The processor 301 is electrically connected to the touch display screen 303, the radio frequency circuit 304, the audio circuit 305, the input unit 306, and the power supply 307. Those skilled in the art will understand that... Figure 3 The electronic device structure shown does not constitute a limitation on the electronic device and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0264] The touch display screen 303 can be used to display a graphical user interface (GUI) and receive operation commands generated by the user interacting with the GUI. The touch display screen 303 may include a display panel and a touch panel. The display panel can be used to display information input by the user or information provided to the user, as well as various graphical user interfaces of the electronic device. These graphical user interfaces can be composed of graphics, text, icons, video, and any combination thereof. Optionally, the display panel can be configured using a liquid crystal display (LCD), organic light-emitting diode (OLED), or other similar technologies. The touch panel can be used to collect touch operations performed by the user on or near it (such as operations performed by the user using a finger, stylus, or any suitable object or accessory on or near the touch panel), generate corresponding operation commands, and execute the corresponding program according to the operation commands. Optionally, the touch panel may include two parts: a touch detection device and a touch controller. The touch detection device detects the user's touch location and the signal generated by the touch operation, transmitting the signal to the touch controller. The touch controller receives touch information from the touch detection device, converts it into touch point coordinates, and sends it to the processor 301. It can also receive and execute commands from the processor 301. The touch panel can cover the display panel. When the touch panel detects a touch operation on or near it, it transmits the information to the processor 301 to determine the type of touch event. Subsequently, the processor 301 provides corresponding visual output on the display panel based on the type of touch event. In this embodiment, the touch panel and the display panel can be integrated into the touch display screen 303 to achieve input and output functions. However, in some embodiments, the touch panel and the touch display screen 303 can be implemented as two independent components to achieve input and output functions. That is, the touch display screen 303 can also be used as part of the input unit 306 to achieve input functions.

[0265] The radio frequency circuit 304 can be used to transmit and receive radio frequency signals to establish wireless communication with network devices or other electronic devices, and to transmit and receive signals with network devices or other electronic devices.

[0266] Audio circuitry 305 can be used to provide an audio interface between a user and an electronic device via a speaker and a microphone. Audio circuitry 305 converts received audio data into electrical signals, transmits them to the speaker, and the speaker converts them into sound signals for output. Conversely, the microphone converts collected sound signals into electrical signals, which are then received by audio circuitry 305, converted back into audio data, and then processed by processor 301 before being transmitted via radio frequency circuitry 304 to, for example, another electronic device, or output to memory 302 for further processing. Audio circuitry 305 may also include an earphone jack to facilitate communication between peripheral headphones and electronic devices.

[0267] The input unit 306 can be used to receive input numbers, characters, or user characteristic information (such as fingerprints, iris, facial information, etc.), and to generate keyboard, mouse, joystick, optical, or trackball signal inputs related to user settings and function control.

[0268] Power supply 307 is used to supply power to various components of electronic device 300. Optionally, power supply 307 can be logically connected to processor 301 through a power management system, thereby enabling functions such as charging, discharging, and power consumption management through the power management system. Power supply 307 may also include one or more DC or AC power supplies, charging systems, power fault detection circuits, power converters or inverters, power status indicators, and other arbitrary components.

[0269] although Figure 3 As not shown in the diagram, the electronic device 300 may also include a camera, sensor, wireless fidelity module, Bluetooth module, etc., which will not be described in detail here.

[0270] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0271] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be performed by a computer program, or by a computer program controlling related hardware. The computer program can be stored in a computer-readable storage medium and loaded and executed by a processor.

[0272] Therefore, embodiments of this application provide a computer-readable storage medium storing multiple computer programs that can be loaded by a processor to execute any of the information recognition methods provided in embodiments of this application. For example, the computer program can perform the following steps: Obtain the target user's operation records, wherein the operation records include at least one adjustment intention obtained by the target user in adjusting the task content, and the response results made by the target user regarding the task content before and after the content adjustment; Based on the adjustment intent and response results, the weaknesses of the target users were identified.

[0273] As can be seen, the computer program can be loaded by the processor to execute any of the information recognition methods provided in the embodiments of this application, thereby bringing about the following technical effects: meeting the personalized learning needs of different users.

[0274] For details on the implementation of each of the above operations, please refer to the previous examples, which will not be repeated here.

[0275] The computer-readable storage medium may include: read-only memory (ROM), random access memory (RAM), disk or optical disk, etc.

[0276] Since the computer program stored in the computer-readable storage medium can execute any of the information identification methods provided in the embodiments of this application, the beneficial effects that any of the information identification methods provided in the embodiments of this application can achieve can be realized. For details, please refer to the previous embodiments, which will not be repeated here.

[0277] The above provides a detailed description of an information identification method, apparatus, electronic device, and computer-readable storage medium provided in the embodiments of this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. An information identification method, characterized in that, The method includes: Obtain the operation records of the target user, wherein the operation records include at least one adjustment intention obtained by the target user to adjust the task content, and the response results made by the target user to the task content before and after the content adjustment; Based on the stated adjustment intent and the response results, the weaknesses of the target user were identified.

2. The method as described in claim 1, characterized in that, The results of locating the target user's weaknesses based on the adjustment intention and the response results include: Based on the response results and the adjustment intention, at least one moderating effect value of the target user on the task content is calculated, so as to obtain the weakness result according to each of the moderating effect values.

3. The method as described in claim 2, characterized in that, The adjustment intent includes at least one content expression dimension, which indicates the direction of content adjustment for the task content; calculating at least one moderating effect value of the target user on the task content based on the response result and the adjustment intent, and obtaining the weakness result based on the moderating effect value, includes: For each of the aforementioned content expression dimensions, the moderating effect value of the target user on the task content is calculated based on the response results; The weaknesses of the target user are determined based on the moderating effect value corresponding to each of the content expression dimensions.

4. The method as described in claim 3, characterized in that, The adjustment intention also includes a difficulty adjustment trend corresponding to the content expression dimension, and the response result includes a first response to the task content before the adjustment and a second response to the task content after the adjustment. The calculation of the target user's moderating effect value on the task content based on the response results includes: Determine the response conclusions of the first response content and the second response content, wherein the response conclusions are used to indicate the score of the target user's response results to the task content before and after the adjustment; Based on each of the aforementioned content expression dimensions, the aforementioned difficulty adjustment trend, and the aforementioned response conclusion, the adjustment effect value of the target user on the task content under each of the aforementioned content expression dimensions is calculated.

5. The method as described in claim 4, characterized in that, The response result also includes the response behavior of the first response content and the second response content. The calculation of the moderating effect value of the target user on the task content based on the response result includes: Based on each of the aforementioned content expression dimensions, the aforementioned difficulty adjustment trend, the aforementioned response conclusion, and the aforementioned response behavior, the moderating effect value of the target user on the task content under each of the aforementioned content expression dimensions is calculated.

6. The method as described in claim 3, characterized in that, The method further includes: Based on the moderating effect value, determine the target expression dimension that the target user has mastered in the task content; In response to the content adjustment operation, the task content is adjusted based on at least one content expression dimension other than the target expression dimension to obtain the adjusted task content.

7. The method as described in claim 1, characterized in that, The response result includes the content of the first response and the content of the second response. Obtaining the target user's operation records includes: Obtain at least one first response from the target user regarding the task content before the adjustment; In response to a content adjustment event for the task content, determine at least one adjustment intent for the task content; The task content is adjusted based on the stated adjustment intention to obtain the adjusted task content, and at least one second response from the target user to the adjusted task content is obtained.

8. The method as described in claim 7, characterized in that, The adjustment intent includes a content expression dimension and a difficulty adjustment trend corresponding to the content expression dimension. Adjusting the task content based on the adjustment intent to obtain the adjusted task content includes: Based on the content expression dimension and the difficulty adjustment trend, the task content is adjusted to obtain the adjusted task content, wherein the content expression of the task content before and after the adjustment remains consistent.

9. The method as described in claim 1, characterized in that, The results of locating the target user's weaknesses based on the adjustment intention and the response results include: The adjustment intention and the response result are input into the neural network model; The neural network model is used to process the adjustment intention and the response result to obtain the weakness result.

10. The method according to any one of claims 1 to 9, characterized in that, After identifying the target user's weaknesses based on the adjustment intent and the response results, the method further includes: Based on the results of the identified weaknesses, learning recommendations for the target user are displayed on the graphical user interface.

11. An information identification device, characterized in that, The device includes: The record acquisition module is used to acquire the operation records of the target user, wherein the operation records include at least one adjustment intention obtained by the target user in adjusting the task content, and the response results made by the target user to the task content before and after the content adjustment; The user identification module is used to locate the target user's weaknesses based on the adjustment intention and the response result.

Images