Personalized learning path generation method and system

Abstract

The invention relates to the technical field of personalized learning, in particular to a personalized learning path generation method and system, and the method comprises the following steps: collecting data when a learner learns a target concept, so as to determine a causal association type; calculating to obtain a cognitive state evaluation value of the learner; synchronously determining and outputting a learning path of the next stage; wherein the step of determining the learning path comprises the sub-steps of selecting an adaptive living example for a learner according to a cognitive cause indicated by the causal association type, and increasing or decreasing the familiarity level of the living example according to the cognitive state evaluation value; meanwhile, an experiment task is matched for the learner, and the complexity level of the experiment task is correspondingly adjusted according to the cognitive state evaluation value and the cognitive demand indicated by the causal association type. According to the method and the device, personalized learning path generation can be realized from a cognitive source by identifying and correcting life analogy errors which cause operation errors.

Description

Technical Field

[0001] This application relates to the technical field of personalized learning, and in particular to a method and system for generating personalized learning paths. Background Technology

[0002] In the field of personalized learning technology, systems and methods for dynamically planning and adapting content and sequence to learner differences have become an important direction for the development of educational technology. This technology has significant application value, especially in the teaching of scientific concepts in the lower grades of primary school. Students at this stage are in a critical period of transition from the preoperational stage to the concrete operational stage, and their thinking is characterized by a strong concreteness, relying primarily on concrete images and direct experience for cognitive construction. For abstract concepts, students often need to use familiar life scenarios for analogy to achieve initial understanding. Therefore, effectively integrating and guiding this analogical learning based on life experience in learning path generation has become a core element in improving the concept learning effectiveness of lower-grade students. Existing personalized learning systems generally collect learners' explicit behavioral data, such as answer accuracy, task completion time, and operational errors, and rely on algorithmic models to conduct ability assessments and recommend learning paths, aiming to achieve the ideal goal of individualized instruction.

[0003] However, despite some progress in behavioral data analysis and path recommendation, existing methods, when dealing with abstract concept learning scenarios for lower elementary school students, can only perceive and respond to the explicit results of students' operational behaviors, such as the correctness of experimental steps, but are completely unable to identify and intervene in their implicit cognitive understanding process, especially failing to address the cognitive root cause of operational errors, namely, incorrect life analogies. This deficiency is not a simple functional insufficiency, but stems from a fundamental disconnect between the design logic of existing technologies and the cognitive patterns of lower elementary school students. Specifically, when learning abstract concepts such as "simple circuits," students naturally use personal life experiences to make analogies to aid understanding, such as comparing "current" to "water flow," "wind," or "electricity in a toy flashlight." The familiarity with the life scenarios upon which these analogies are based directly affects the effectiveness of the analogies and the correctness of the mental models constructed by students. A key problem that often arises in actual teaching is that students' experimental operational errors are often not due to insufficient operational skills, but rather directly caused by inherent incorrect analogical associations. For example, if a student incorrectly compares "current" to "wind," believing it to be invisible and easily dissipated, they might misunderstand that "current will find its own path" when connecting circuits, leading to a confusion in the order of wire connections. Current systems can only record the result of "incorrect wire connection" and may mechanically recommend more circuit connection exercises or repetitive explanations of concepts based on this. This path adjustment is essentially "blind" because it fails to recognize the cognitive behavior of "students using wind as an analogy."

[0004] Therefore, the paths generated by the system can only inefficiently repeat superficial operational proficiency, failing to implement targeted cognitive correction. As a result, students' incorrect analogies persist, and corresponding operational errors recur. The system, however, continuously adjusts at a superficial level, significantly diminishing the effectiveness of personalized guidance and failing to achieve the learning goal of promoting concept understanding from its cognitive roots.

[0005] In summary, the core flaw of existing personalized learning path generation technologies lies in their lack of perception and association of students' cognitive understanding processes, especially their ability to make analogies and associations in life, at the logical level. They rely solely on behavioral outcome data to make path decisions, which makes it impossible to achieve truly personalized learning that corrects deviations from the cognitive root and effectively deepens the understanding of concepts when dealing with the abstract concept learning of lower elementary school students. Summary of the Invention

[0006] In order to generate personalized learning paths from the cognitive root by identifying and correcting real-life analogy errors that lead to operational mistakes, this application provides a method and system for generating personalized learning paths.

[0007] Firstly, this application provides a personalized learning path generation method, which adopts the following technical solution: A personalized learning path generation method includes the following steps: Collect the analogy descriptions provided by learners when learning target concepts, the familiarity level of the life scenarios corresponding to the analogy descriptions, and the operational results of the learners in related experimental tasks, and analyze the analogy descriptions and operational results to determine the type of causal relationship between the two; Based on the analogy description, the familiarity level of the life scenario, and the operation result, the learner's cognitive state assessment value is calculated; Based on the cognitive state assessment value and the causal association type, the learning path for the next stage is simultaneously determined and output. Determining the learning path includes: selecting suitable life examples for the learner based on the cognitive causes indicated by the causal association type, with the familiarity level of the life examples being increased or decreased according to the cognitive state assessment value; simultaneously, matching experimental tasks for the learner, with the complexity level of the experimental tasks being adjusted accordingly based on the cognitive state assessment value and the cognitive needs indicated by the causal association type, so that the selected life examples and the matched experimental tasks form a synergistic intervention targeting the cognitive causes.

[0008] Optionally, the analysis of the analogy description and the operation result to determine the type of causal relationship between them includes: Based on the familiarity level of the life scenario corresponding to the analogy description, and the operation errors identified from the operation results, predefined causal rules are matched to determine the causal association type; The causal association types include at least: a first type, which corresponds to an operational error directly caused by an incorrect analogy description associated with a low level of familiarity; and a second type, which corresponds to an operational error caused by a correct analogy description associated with a high level of familiarity failing to translate into correct operation.

[0009] Optionally, the calculation of the learner's cognitive state assessment value includes: Based on the causal association type, the analogy accuracy score corresponding to the analogy description, the familiarity score corresponding to the familiarity level of the life scene, and the operation conversion score corresponding to the operation result are obtained from the predefined evaluation rules. A pre-configured weighted fusion rule is used to synthesize the analogy accuracy score, familiarity score, and operation conversion score into the cognitive state assessment value; wherein, the weighted fusion rule is configured to assign the highest weight factor to the familiarity score.

[0010] Optionally, the step of simultaneously determining and outputting the learning path for the next stage based on the cognitive state assessment value and the causal association type includes: If the cognitive state assessment value is lower than the first threshold and the causal association type is the first type, then the following collaborative adjustment is performed to generate the learning path: the life scene associated with the current analogy description is replaced with a life instance with the highest level of familiarity, while the complexity level of the experimental task is adjusted to the predefined lowest level.

[0011] Optionally, the step of simultaneously determining and outputting the learning path for the next stage based on the cognitive state assessment value and the causal association type further includes: If the cognitive state assessment value is between the first threshold and the second threshold and the causal association type is the second type, then the following association adjustment is performed to generate the learning path: maintain the life instance currently associated with the analogy description, adjust the complexity level of the experimental task to a predefined intermediate level, decompose the experimental task into multiple sequential operation stages, and configure operation guidelines for each operation stage that are logically associated with the analogy description. The second threshold is greater than the first threshold.

[0012] Optionally, the step of simultaneously determining and outputting the learning path for the next stage based on the cognitive state assessment value and the causal association type further includes: If the cognitive state assessment value is higher than the second threshold, the following cognitive adjustments are performed to generate the learning path: the familiarity level of the life instance is reduced, or the attribute dimensions of the life instance associated with the current analogy description are expanded; at the same time, the complexity level of the experimental task is increased, and abstract symbolic representations are introduced in the experimental task to partially replace or assist the specific physical representations.

[0013] Optionally, after outputting the learning path for the next stage, a path iteration optimization step is also included: Obtain feedback data after the learner executes the learning path. The feedback data includes at least: the updated analogy description and its corresponding familiarity level with the life scene, the operational accuracy in subsequent experimental tasks, and the test results of the conceptual understanding of the target concept. Based on the feedback data, update the cognitive state assessment value and the causal association type; Based on the updated cognitive state assessment value and causal association type, the step of synchronously determining the learning path based on the cognitive state assessment value and causal association type is executed again to generate a new learning path for the next stage.

[0014] Secondly, this application provides a personalized learning path generation system, which adopts the following technical solution: A personalized learning path generation system, comprising: The data acquisition and causal analysis module is used to collect learners' analogy descriptions, the familiarity level of the life scenarios corresponding to the analogy descriptions, and the operation results, and analyze the analogy descriptions and operation results according to predefined causal rules to determine the causal relationship type. The cognitive state assessment module is used to generate a cognitive state assessment value through weighted fusion calculation based on the analogy description, the level of familiarity with the life scene, and the operation result. The collaborative path generation module is used to synchronously generate and output the learning path for the next stage based on the cognitive state assessment value and the causal association type. The collaborative path generation module is configured to adjust the familiarity level of the life instance and match the complexity level of the experimental task according to the cognitive cause indicated by the causal association type, so as to form a collaborative intervention for the cognitive cause. The path optimization feedback module is used to update the cognitive state assessment value and causal association type based on the feedback data after the learner executes the learning path, and to trigger the collaborative path generation module to generate a new learning path.

[0015] In summary, this application includes the following beneficial technical effects: This method, by collecting and analyzing learners' analogical descriptions and their corresponding levels of familiarity with life scenarios and experimental operation results, can accurately identify the implicit cognitive root causes of operational errors, namely, incorrect life analogy associations. Based on the type of causal relationship, it can implement targeted corrections from the cognitive level, effectively solving the fundamental deficiency of existing technologies that can only respond to explicit behaviors and cannot reach and intervene in the cognitive understanding process. It achieves truly personalized guidance that corrects deviations from the source and promotes in-depth understanding of concepts.

[0016] Based on the calculated cognitive state assessment value and causal association type, this method synchronously and collaboratively adjusts the familiarity level of life examples and the complexity level of experimental tasks, so that life analogy guidance and practical task training form a two-dimensional intervention combination targeting specific cognitive causes. This overcomes the limitations of existing methods that can only recommend single points or surface content, and significantly improves the adaptation accuracy and learning efficiency of personalized paths.

[0017] After outputting the learning path, this method continuously updates the cognitive state assessment value and causal relationship type by collecting learner feedback data, and dynamically generates new adaptive paths based on the update results, forming a closed-loop iterative optimization mechanism. This allows the learning path to continuously evolve along with changes in learners' cognitive levels, fundamentally avoiding the problems of rigid path adjustments and inability to adapt to learners' cognitive development in the long term, as seen in existing systems. Attached Figure Description

[0018] Figure 1 It is a flowchart of the overall logical closed loop of personalized learning path generation; Figure 2 This is a flowchart of the logic for determining the type of causal relationship; Figure 3 It is a flowchart of path coordination adjustment based on cognitive state and causal type. Detailed Implementation

[0019] The following is in conjunction with the appendix Figure 1-3 This application will be described in further detail.

[0020] This application discloses a method for generating personalized learning paths. For example... Figure 1 As shown, a personalized learning path generation method is applicable to the learning scenarios of abstract scientific concepts for primary school students. Based on the concrete thinking theory of educational psychology and computer software technology, it achieves personalized guidance by accurately linking cognitive data and dynamically adjusting learning elements, thereby optimizing the learning path from the cognitive root.

[0021] First, it is necessary to complete the construction of the basic association database and the calibration of key parameters. All basic data are derived from the cognitive test experiment of 100 primary school students and the conclusions of educational psychology research. Basic library construction: 1. Familiarity Level Library for Everyday Scenarios: Used to quickly calibrate the familiarity level of analogical descriptions. L1 High-Frequency Familiarity: Personal items, daily family life; L2 Medium-Frequency Familiarity: School scenarios; L3 Low-Frequency Unfamiliarity: Industrial / Public scenarios; 2. Experiment Complexity - Cognitive Requirements Library, used to match different cognitive needs. C1 (lowest level): 1-component experiment; C2 (intermediate level): 2-component experiments; C3 (higher level): 3-component experiments + abstract symbols. 3. Concept-Analogy-Causality Rule Base: Predefined correct / incorrect analogy examples and causal determination criteria to provide a basis for causal association analysis.

[0022] Key parameter calibration: 1. Cognitive assessment scoring criteria, where analogy is scored as follows: 90 points for correct and 30 points for incorrect; L1=100 points, L2=80 points, L3=50 points; 2. The conversion rate of an operation is assigned a value based on the matching degree between the analogy and the operation; 3. Weighted fusion rule weights, among which familiarity is 0.4, analogy accuracy is 0.3, and operation conversion is 0.3. Since the cognition of lower grade students relies on life experience, familiarity has the highest weight. 4. Cognitive thresholds, with the first threshold at 60 points and the second threshold at 85 points, are calibrated through student learning effect tracking experiments to accurately distinguish between three states: cognitive deviation, need for improvement, and achievement of standards.

[0023] This method includes the following steps: S1 Data Collection and Causal Relationship Type Determination like Figure 1 As shown, this step is based on the basic association library and calibrated parameters built in the basic preparation stage. By collecting learners' core cognitive and operational data, it analyzes the intrinsic relationship between analogical descriptions and operational results, providing basic data support for subsequent cognitive state assessment.

[0024] S11 Multi-dimensional Data Acquisition The system collects three types of core data step by step through a software interface adapted to the cognitive characteristics of students in grades 1-3 of primary school.

[0025] S111 Collection Analog Description After learners have an initial exposure to target concepts such as "simple circuits," the system uses an interactive software interface to pose guided questions, such as, "What in everyday life do you think the core element of the target concept (such as electric current) resembles? Please describe it in one sentence." This encourages learners to provide analogies. For example, when learners describe "electric current" as "like wind" or "like water flowing from a pipe in your house," the system records these analogies in real time and stores them in the data acquisition module.

[0026] S112 collects familiarity levels of everyday life scenarios After learners provide analogy descriptions, the system displays multiple-choice questions about relevant life scenarios through an interactive software interface. The question options are designed based on the familiarity level classification rules of life scenarios in the basic association library, covering examples corresponding to L1, L2, and L3 scenarios. For example, if a learner's analogy description is "electric current is like water flowing in a pipe," the multiple-choice options would be: "1. The wires of a toy flashlight, 2. The water pipes connected to a faucet at home, 3. The water pipes of a water dispenser at school, 4. Industrial pipes seen on television." Based on the learner's selections and the scenario classification standards in the basic association library, the system labels the corresponding familiarity level of the life scenario. Specifically, L1 level scenarios are labeled based on the daily contact frequency statistics of 100 students, with learners encountering these scenarios ≥3 times per day; L2 level scenarios are encountered 1-2 times per day; and L3 level scenarios are encountered ≤1 time per week. This classification aligns with the actual life experiences of lower-grade students, ensuring the rationality of the familiarity level labeling.

[0027] S113 Collects Experimental Results The system pushes experimental tasks related to the target concept to learners. Based on the initial operational ability test results of 100 students, level C2 is suitable for the initial operational ability of most learners. Therefore, the initial complexity of the experimental tasks is set to the middle level of C1-C3 in the basic association library, i.e., level C2. During the learners' execution of the experimental tasks, the system records key data of the experimental operation in real time through the software's built-in operation monitoring module, including the completeness of the operation steps, the correctness of the component connection sequence, and the specific type of erroneous operation. All operation data is stored in a structured form, providing traceable operational evidence for subsequent causal association analysis.

[0028] S12 Causal Relationship Type Analysis like Figure 1 and 2 As shown, the system invokes the concept-analogy-causality rule library built during the basic preparation phase. Taking the familiarity level of the life scenario, the accuracy of the analogy description, and the error type in the operation result collected in S11 as input, it determines the causal relationship type between the analogy description and the operation result by matching predefined causal rules. This process can accurately locate the cognitive causes leading to operation errors, avoiding the blindness of existing technologies that only adjust paths based on operation results.

[0029] S121 matches predefined causal rules The system first preprocesses the data collected by S11: based on the concept-analogy-causality rule base, it determines the accuracy of the analogy descriptions; for example, "current is like water flowing in a pipe" is determined to be a correct analogy, while "current is like wind" is determined to be an incorrect analogy. It extracts the specific category of the familiarity level of the life scenario, i.e., L1, L2, or L3; and it identifies clear operational errors from the operation results, such as "incorrectly connecting the negative terminal of the battery" or "disordered wire connection." Subsequently, the system matches the three types of preprocessed data one by one with the predefined causal association conditions in the rule base, filtering out causal association types that match the current learner's data characteristics.

[0030] S122 Determines the first type When the system matches all of the following conditions, the causal association type is determined to be Type I: the analogy is judged as an incorrect analogy, the familiarity level of the corresponding life scenario is L3, and there is a clear operational error in the operation result. The core characteristic of this type is that "low-familiarity incorrect analogy directly leads to operational error." For example, if a learner compares "current" to "the wind in an industrial pipe seen on TV" (incorrect analogy), the corresponding life scenario familiarity level is L3 (interaction ≤ 1 time per week). When performing a circuit connection experiment, if an operational error occurs such as "connecting wires arbitrarily and not distinguishing between positive and negative terminals," the system will determine it to be Type I. This determination process can directly locate the cognitive root of "incorrect analogy in an unfamiliar scenario." Compared with existing technologies that only focus on the operational result of "incorrect wire connection," it can more accurately identify the essence of cognitive bias.

[0031] S123 determines the second type. When the system matches all of the following conditions, the causal association type is determined to be Type II: the analogy description is judged to be a correct analogy, the corresponding life scenario familiarity level is L1 or L2, and there is a clear operational error in the operation result. The core characteristic of this type is "highly familiar correct analogy not translated into correct operation." For example, if a learner compares "current" to "water flow from a tap at home" (correct analogy), the corresponding life scenario familiarity level is L1 (daily contact ≥3 times), and an operational error occurs during the circuit connection experiment, such as "the switch is connected to the negative terminal of the battery, and the circuit cannot be controlled," the system will determine it to be Type II. This determination process can identify the problem of "disconnect between cognition and operation," providing a clear direction for subsequent strengthening of the conversion association between analogy and operation, and making up for the deficiency of existing technologies in being unable to connect cognition and operation.

[0032] S2 Cognitive Status Assessment Value Calculation like Figure 1As shown, this step takes the analogy descriptions collected in S1, the level of familiarity with life scenarios, the results of experimental operations, and the causal relationship type determined in S1 as the core basis. Through multi-dimensional score extraction and weighted fusion, the learner's cognitive state assessment value is calculated to objectively reflect their cognitive level of the target concept.

[0033] S21 Base Score Acquisition Based on the causal relationship type determined in S12, and combined with the predefined evaluation rules (which originated from a cognitive test experiment of 100 students in grades 1-3 and were optimized through three rounds of teaching verification), the system extracts the analogy accuracy score, familiarity score, and operation conversion score from the three types of data collected in S1, ensuring that each score is directly related to the core data in S1.

[0034] S211 Extraction of Analogy Accuracy Score Based on the accuracy assessment of the analogy description in S12, the system extracts the corresponding analogy accuracy score from predefined evaluation rules. 90 points are awarded when the analogy description is judged correct, and 30 points are awarded when it is judged incorrect. This scoring system is based on experimental data, which shows that correct analogies contribute approximately three times more to conceptual understanding than incorrect analogies. The difference between 90 and 30 points effectively distinguishes the cognitive value of the two types of analogies.

[0035] S212 Extract Familiarity Score The system extracts corresponding familiarity scores from predefined assessment rules based on the familiarity levels of life scenarios labeled in S112. When the familiarity level is L1, 100 points are extracted; at L2, 80 points are extracted; and at L3, 50 points are extracted. This gradient score design accurately quantifies the auxiliary role of life scenarios in cognitive construction.

[0036] S213 Extraction Operation Conversion Score Based on the causal relationship type determined in S12 and the experimental operation results recorded in S113, the system judges the matching degree between the analogy description and the operation result, and then extracts the corresponding operation conversion score from the predefined evaluation rules. Specifically, 100 points are extracted when both the analogy description and the operation result are correct; 60 points are extracted when the analogy description is correct but the operation result is incorrect; 30 points are extracted when both the analogy description and the operation result are incorrect; and 40 points are extracted when the analogy description is incorrect but the operation result is correct. Experimental follow-up revealed that the matching degree between analogy and operation directly affects the stability of concept understanding. Correctly matched operations contribute significantly more to cognitive deepening than mismatched operations, and this score setting is based on this conclusion.

[0037] S22 Weighted Fusion Calculation The system employs a pre-configured weighted fusion rule to synthesize the analogy accuracy score, familiarity score, and operational conversion score extracted by S21 into a cognitive state assessment value. This weighted rule is derived from an analysis of the cognitive characteristics of 100 students in grades 1-3. Experimental data shows that familiarity with everyday scenarios has the highest impact on the effectiveness of analogies among lower-grade students, with a correlation coefficient of 0.72, higher than the 0.58 for analogy accuracy and the 0.55 for operational conversion. Therefore, the familiarity score is assigned the highest weight factor.

[0038] S221 Determine the weighting factors The weighting factors for each score in the weighted fusion rule are clearly defined as follows: familiarity score weighting factor 0.4, analogy accuracy score weighting factor 0.3, and operational conversion score weighting factor 0.3. This weighting allocation prioritizes the supporting role of real-life scenarios in cognition, while also taking into account the correctness of analogical logic and operational conversion ability, which aligns with the cognitive pattern of lower-grade students transitioning from concrete experience to abstract concepts.

[0039] S222 Calculate cognitive state assessment value The cognitive state assessment score is calculated using the following formula: Cognitive State Assessment S = Analogy Accuracy Score × 0.3 + Familiarity Score × 0.4 + Operational Conversion Score × 0.3. For example, if a learner provides a correct analogy description in S1 (corresponding to an analogy accuracy score of 90) and associates it with a Level L1 life scenario (corresponding to a familiarity score of 100), but makes an error in the experimental operation (corresponding to an operation conversion score of 60), then their cognitive state assessment score S = 90 × 0.3 + 100 × 0.4 + 60 × 0.3 = 27 + 40 + 18 = 85. This result accurately reflects the learner's state of "correct cognitive logic but insufficient operational conversion," providing a clear quantitative direction for subsequent path adjustments and avoiding path recommendation biases caused by the single assessment dimension in existing technologies.

[0040] S3 Learning Path Generation like Figure 1 and Figure 3 As shown, this step uses the cognitive state assessment value calculated by S2 and the causal relationship type determined by S1 as the core basis, and simultaneously adjusts the familiarity level of life examples and the complexity level of experimental tasks to generate targeted personalized learning paths.

[0041] S31 Cognitive Bias Scenario Path Generation When a learner's cognitive state assessment score is below 60 points, this threshold was determined through a learning effect tracking experiment of 100 students in grades 1-3. The data showed that 87% of students with scores below this value had clear cognitive biases and needed to correct the erroneous cognitions at their root. When the causal relationship type was the first type, that is, the low familiarity error analogy directly led to the operational error, the system performed coordinated adjustments around the two core objectives of correcting the error analogy and reducing the cognitive load, and generated an appropriate learning path.

[0042] S311 Real-life Examples Adjustment The system first extracts the learner's current incorrect analogy description from S111, along with the associated L3-level life scenario, and then replaces it with an L1-level life instance defined in the basic preparation phase. For example, if the learner's original analogy description is "electricity is like wind" (associated with an industrial pipeline scenario on TV, belonging to L3 level), the system will replace the life instance with "electricity is like the electricity in your toy flashlight" or "electricity is like the electricity in your mother's charging cable," while simultaneously displaying a guiding prompt through the software interface: "Electricity is like the 'small flow of energy' in a toy flashlight, which can make the light bulb light up, and it's different from the 'wind' you described." This adjustment reinforces the correct analogy cognition through highly familiar scenarios, fundamentally reducing the interference of incorrect associations on subsequent operations.

[0043] S312 Experimental Task Adjustment The system lowers the complexity level of the initial experimental tasks in S113 to the C1 level (the lowest complexity level, containing only one component) defined in the basic preparation phase. For example, if the original experimental task was a three-component operation involving "battery + switch + light bulb + wire," the system simplifies it to a one-component experiment involving only "battery + light bulb." Simultaneously, it guides learners to disassemble their toy flashlight and observe the direct connection method of "battery positive terminal → light bulb bottom → battery negative terminal." The software also provides real-time annotations such as "electricity 'flows' from the positive terminal to the light bulb, just like the electricity in the flashlight makes the light bulb light up." This simplification strongly associates the operational process with the newly replaced L1 level real-life examples, avoiding cognitive overload caused by repetitive complex experiments in existing technologies, and helping learners establish a positive association between "correct analogy" and "simple operation."

[0044] S32 Cognitive Improvement Scenario Path Generation When learners' cognitive state assessment scores are between 60 and 85, experimental data show that 91% of students in this range can understand the correct analogy, but cannot translate cognition into actual operation. When the causal relationship type is type II, the system adjusts the relationship between maintaining the cognitive foundation and strengthening the operation transformation to generate an appropriate learning path.

[0045] S321 Real-life Examples Adjustment The system retains the learner's current correct analogy description in S111, as well as the L1 or L2 level life scenario associated with that description, without changing the life instance. For example, when the learner's analogy description is "electric current is like water flowing in the pipes at home," the system continues to use that life instance to avoid disrupting the established correct analogy cognition due to a change in scenario.

[0046] S322 Experiment Task Adjustment The system adjusts the complexity level of the experimental task to C2 level (intermediate complexity level, including experiments with two components: battery, switch, and light bulb) as defined in the basic preparation phase. It also breaks down the experimental task into multiple sequential operation stages, configuring operational guidance for each stage with a logical association with the correct analogy. For example, the system breaks down the experiment into two stages: "Battery positive terminal → Switch → Light bulb" and "Light bulb → Battery negative terminal." In the first stage, the software prompts, "The switch is like the handle of a faucet; it needs to be installed at the starting point of the 'water flow' (battery positive terminal) to control the flow." In the second stage, the software prompts, "Now send the 'water flow' back to the ending point (battery negative terminal), so the 'water flow' can cycle and light up the light bulb." This breakdown, through the logical mapping of analogy and operation, helps learners transform correct understanding into practical operation, overcoming the deficiency of existing technologies in connecting understanding and operation.

[0047] S33 Cognitive Achievement Scenario Path Generation When learners' cognitive status assessment scores are above 85, experimental data show that 89% of students above this threshold have mastered the basic content of the target concept and have the ability to upgrade to abstract cognition. The system performs cognitive adjustments around expanding cognitive dimensions and promoting abstract leaps, generating suitable learning paths.

[0048] S331 Real-life Examples Adjustment The system employs two methods to adjust real-life examples: First, it upgrades the current L1-level real-life example to an L2-level one, such as replacing "a water tap at home" with "a light bulb in a school classroom," while guiding learners to understand that "the switch for the classroom light is like a water tap at home, controlling the 'small current' of electricity for the entire classroom." Second, it expands the attribute dimensions of the current L1-level real-life examples. For instance, regarding the example of "water pipes at home," the system adds the prompt, "Water pipes at home come in different thicknesses, just like the difference between one and two batteries; two batteries have a 'thicker' current, making the light bulb brighter." These two adjustments, by expanding cognitive dimensions, avoid the limitations of existing technology that remains at the basic cognitive level, helping learners deepen their understanding of the concepts.

[0049] S332 Experimental Task Adjustment The system elevates the complexity level of experimental tasks to C3 level (a higher complexity level, including experiments with three components: battery, switch, light bulb, and wires) as defined in the basic preparation stage. Simultaneously, it introduces abstract symbols to represent these components, aiding in concrete physical operations. For example, the system first displays an animation of the transformation from "physical circuit" to "simple symbol drawing," labeling the battery with "+" and "-" symbols, the switch with "□" symbols, and the light bulb with "○" symbols. It then guides learners to draw the complete circuit using simple symbols, with the software prompting, "The '□' you drew represents a switch, and like a physical switch, it controls a small current." This adjustment facilitates the transition from concrete physical operations to abstract symbolic cognition, helping learners gradually build abstract conceptual thinking.

[0050] S4 path iteration optimization like Figure 1 As shown, this step is based on the learning path generated by S3. By collecting feedback data after the learner executes the path, the cognitive state assessment value and causal relationship type are updated, and a new adapted path is generated, forming a closed-loop mechanism of "path generation-feedback-optimization".

[0051] S41 Feedback Data Acquisition The system collects three types of core feedback data based on the learner's performance after executing the learning path output by S3.

[0052] Episode S411 Updated Analog Descriptions and Familiarity Levels The system uses the same guided questioning as S111, such as "What in life do you think electric current is like now?", to collect the learner's updated analogy description. Then, using a multiple-choice question format consistent with S112, the learner confirms the corresponding life scenario for the analogy description. This is then combined with the familiarity level standards determined during the basic preparation phase: L1 level is ≥3 times per day, L2 level is 1-2 times per day, and L3 level is ≤1 time per week, to label the updated familiarity level of the life scenario. For example, after executing the S3 path, if the learner's updated analogy description is "electric current is like the electricity in a classroom light," and the selected scenario is "school classroom," the system would label their familiarity level as L2.

[0053] S412 data acquisition accuracy of subsequent experimental operations The system pushes subsequent experimental tasks to learners that match the complexity level adjusted in S3. For example, if S3 adjusts an experiment to level C2, it will push circuit experiments of the same level. Using the same operation monitoring module as S113, it records the total number of experimental operations and the number of correct operations. The operation accuracy rate is then calculated using the formula: "Operation Accuracy Rate = Number of Correct Operations / Total Number of Operations × 100%". For example, if a learner completes 5 experimental operations and 4 of them are correct, the system calculates the operation accuracy rate as 80%. This indicator directly reflects the effect of the S3 path on improving the learner's operational conversion ability.

[0054] S413 Collects Conceptual Understanding Test Results The system designs five core comprehension questions around the target concept, such as "Why is the switch connected to the positive terminal of the battery?" and "Where does the current start from in the battery and where does it return?" Each question is worth 2 points, for a maximum of 10 points. Learners answer these questions through an interactive software interface. The system uses predefined scoring criteria, such as "being able to accurately explain the principle of how a switch controls the flow of current" for 2 points and "knowing that a switch works but not being able to explain the principle" for 1 point. The system then calculates the concept comprehension test results, which can supplement and verify changes in cognitive state, avoiding the one-sidedness of relying solely on operational data to assess cognitive level.

[0055] S42 Cognitive Data Update Based on the feedback data collected in S41, the system calculates a new cognitive state assessment value using the cognitive assessment rules in S2. At the same time, it reiterates the causal relationship type with reference to the causal determination criteria in S1, ensuring that the logic of updating cognitive data is completely consistent with the previous steps and avoiding deviations caused by fluctuations in assessment criteria.

[0056] S421 Updated Cognitive Status Assessment Values The system extracts updated analogy accuracy, familiarity, and operational conversion scores from the feedback data collected in S41: 90 points are extracted for correct analogy descriptions and 30 points for incorrect ones, consistent with the standard in S211; 100 points are extracted for familiarity levels L1, 80 points for L2, and 50 points for L3, consistent with the standard in S212; the operational conversion score is determined based on the matching between the analogy and the operation, for example, 100 points for both correct analogy and correct operation, exactly the same as the standard in S213. The system then uses the weighted fusion rule determined in S22: a weight of 0.4 for familiarity score, 0.3 for analogy accuracy score, and 0.3 for operational conversion score, to calculate a new cognitive state assessment value. For example, if a learner's updated analogy description is correct (90 points), their familiarity level is L2 (80 points), and their operational accuracy is 100% (100 points), the new cognitive state assessment value S = 90 × 0.3 + 80 × 0.4 + 100 × 0.3 = 27 + 32 + 30 = 89 points.

[0057] S422 Reiterates Types of Causal Relationships Referring to the causal determination rules in S12, and combining the updated analogy description, familiarity level, and operational accuracy collected in S41, the system reiterates the causal relationship type. For example, if the learner's updated analogy description is correct, the familiarity level is L2, and the operational accuracy is 100% with no errors, the system determines that there is no clear causal relationship type. If the learner's updated analogy description is correct, the familiarity level is L1, but the operational accuracy is 60% and there are errors, the system determines it to be the second type, that is, a highly familiar and correct analogy has not been converted into a correct operation. This determination provides a clear cognitive causal basis for the subsequent generation of new paths.

[0058] S43 New Learning Path Generation Based on the updated cognitive state assessment value and causal relationship type in S42, the system fully adopts the path generation logic of S3, and synchronously adjusts the familiarity level of life examples and the complexity level of experimental tasks to generate a new learning path for the next stage.

[0059] S431 adjusts logic based on updated cognitive value matching path If the updated cognitive state assessment score is below 60 points and the causal relationship type is Type I, the system performs the same collaborative adjustment as in S31: replacing the life instance with an L1 level and downgrading the experimental task to a C1 level. If the assessment score is between 60 and S < 85 points and the causal relationship type is Type II, the system performs the same association adjustment as in S32: maintaining highly familiar life instances, adjusting the experimental task to a C2 level and breaking down the operational steps. If the assessment score is ≥ 85 points, the system performs the same cognitive adjustment as in S33: appropriately reducing the familiarity level of the life instance or expanding the attribute dimensions, upgrading the experimental task to a C3 level and introducing abstract symbols. For example, if the learner's updated cognitive state assessment score is 89 points (≥ 85 points), the system will adjust the life instance from a L2 level to a L3 level, upgrade the experimental task from a C2 level to a C3 level, and introduce a standard circuit symbol diagram to assist in the operation.

[0060] S432 outputs a new learning path The system integrates the adjusted real-life examples and experimental tasks into a new learning path, clearly presented through the software interface. The presentation format is consistent with the path output by S3, including prompts based on real-life examples and guidance on experimental steps. For example, the new path prompts, such as "Now let's learn about simplified electrical piping in a factory (Level 3 scenario). Try drawing the circuit using symbols like '+', '□', and '○', and then connect it with physical components," ensuring learners can clearly understand and execute the new path, continuously driving their cognitive development to higher levels.

[0061] The implementation principle of a personalized learning path generation method in this application embodiment is as follows: The implementation principle of this personalized learning path generation method is to identify the cognitive root causes of operational errors (such as incorrect analogical associations) by collecting and analyzing learners' analogical descriptions, familiarity levels of life scenarios, and experimental operation results. Based on this, the cognitive state assessment value and causal association type are dynamically calculated, and the familiarity level of life examples and the complexity level of experimental tasks are adjusted simultaneously to form a collaborative intervention path targeting cognitive causes. This process not only realizes the perception and guidance of the learning process from the cognitive level rather than just the behavioral level, but also continuously optimizes the path through iterative feedback, thereby effectively solving the problem that existing systems only rely on explicit operational data and cannot identify and correct implicit cognitive errors, truly realizing the promotion of students' deep understanding and operational transformation of abstract concepts from the cognitive root causes.

[0062] This embodiment also discloses a personalized learning path generation system, which is suitable for abstract scientific concept learning scenarios for students in grades 1-3 of primary school. It includes a data collection and causal analysis module, a cognitive state assessment module, a collaborative path generation module, and a path optimization feedback module.

[0063] The data acquisition and causal analysis module uses a software interface adapted to the cognitive characteristics of lower-grade students to collect learners' analogical descriptions, the familiarity level of the corresponding life scenarios, and the operational results of related experimental tasks. The acquisition logic is consistent with the multi-dimensional data acquisition in S11 of the aforementioned method. Simultaneously, the module calls a predefined concept-analogy-causal rule library, matching causal rules and determining the causal association type based on the collected familiarity levels and identified operational errors. The analysis logic completely corresponds to the causal association type analysis in S12 of the aforementioned method.

[0064] The cognitive state assessment module receives the analogy description, familiarity level of the life scene, operation result, and causal relationship type output by the data collection and causal analysis module, and adopts the cognitive assessment rules of S2 in the aforementioned method. The module first extracts the analogy accuracy score, familiarity score, and operation conversion score, and then calculates and generates the learner's cognitive state assessment value according to the pre-configured weighted fusion rules, which quantitatively reflects the learner's real-time cognitive level.

[0065] The collaborative path generation module uses the cognitive state assessment value output by the cognitive state assessment module and the causal relationship type determined by the data collection and causal analysis module as a basis to simultaneously adjust the familiarity level of life examples and the complexity level of experimental tasks. Following the path generation logic of S3 in the aforementioned method, the module performs collaborative adjustment, correlation adjustment, or cognitive adjustment for three scenarios: cognitive bias, cognitive need for improvement, and cognitive achievement, respectively, generating learning paths that adapt to both life examples and experimental tasks, thus forming a collaborative intervention on the causes of cognition.

[0066] The path optimization feedback module collects feedback data from learners after they execute the path output by the collaborative path generation module. This includes updated analogy descriptions and corresponding familiarity levels, accuracy rates of subsequent experimental operations, and concept comprehension test results. The data collection dimensions are consistent with those of the feedback data collection in S41 of the aforementioned method. Based on the feedback data, the module updates the cognitive state assessment value using the evaluation rules of S2, reiterates the causal relationship type with reference to the causal determination criteria of S1, and triggers the collaborative path generation module to execute the path generation logic again, outputting a new adapted learning path to achieve closed-loop optimization of the path.

[0067] The system is supported by the basic association library and key parameters established in the basic preparation phase. It completes data input and cognitive cause positioning through the data collection and causal analysis module, realizes cognitive level quantification through the cognitive state assessment module, outputs the initial personalized path through the collaborative path generation module, and then continuously iterates and optimizes through the path optimization feedback module to ensure that the learning path always matches the learner's real-time cognitive needs, effectively solving the shortcomings of existing systems that rely only on explicit behavioral data and blindly adjust paths.

[0068] The above are all preferred embodiments of this application, and are not intended to limit the scope of protection of this application. Therefore, all equivalent changes made in accordance with the structure, shape and principle of this application should be covered within the scope of protection of this application.

Claims

1. A method for generating personalized learning paths, characterized in that, Includes the following steps: Collect the analogy descriptions provided by learners when learning target concepts, the familiarity level of the life scenarios corresponding to the analogy descriptions, and the operational results of the learners in related experimental tasks, and analyze the analogy descriptions and operational results to determine the type of causal relationship between the two; Based on the analogy description, the familiarity level of the life scenario, and the operation result, the learner's cognitive state assessment value is calculated; Based on the cognitive state assessment value and the causal association type, the learning path for the next stage is simultaneously determined and output. Determining the learning path includes: selecting suitable life examples for the learner based on the cognitive causes indicated by the causal association type, with the familiarity level of the life examples being increased or decreased according to the cognitive state assessment value; simultaneously, matching experimental tasks for the learner, with the complexity level of the experimental tasks being adjusted accordingly based on the cognitive state assessment value and the cognitive needs indicated by the causal association type, so that the selected life examples and the matched experimental tasks form a synergistic intervention targeting the cognitive causes.

2. The method according to claim 1, characterized in that, The analysis of the analogy description and the operation result to determine the type of causal relationship between the two includes: Based on the familiarity level of the life scenario corresponding to the analogy description, and the operation errors identified from the operation results, predefined causal rules are matched to determine the causal association type; The causal association types include at least: a first type, which corresponds to an operational error directly caused by an incorrect analogy description associated with a low level of familiarity; and a second type, which corresponds to an operational error caused by a correct analogy description associated with a high level of familiarity failing to translate into correct operation.

3. The method according to claim 2, characterized in that, The calculation yields the learner's cognitive state assessment value, including: Based on the causal association type, the analogy accuracy score corresponding to the analogy description, the familiarity score corresponding to the familiarity level of the life scene, and the operation conversion score corresponding to the operation result are obtained from the predefined evaluation rules. A pre-configured weighted fusion rule is used to synthesize the analogy accuracy score, familiarity score, and operation conversion score into the cognitive state assessment value; wherein, the weighted fusion rule is configured to assign the highest weight factor to the familiarity score.

4. The method according to claim 2, characterized in that, The step of simultaneously determining and outputting the learning path for the next stage based on the cognitive state assessment value and the causal relationship type includes: If the cognitive state assessment value is lower than the first threshold and the causal association type is the first type, then the following collaborative adjustment is performed to generate the learning path: the life scene associated with the current analogy description is replaced with a life instance with the highest level of familiarity, while the complexity level of the experimental task is adjusted to the predefined lowest level.

5. The method according to claim 4, characterized in that, The step of simultaneously determining and outputting the learning path for the next stage based on the cognitive state assessment value and the causal relationship type also includes: If the cognitive state assessment value is between the first threshold and the second threshold and the causal association type is the second type, then the following association adjustment is performed to generate the learning path: maintain the life instance currently associated with the analogy description, adjust the complexity level of the experimental task to a predefined intermediate level, decompose the experimental task into multiple sequential operation stages, and configure operation guidelines for each operation stage that are logically associated with the analogy description. The second threshold is greater than the first threshold.

6. The method according to claim 5, characterized in that, The step of simultaneously determining and outputting the learning path for the next stage based on the cognitive state assessment value and the causal relationship type also includes: If the cognitive state assessment value is higher than the second threshold, the following cognitive adjustments are performed to generate the learning path: the familiarity level of the life instance is reduced, or the attribute dimensions of the life instance associated with the current analogy description are expanded; at the same time, the complexity level of the experimental task is increased, and abstract symbolic representations are introduced in the experimental task to partially replace or assist the specific physical representations.

7. The method according to claim 1, characterized in that, After outputting the learning path for the next stage, the process also includes a path iteration optimization step: Obtain feedback data after the learner executes the learning path. The feedback data includes at least: the updated analogy description and its corresponding familiarity level with the life scene, the operational accuracy in subsequent experimental tasks, and the test results of the conceptual understanding of the target concept. Based on the feedback data, update the cognitive state assessment value and the causal association type; Based on the updated cognitive state assessment value and causal association type, the step of synchronously determining the learning path based on the cognitive state assessment value and causal association type is executed again to generate a new learning path for the next stage.

8. A personalized learning path generation system, characterized in that, include: The data acquisition and causal analysis module is used to collect learners' analogy descriptions, the familiarity level of the life scenarios corresponding to the analogy descriptions, and the operation results, and analyze the analogy descriptions and operation results according to predefined causal rules to determine the causal relationship type. The cognitive state assessment module is used to generate a cognitive state assessment value through weighted fusion calculation based on the analogy description, the level of familiarity with the life scene, and the operation result. The collaborative path generation module is used to synchronously generate and output the learning path for the next stage based on the cognitive state assessment value and the causal association type. The collaborative path generation module is configured to adjust the familiarity level of the life instance and match the complexity level of the experimental task according to the cognitive cause indicated by the causal association type, so as to form a collaborative intervention for the cognitive cause. The path optimization feedback module is used to update the cognitive state assessment value and causal association type based on the feedback data after the learner executes the learning path, and to trigger the collaborative path generation module to generate a new learning path.

Images