Preschool education visual interaction method and system

Abstract

The invention provides a preschool education visual interaction method and system, and the method comprises the steps: obtaining the initial interaction data of a preschool child, analyzing the operation complexity and structural features, constructing an initial multi-dimensional cognitive state vector, analyzing the continuous trajectory of the child in the interaction process, decomposing the continuous trajectory into behavior sequences of exploration, delay, correction and confirmation elements, and carrying out the recognition of the initial multi-dimensional cognitive state vector. Calculating a behavior primitive transition probability matrix, updating a cognitive state vector according to the matrix and a calculation gain value, when the cyclic probability formed by the delayed primitive and the correction primitive exceeds a preset threshold value, reconstructing a task visual element in a non-prompt manner, and based on the updated cognitive state vector, calculating the cognitive state vector. And calculating a cognitive coupling degree between the state vector and a task preset cognitive load vector in a teaching task library, and selecting a task of which the coupling degree is in a target interval as a next round of interaction task, thereby realizing personalized teaching recommendation.

Description

Technical Field

[0001] This application belongs to the field of visualization and interactive technology, and in particular relates to a visualization and interactive method and system for preschool education. Background Technology

[0002] Gamification and visualization can attract the attention of preschool children and stimulate their interest in learning. Most interactive preschool education applications are based on pre-set teaching content and difficulty levels, and the assessment of children's learning outcomes relies on explicit, outcome-based indicators such as task completion accuracy and reaction time. These methods, to some extent, ignore the individual differences in children's cognitive development and struggle to understand their true thought processes and cognitive states during problem-solving. For example, it's impossible to reliably distinguish whether a child truly understands the knowledge point after exploratory attempts or simply guesses the answer by chance. When tasks are too difficult, they can dampen children's enthusiasm, while tasks that are too easy fail to provide reliable intellectual stimulation, making it difficult to achieve truly personalized education.

[0003] User modeling techniques offer coarse-grained analysis of children's behavior, typically focusing on simple statistics of operational sequences without delving into the underlying cognitive intentions. For example, they fail to identify cognitive processes such as hesitation and correction exhibited by children during interactions. When children fall into a cognitive trap, displaying "cognitive fixedness" by repeatedly trying incorrect paths, existing systems often lack reliable intervention mechanisms. Current systems primarily provide direct prompts or answers, which, while helpful in completing the current task, may inhibit the development of children's independent exploration and problem-solving abilities. Furthermore, existing systems often make rough adjustments to difficulty based on historical performance, failing to establish a precise matching model between children's cognitive state and the cognitive load of the teaching task. This results in recommended tasks often not falling precisely within children's zone of proximal development, thus impacting overall teaching efficiency. Therefore, a method is urgently needed that can analyze children's interactive behavior, assess and update their cognitive state in real time, and provide non-prompting interventions. Summary of the Invention

[0004] This invention proposes a visual interactive method for preschool education to address the limitations of existing technologies in deeply understanding children's true thought processes and cognitive states during problem-solving, and in exploring the cognitive intentions behind their behaviors. The method includes the following steps:

[0005] A multimodal exploration space is presented to obtain initial interaction data of preschool children. By analyzing the operational complexity and structural features of the data, an initial multidimensional cognitive state vector representing the child's current cognitive level is constructed.

[0006] The continuous interaction trajectory of children in teaching tasks is obtained, the trajectory is decomposed into a sequence of behaviors consisting of exploration, hesitation, correction and confirmation primitives, and a transition probability matrix of behavior primitives is calculated based on the sequence of behaviors.

[0007] A gain value for updating the multidimensional cognitive state vector is calculated based on the behavioral primitive transition probability matrix; the multidimensional cognitive state vector is updated using the gain value; when the probability of a loop consisting of hesitant primitives and corrective primitives exceeds a preset cognitive fixation threshold, the visual elements of the teaching task are reconstructed non-cue.

[0008] Based on the updated multidimensional cognitive state vector, the cognitive coupling degree between the state vector and the preset multidimensional cognitive load vector of each teaching task in the teaching task library is calculated, and the teaching task with the cognitive coupling degree in the preset target range is selected as the next round of interactive task.

[0009] Optionally, constructing an initial multidimensional cognitive state vector representing the child's current cognitive level includes:

[0010] The dimensions of the multidimensional cognitive state vector are defined as spatial perception ability, logical reasoning ability, and hand-eye coordination ability.

[0011] For each dimension, an initial score is calculated based on the task completion time, number of erroneous operations, and operation path length in the initial interaction data, with a value range of 0 to 100.

[0012] The initial scores together constitute the initial multidimensional cognitive state vector.

[0013] Optionally, decomposing the trajectory into a sequence of behaviors consisting of exploration, hesitation, correction, and confirmation primitives includes:

[0014] Children's continuous dragging actions on the touch screen for more than 1.5 seconds and a distance of more than 100 pixels are identified as exploration primitives;

[0015] Behaviors that remain in the same area for more than 2 seconds without any target-oriented actions are identified as hesitant primitives.

[0016] The action of abandoning the current operation object and selecting a new operation object is identified as a modification primitive.

[0017] The operation of successfully placing the target object in the designated area and receiving correct feedback from the system is identified as the confirmation primitive.

[0018] Optionally, calculating a gain value for updating the multidimensional cognitive state vector based on the behavioral primitive transition probability matrix includes:

[0019] Extract the probability of a direct transition from an exploration primitive to a confirmation primitive from the behavior primitive transition probability matrix. ;

[0020] Extract the probability of shifting from a hesitant primitive to a corrective primitive. And the probability of shifting from a modified primitive to a hesitant primitive. ;

[0021] The gain value G is calculated using the following formula: ,in, and These are the preset weighting coefficients.

[0022] Optionally, updating the multidimensional cognitive state vector using the gain value includes:

[0023] Obtain the preset multidimensional cognitive load vector of the current teaching task. ;

[0024] The multidimensional cognitive state vector is updated using the following formula: ,in, This is the updated multidimensional cognitive state vector. Let G be the current multidimensional cognitive state vector, and G be the gain value. This is the preset learning rate.

[0025] Optionally, when the probability of a loop consisting of hesitation and correction primitives exceeds a preset cognitive fixation threshold, the visual elements of the teaching task are reconstructed non-cueably, including:

[0026] The probability of a cycle formed by hesitant and corrective primitives, i.e., the probability of transitioning from a hesitant primitive to a corrective primitive. The probability of shifting from a modified primitive to a hesitant primitive The product of these factors exceeds the preset cognitive fixation threshold. When this occurs, the reconstruction is triggered;

[0027] The reconstruction specifically involves reducing the color saturation of interfering elements in the current teaching task and randomly shifting the initial position of the target element to different areas of the screen without changing the shape and size of the target element.

[0028] Optionally, the step of calculating the cognitive coupling degree between the updated multidimensional cognitive state vector and the preset multidimensional cognitive load vector of each teaching task in the teaching task library includes:

[0029] Let the updated multidimensional cognitive state vector be denoted as... The preset multidimensional cognitive load vector of any teaching task in the teaching task database is denoted as... ;

[0030] The cosine similarity formula is used to calculate the cognitive coupling degree C between the two vectors.

[0031] Optionally, selecting teaching tasks with cognitive coupling within a preset target range as the next round of interactive tasks includes:

[0032] Set the preset target interval as ;

[0033] All cognitive coupling degrees C satisfying the following criteria are selected from the teaching task database. The teaching tasks;

[0034] From the selected tasks, one task is randomly chosen as the next round of interaction task.

[0035] Furthermore, this invention also relates to a preschool education visualization and interactive system, comprising the following modules:

[0036] The module is used to present a multimodal exploration space, acquire initial interaction data of preschool children, and construct an initial multidimensional cognitive state vector representing the child's current cognitive level by analyzing the operational complexity and structural features of the data.

[0037] The calculation module is used to acquire the continuous interaction trajectory of children in the teaching task, decompose the trajectory into a sequence of behaviors consisting of exploration, hesitation, correction and confirmation primitives, and calculate a transition probability matrix of behavior primitives based on the sequence of behaviors.

[0038] The update module is used to calculate a gain value for updating the multidimensional cognitive state vector based on the behavioral primitive transition probability matrix; update the multidimensional cognitive state vector using the gain value; and perform non-cue reconstruction of the visual elements of the teaching task when the probability of a loop consisting of hesitation primitives and correction primitives exceeds a preset cognitive fixation threshold.

[0039] The selection module is used to calculate the cognitive coupling degree between the updated multidimensional cognitive state vector and the preset multidimensional cognitive load vector of each teaching task in the teaching task library, and select the teaching task whose cognitive coupling degree is in the preset target range as the next round of interaction task.

[0040] Preferably, constructing the initial multidimensional cognitive state vector representing the child's current cognitive level includes:

[0041] The dimensions of the multidimensional cognitive state vector are defined as spatial perception ability, logical reasoning ability, and hand-eye coordination ability.

[0042] For each dimension, an initial score is calculated based on the task completion time, number of erroneous operations, and operation path length in the initial interaction data, with a value range of 0 to 100.

[0043] The initial scores together constitute the initial multidimensional cognitive state vector.

[0044] Preferably, the step of decomposing the trajectory into a sequence of behaviors consisting of exploration, hesitation, correction, and confirmation primitives includes:

[0045] Children's continuous dragging actions on the touch screen for more than 1.5 seconds and a distance of more than 100 pixels are identified as exploration primitives;

[0046] Behaviors that remain in the same area for more than 2 seconds without any target-oriented actions are identified as hesitant primitives.

[0047] The action of abandoning the current operation object and selecting a new operation object is identified as a modification primitive.

[0048] The operation of successfully placing the target object in the designated area and receiving correct feedback from the system is identified as the confirmation primitive.

[0049] Preferably, calculating a gain value for updating the multidimensional cognitive state vector based on the behavioral primitive transition probability matrix includes:

[0050] Extract the probability of a direct transition from an exploration primitive to a confirmation primitive from the behavior primitive transition probability matrix. ;

[0051] Extract the probability of shifting from a hesitant primitive to a corrective primitive. And the probability of shifting from a modified primitive to a hesitant primitive. ;

[0052] The gain value G is calculated using the following formula: ,in, and These are the preset weighting coefficients.

[0053] Preferably, updating the multidimensional cognitive state vector using the gain value includes:

[0054] Obtain the preset multidimensional cognitive load vector of the current teaching task. ;

[0055] The multidimensional cognitive state vector is updated using the following formula: ,in, This is the updated multidimensional cognitive state vector. Let G be the current multidimensional cognitive state vector, and G be the gain value. This is the preset learning rate.

[0056] Preferably, when the probability of a loop consisting of hesitation and correction primitives exceeds a preset cognitive fixation threshold, the visual elements of the teaching task are reconstructed non-cueably, including:

[0057] The probability of a cycle formed by hesitant and corrective primitives, i.e., the probability of transitioning from a hesitant primitive to a corrective primitive. The probability of shifting from a modified primitive to a hesitant primitive The product of these factors exceeds the preset cognitive fixation threshold. When this occurs, the reconstruction is triggered;

[0058] The reconstruction specifically involves reducing the color saturation of interfering elements in the current teaching task and randomly shifting the initial position of the target element to different areas of the screen without changing the shape and size of the target element.

[0059] Preferably, the step of calculating the cognitive coupling degree between the updated multidimensional cognitive state vector and the preset multidimensional cognitive load vector of each teaching task in the teaching task library includes:

[0060] Let the updated multidimensional cognitive state vector be denoted as... The preset multidimensional cognitive load vector of any teaching task in the teaching task database is denoted as... ;

[0061] The cosine similarity formula is used to calculate the cognitive coupling degree between the two vectors.

[0062] Preferably, selecting teaching tasks with cognitive coupling within a preset target range as the next round of interactive tasks includes:

[0063] Set the preset target interval as ;

[0064] All cognitive coupling degrees C satisfying the following criteria are selected from the teaching task database. The teaching tasks;

[0065] From the selected tasks, one task is randomly chosen as the next round of interaction task.

[0066] This invention decomposes children's continuous interaction trajectories into behavioral primitives such as exploration, hesitation, correction, and confirmation, achieving a refined insight into children's true thought processes and cognitive processes, surpassing traditional, coarse evaluation methods that rely on outcome-based indicators such as task accuracy. It constructs and updates a more accurate multidimensional cognitive state vector, objectively representing children's current ability level. Especially when a child is identified as being stuck in a cognitive fixed state, intervention through non-cue reconstruction of visual elements breaks the mental deadlock, avoids the harm to children's independent exploration caused by directly providing answers, and promotes the development of children's problem-solving abilities. By calculating the coupling degree between children's cognitive states and the cognitive load of teaching tasks to recommend subsequent tasks, it ensures that the teaching content accurately matches children's cognitive frontiers, thereby improving the reliability of teaching. Attached Figure Description

[0067] Figure 1 A flowchart of the first embodiment;

[0068] Figure 2 This is a schematic diagram of a multidimensional cognitive state vector representation. Detailed Implementation

[0069] Many specific details are set forth in the following description to provide a full understanding of this specification. However, this specification can be implemented in many other ways than those described herein, and those skilled in the art can make similar extensions without departing from the spirit of this specification. Therefore, this specification is not limited to the specific implementations disclosed below.

[0070] The terminology used in one or more embodiments of this specification is for the purpose of describing particular embodiments only and is not intended to be limiting of the one or more embodiments of this specification. The singular forms “a,” “described,” and “the” as used in one or more embodiments of this specification and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used in one or more embodiments of this specification refers to and includes any or all possible combinations of one or more associated listed items.

[0071] It should be understood that although the terms first, second, etc., may be used to describe various information in one or more embodiments of this specification, such information should not be limited to these terms. These terms are only used to distinguish information of the same type from one another. For example, first may also be referred to as second without departing from the scope of one or more embodiments of this specification, and similarly, second may also be referred to as first. Depending on the context, the word "if" as used herein may be interpreted as "when," "when," or "in response to a determination."

[0072] In the first embodiment, the present invention proposes a visual interactive method for preschool education, such as... Figure 1 This includes the following steps:

[0073] S1 presents a multimodal exploration space, acquires initial interaction data of preschool children, and constructs an initial multidimensional cognitive state vector representing the child's current cognitive level by analyzing the operational complexity and structural features of the data.

[0074] Specifically, the multimodal exploration space is a virtual block scene without fixed task objectives, where children can freely drag, rotate, and stack blocks of different shapes and colors. During this time, the child's touch or mouse trajectory sequence, timestamps, and all operation events on the block objects are continuously recorded. Operational complexity is represented by calculating the number of operations per unit time, the diversity of operation types, and the average operation sequence length. Structural features are evaluated by analyzing spatial attributes such as the height, symmetry, and color clustering of the stacked blocks. The operational complexity index is mapped to the attention concentration and operational finesse dimensions in a multidimensional cognitive state vector, and the structural feature index is mapped to the spatial imagination and logical planning dimensions, thus forming an initial cognitive state numerical vector that includes, for example, these four dimensions.

[0075] In an optional embodiment, constructing an initial multidimensional cognitive state vector representing the child's current cognitive level includes:

[0076] The dimensions of the multidimensional cognitive state vector are defined as spatial perception ability, logical reasoning ability, and hand-eye coordination ability.

[0077] For each dimension, an initial score is calculated based on the task completion time, number of erroneous operations, and operation path length in the initial interaction data, with a value range of 0 to 100.

[0078] The initial scores together constitute the initial multidimensional cognitive state vector.

[0079] A child's cognitive level can be abstracted into a three-dimensional vector, namely the cognitive state vector, with the three dimensions corresponding to spatial perception, logical reasoning, and hand-eye coordination, respectively. Each component of this vector is a value between 0 and 100, intuitively representing the child's level in the corresponding ability dimension. For example, a vector [85, 70, 90] indicates that the child performs well in spatial perception, has moderate logical reasoning ability, and excellent hand-eye coordination.

[0080] In the initial assessment phase, a series of benchmark tasks are provided, and children's interaction data is recorded to calculate the initial vector. Taking the spatial perception dimension as an example, suppose in a jigsaw puzzle task, the standard completion time is 30 seconds, the standard number of errors is 1, and the standard path length is 500 pixels. A child takes 45 seconds, makes 3 errors, and the path length is 700 pixels. Points are deducted based on the degree of deviation from the standards; for example, 0.5 points are deducted for every second exceeding the time limit, 10 points are deducted for each additional error, and 0.1 points are deducted for every 10 pixels the path exceeds the limit. Therefore, the child's initial spatial perception score is 70.5. By performing similar calculations and weighted averaging on data from different tasks, the initial scores for the three dimensions are obtained, collectively forming the initial multidimensional cognitive state vector, such as... Figure 2 .

[0081] S2, obtain the continuous interaction trajectory of the child in the teaching task, decompose the trajectory into a behavior sequence composed of exploration, hesitation, correction and confirmation primitives, and calculate a behavior primitive transition probability matrix based on the behavior sequence;

[0082] Specifically, taking a graphic matching teaching task as an example, the continuous interactive trajectory is the sequence of screen coordinates of a child dragging a shape object. When the trajectory shows rapid movement above multiple target slots without release, this segment is marked as an exploration primitive. When the trajectory's movement speed in a certain area is lower than a preset speed threshold and the dwell time exceeds a preset duration threshold, such as 1.5 seconds, it is marked as a hesitation primitive. When a child places a shape object in an incorrect slot and then picks it up again within a short time, this continuous action of placing and picking up is marked as a correction primitive. When a child successfully places a shape object in the correct slot and starts dragging the next shape object, this successful placement action is marked as a confirmation primitive. By segmenting and marking the trajectory throughout the entire task process, a complete sequence of behavioral primitives is obtained, such as exploration-exploration-hesitation-correction-hesitation-exploration-confirmation. All cases of jumping from one primitive to another in the sequence are statistically analyzed, constructing a 4×4 transition probability matrix, where the matrix elements... This represents the probability of transitioning from primitive i to primitive j.

[0083] In an optional embodiment, decomposing the trajectory into a sequence of behaviors consisting of exploration, hesitation, correction, and confirmation primitives includes:

[0084] Children's continuous dragging actions on the touch screen for more than 1.5 seconds and a distance of more than 100 pixels are identified as exploration primitives;

[0085] Behaviors that remain in the same area for more than 2 seconds without any target-oriented actions are identified as hesitant primitives.

[0086] The action of abandoning the current operation object and selecting a new operation object is identified as a modification primitive.

[0087] The operation of successfully placing the target object in the designated area and receiving correct feedback from the system is identified as the confirmation primitive.

[0088] A trajectory analysis module is provided that receives and processes raw input data streams from the touchscreen in real time. This data includes the coordinates, timestamps, and event types (e.g., graphic drag, click, or release) for each touch point. The module segments and categorizes these continuous data points according to preset rules, thereby transforming the child's entire operational trajectory into a discrete sequence composed of four behavioral primitives. This process does not rely on complex models but is achieved by setting explicit thresholds.

[0089] For example, in a shape matching task, a child presses and drags a circular block for 2 seconds, covering a distance of 250 pixels. Both the duration and distance of this action exceed a threshold, so it is recorded as an exploration primitive. The child drags the circular block near a square groove, stops moving but doesn't lift their finger off the screen, and remains there for 2.5 seconds; this is identified as a hesitation primitive. The child releases the circular block and instead clicks and drags a square block; this switching of objects is identified as a correction primitive. The child successfully places the square block into the square groove, providing a correct prompt sound; this operation is marked as a confirmation primitive. Thus, a complete attempt by the child is decoded as a sequence: exploration, hesitation, correction, and confirmation.

[0090] S3, calculate a gain value for updating the multidimensional cognitive state vector based on the behavioral primitive transition probability matrix; update the multidimensional cognitive state vector using the gain value; when the probability of a loop consisting of hesitant primitives and corrective primitives exceeds a preset cognitive fixation threshold, perform non-cue reconstruction of the visual elements of the teaching task.

[0091] Specifically, the formula for calculating the gain value can be set as follows: Gain value = Weight coefficient 1 multiplied by the transition probability from exploration to confirmation, minus weight coefficient 2 multiplied by the product of the transition probability from hesitation to correction and the transition probability from correction to hesitation. This gain value is multiplied by a preset learning rate and added to the cognitive dimension focused on by the current task. For example, for a picture matching task, the spatial imagination dimension is mainly updated, thereby updating the entire multidimensional cognitive state vector. At the same time, the cycle probability between hesitation and correction primitives is monitored, that is, the product of the transition probability from hesitation to correction and the transition probability from correction to hesitation. When this product value exceeds a set cognitive fixation threshold, such as 0.6, it is determined that the child is in a thinking dilemma. At this time, the task interface will be non-cue-basedly reconstructed, such as uniformly changing all the candidate colors or rotating the background pattern by 90°, to break the child's fixed thinking pattern with new visual stimuli.

[0092] In an optional embodiment, calculating a gain value for updating the multidimensional cognitive state vector based on the behavioral primitive transition probability matrix includes:

[0093] Extract the probability of a direct transition from an exploration primitive to a confirmation primitive from the behavior primitive transition probability matrix. ;

[0094] Extract the probability of shifting from a hesitant primitive to a corrective primitive. And the probability of shifting from a modified primitive to a hesitant primitive. ;

[0095] The gain value G is calculated using the following formula: ;in, and These are the preset weighting coefficients.

[0096] A first-order Markov model is used to represent a child's learning efficiency and confusion level in a task. Internally, a 4×4 behavioral primitive transition probability matrix is ​​maintained, where the rows and columns correspond to four primitives: exploration, hesitation, correction, and confirmation. Whenever a child completes a primitive operation and moves to the next primitive, the frequency count in the matrix is ​​updated, and the transition probability is recalculated. For example, if the previous primitive was exploration and the current primitive is confirmation, the count in the exploration row and confirmation column of the matrix is ​​incremented by one.

[0097] After completing a teaching task, the gain value is calculated based on the transition probability matrix generated during the task. Assume the weighting coefficients are set as follows: , The probability of directly transitioning from the exploration primitive to the confirmation primitive can be found from the matrix. A value of 0.7 represents the child's direct problem-solving ability. The probability of shifting from hesitation to correction. The probability is 0.3, indicating a shift from correction back to hesitation. The product of these two values ​​is 0.4, representing the degree to which the child gets caught in a cycle of ineffective attempts. Substituting these values ​​into the formula, the gain value G is calculated to be 0.93.

[0098] In an optional embodiment, updating the multidimensional cognitive state vector using the gain value includes:

[0099] Obtain the preset multidimensional cognitive load vector of the current teaching task. ;

[0100] The multidimensional cognitive state vector is updated using the following formula: ,in, This is the updated multidimensional cognitive state vector. Let G be the current multidimensional cognitive state vector, and G be the gain value. This is the preset learning rate.

[0101] The cognitive state vector is adjusted based on the child's performance on a specific task, i.e., the magnitude of the gain value G. The direction of the update is determined by the cognitive load vector of the currently completed task. This decision ensures that children's ability development is targeted. For example, if a task primarily tests spatial perception, then the dimension of the cognitive state vector corresponding to spatial perception will receive the most updates.

[0102] Assume the child's current cognitive state vector is The task we just completed was one that required a high level of logical reasoning and a moderate level of hand-eye coordination. The preset cognitive load vector for this task was... The calculated gain value is G=0.93, with a preset learning rate. Calculate the unit vector of the task load vector, with a magnitude of... Update the step size vector as follows: New cognitive state vector .

[0103] To break the cognitive fixation that children may exhibit, in an optional embodiment, when the probability of a loop consisting of hesitation and correction primitives exceeds a preset cognitive fixation threshold, the visual elements of the teaching task are reconstructed non-cueably, including:

[0104] The probability of a cycle formed by hesitant and corrective primitives, i.e., the probability of transitioning from a hesitant primitive to a corrective primitive. The probability of shifting from a modified primitive to a hesitant primitive The product of these factors exceeds the preset cognitive fixation threshold. When this occurs, the reconstruction is triggered;

[0105] The reconstruction specifically involves reducing the color saturation of interfering elements in the current teaching task and randomly shifting the initial position of the target element to different areas of the screen without changing the shape and size of the target element.

[0106] Continuously monitor the strength of specific cyclic patterns in the behavioral primitive transition probability matrix, the strength of which changes from hesitation to correction transition probability. With the transition probability adjusted to hesitation The product of these factors is used to represent cognitive fixation. A higher product value indicates a greater likelihood that the child is trapped in an ineffective cycle. A pre-set cognitive fixation threshold is also included. The value is 0.25. In one task, the child repeatedly tried to place a triangular block into a circular hole and then into a square hole, resulting in frequent alternations of hesitation and correction primitives. The value at this moment was calculated based on the interaction data. and The product of the two is 0.3, exceeding the threshold of 0.25, thus triggering the reconstruction mechanism. In addition to the target triangular block and triangular hole, the current interface also contains distracting circular and square blocks and holes. Reconstruction is immediately performed, reducing the color saturation of the circular and square blocks and holes from 100% to 40%, making them less visually prominent. The target triangular block is instantly moved from the top left corner of the screen to the bottom right corner; this abrupt change in position prompts the child to re-examine the entire task interface, potentially uncovering the correct solution previously overlooked.

[0107] S4. Based on the updated multidimensional cognitive state vector, calculate the cognitive coupling degree between the state vector and the preset multidimensional cognitive load vector of each teaching task in the teaching task library, and select the teaching task with the cognitive coupling degree in the preset target range as the next round of interaction task.

[0108] Specifically, the teaching task library contains hundreds of pre-stored teaching tasks. Each task is pre-labeled with a multi-dimensional cognitive load vector by educational experts, and the dimensions of this vector are completely consistent with the child's cognitive state vector. For example, in a simple color classification task, the cognitive load vector has a high value in the logical planning dimension and a low value in the spatial imagination dimension. Cognitive coupling is calculated using the cosine similarity of the two vectors. In another embodiment, cosine similarity and Euclidean distance are combined to calculate cognitive coupling, for example, by using their product or weighted sum as the cognitive coupling. The cosine similarity value between the child's currently updated cognitive state vector and the cognitive load vector of each task is calculated. A target range is preset, for example, 0.7 to 0.9, which represents a level of difficulty that is challenging yet achievable for the child. From all tasks whose cognitive coupling calculation results fall within this range, one task is randomly selected or selected according to a specific strategy as the next interactive task pushed to the child.

[0109] To match the next most suitable teaching task for a child, in an optional embodiment, the calculation of the cognitive coupling degree between the updated multidimensional cognitive state vector and the preset multidimensional cognitive load vector of each teaching task in the teaching task library includes:

[0110] Let the updated multidimensional cognitive state vector be denoted as... The preset multidimensional cognitive load vector of any teaching task in the teaching task database is denoted as... ;

[0111] The cognitive coupling degree C between the two vectors is calculated using the cosine similarity formula: .

[0112] The updated cognitive state vector of the child is considered as the location of their current ability in a multidimensional cognitive space. Simultaneously, a teaching task library is maintained, where each task is pre-labeled with a multidimensional cognitive load vector, representing the ability ratio required to complete the task. Cognitive coupling is calculated by using cosine similarity to represent the directional consistency between the child's current ability vector and the task requirement vectors.

[0113] Taking specific data as an example, the child's updated cognitive state vector is: The task pool contains two candidate tasks. Task A is a complex maze with a cognitive load vector of... This primarily tests spatial awareness. Task B is a sorting game with a cognitive load vector of... This primarily tests logical reasoning ability. Cognitive coupling is calculated separately. For task A, For task B, .

[0114] In an optional embodiment, selecting a teaching task with cognitive coupling within a preset target range as the next round of interaction task includes:

[0115] Set the preset target interval as ;

[0116] All cognitive coupling degrees C satisfying the following criteria are selected from the teaching task database. The teaching tasks;

[0117] From the selected tasks, one task is randomly chosen as the next round of interaction task.

[0118] A very high cognitive coupling degree, such as close to 1, indicates that the task's required abilities are highly matched to the child's current abilities, potentially making it too simple and lacking challenge. Conversely, a very low coupling degree suggests that the task may be too difficult and could discourage the child. Therefore, it is important to set an optimal cognitive coupling degree range, rather than simply choosing the task with the highest coupling degree.

[0119] Continuing with the example above, the target coupling range is set to [0.90, 0.98]. After calculating the coupling of all tasks, the coupling of task A is 0.88, which is lower than the lower limit of the range. Therefore, it is excluded. Task B has a coupling degree of 0.97, which falls exactly within this range. Assume there are also tasks C and D, with coupling degrees of 0.92 and 0.99 respectively. Then, task D is excluded because its coupling degree exceeds the upper limit. Also excluded. The set of candidate tasks that meet the criteria is Task B and Task C. To increase the diversity of teaching paths, one of them will not be fixed. Instead, a task will be selected from the candidate set with equal probability using a random number generator. For example, Task B will be selected as the interactive task presented to the children in the next round.

[0120] In a second embodiment, the present invention also provides a preschool education visualization and interactive system, comprising the following modules:

[0121] The module is used to present a multimodal exploration space, acquire initial interaction data of preschool children, and construct an initial multidimensional cognitive state vector representing the child's current cognitive level by analyzing the operational complexity and structural features of the data.

[0122] The calculation module is used to acquire the continuous interaction trajectory of children in the teaching task, decompose the trajectory into a sequence of behaviors consisting of exploration, hesitation, correction and confirmation primitives, and calculate a transition probability matrix of behavior primitives based on the sequence of behaviors.

[0123] The update module is used to calculate a gain value for updating the multidimensional cognitive state vector based on the behavioral primitive transition probability matrix; update the multidimensional cognitive state vector using the gain value; and perform non-cue reconstruction of the visual elements of the teaching task when the probability of a loop consisting of hesitation primitives and correction primitives exceeds a preset cognitive fixation threshold.

[0124] The selection module is used to calculate the cognitive coupling degree between the updated multidimensional cognitive state vector and the preset multidimensional cognitive load vector of each teaching task in the teaching task library, and select the teaching task whose cognitive coupling degree is in the preset target range as the next round of interaction task.

[0125] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, compact disc read-only memory (CD-ROM), optical storage, etc.) containing computer-usable program code.

[0126] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0127] The above description is merely an embodiment of this application and is not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principle of this application should be included within the scope of the claims of this application.

Claims

1. A pre-school education visual interactive method, characterized in that, Includes the following steps: A multimodal exploration space is presented to obtain initial interaction data of preschool children. By analyzing the operational complexity and structural features of the data, an initial multidimensional cognitive state vector representing the child's current cognitive level is constructed. The continuous interaction trajectory of children in teaching tasks is obtained, the trajectory is decomposed into a sequence of behaviors consisting of exploration, hesitation, correction and confirmation primitives, and a transition probability matrix of behavior primitives is calculated based on the sequence of behaviors. A gain value for updating the multidimensional cognitive state vector is calculated based on the behavioral primitive transition probability matrix. The multidimensional cognitive state vector is updated using the gain value; when the probability of a loop consisting of hesitant primitives and corrective primitives exceeds a preset cognitive fixation threshold, the visual elements of the teaching task are reconstructed non-cue. Based on the updated multidimensional cognitive state vector, the cognitive coupling degree between the state vector and the preset multidimensional cognitive load vector of each teaching task in the teaching task library is calculated, and the teaching task with the cognitive coupling degree in the preset target range is selected as the next round of interactive task.

2. The method of claim 1, wherein, The construction of the initial multidimensional cognitive state vector representing the child's current cognitive level includes: The dimensions of the multidimensional cognitive state vector are defined as spatial perception ability, logical reasoning ability, and hand-eye coordination ability. For each dimension, an initial score is calculated based on the task completion time, number of erroneous operations, and operation path length in the initial interaction data, with a value range of 0 to 100. The initial scores together constitute the initial multidimensional cognitive state vector.

3. The method of claim 1, wherein, The step of decomposing the trajectory into a sequence of behaviors consisting of exploration, hesitation, correction, and confirmation primitives includes: Children's continuous dragging actions on the touch screen for more than 1.5 seconds and a distance of more than 100 pixels are identified as exploration primitives; Behaviors that remain in the same area for more than 2 seconds without any target-oriented actions are identified as hesitant primitives. The action of abandoning the current operation object and selecting a new operation object is identified as a modification primitive. The operation of successfully placing the target object in the designated area and receiving correct feedback from the system is identified as the confirmation primitive.

4. The method of claim 1, wherein, The step of calculating a gain value for updating the multidimensional cognitive state vector based on the behavioral primitive transition probability matrix includes: The probability of directly switching from the exploration primitive to the confirmation primitive is positively correlated with the gain value, while the probability of forming a loop between the hesitant primitive and the correction primitive is negatively correlated with the gain value. extracting from the behavior primitive transition probability matrix a probability of directly transitioning from an exploration primitive to a confirmation primitive ; probability of shifting from a hesitation primitive to a correction primitive and probability of shifting from a correction primitive to a hesitation primitive ; The gain value G is calculated using a formula: ; wherein, and is a preset weight coefficient.

5. The method of claim 4, wherein, Updating the multidimensional cognitive state vector using the gain value includes: Obtain the preset multidimensional cognitive load vector of the current teaching task. ; The multidimensional cognitive state vector is updated using the following formula: ;in, This is the updated multidimensional cognitive state vector. Let G be the current multidimensional cognitive state vector, and G be the gain value. This is the preset learning rate.

6. The method according to claim 1, characterized in that, When the probability of a loop consisting of hesitation and correction primitives exceeds a preset cognitive fixation threshold, the visual elements of the teaching task are reconstructed non-cue, including: The probability of a cycle formed by hesitant and corrective primitives, i.e., the probability of transitioning from a hesitant primitive to a corrective primitive. The probability of shifting from a modified primitive to a hesitant primitive The product of these factors exceeds the preset cognitive fixation threshold. When this occurs, the reconstruction is triggered; The reconstruction specifically involves reducing the color saturation of interfering elements in the current teaching task and randomly shifting the initial position of the target element to different areas of the screen without changing the shape and size of the target element.

7. The method according to claim 1, characterized in that, The calculation of the cognitive coupling degree between the updated multidimensional cognitive state vector and the preset multidimensional cognitive load vector of each teaching task in the teaching task library, based on the updated multidimensional cognitive state vector, includes: Let the updated multidimensional cognitive state vector be denoted as... The preset multidimensional cognitive load vector of any teaching task in the teaching task database is denoted as... The cognitive coupling degree C between the two vectors is calculated using the cosine similarity formula.

8. The method according to claim 7, characterized in that, The selection of teaching tasks with cognitive coupling within a preset target range as the next round of interactive tasks includes: Set the preset target interval as ; All cognitive coupling degrees C satisfying the following criteria are selected from the teaching task database. The teaching tasks; From the selected tasks, one task is randomly chosen as the next round of interaction task.

9. A preschool education visualization and interactive system, characterized in that, Includes the following modules: The module is used to present a multimodal exploration space, acquire initial interaction data of preschool children, and construct an initial multidimensional cognitive state vector representing the child's current cognitive level by analyzing the operational complexity and structural features of the data. The calculation module is used to acquire the continuous interaction trajectory of children in the teaching task, decompose the trajectory into a sequence of behaviors consisting of exploration, hesitation, correction and confirmation primitives, and calculate a transition probability matrix of behavior primitives based on the sequence of behaviors. The update module is used to calculate a gain value for updating the multidimensional cognitive state vector based on the behavioral primitive transition probability matrix. The multidimensional cognitive state vector is updated using the gain value; when the probability of a loop consisting of hesitant primitives and corrective primitives exceeds a preset cognitive fixation threshold, the visual elements of the teaching task are reconstructed non-cue. The selection module is used to calculate the cognitive coupling degree between the updated multidimensional cognitive state vector and the preset multidimensional cognitive load vector of each teaching task in the teaching task library, and select the teaching task whose cognitive coupling degree is in the preset target range as the next round of interaction task.

10. The system according to claim 9, characterized in that, The construction of the initial multidimensional cognitive state vector representing the child's current cognitive level includes: The dimensions of the multidimensional cognitive state vector are defined as spatial perception ability, logical reasoning ability, and hand-eye coordination ability. For each dimension, an initial score is calculated based on the task completion time, number of erroneous operations, and operation path length in the initial interaction data, with a value range of 0 to 100. The initial scores together constitute the initial multidimensional cognitive state vector.

Images