An interactive teaching system for digitized teaching materials

Abstract

The application relates to the technical field of digital teaching system, and discloses an interactive teaching system of digital teaching materials, which is combined with knowledge graph and artificial intelligence auxiliary decision technology and comprises a knowledge node database based on knowledge graph construction, a content presentation unit and a teaching path dynamic branching unit with an embedded artificial intelligence recommendation model. The teaching path dynamic branching unit is used for responding to a node switching request of a learner, inquiring the knowledge node database, dynamically generating a micro teaching guidance interface containing supporting teaching options and linear regular teaching options according to logical association in the graph, providing non-mandatory auxiliary decision suggestions through the intelligent recommendation model, and redirecting the content presentation unit according to active selection of the learner on the interface. The application changes the regular page turning operation into a teaching path selection event actively participated in by the learner, changes the content supply from a fixed linear sequence into a dynamic branching structure, and avoids cognitive hindrance.

Description

Technical Field

[0001] This invention relates to an interactive teaching system for digital teaching materials, belonging to the technical field of digital teaching systems. Background Technology

[0002] Digital textbooks are widely used in current technology, typically embedding interactive elements such as multiple-choice questions and trending links into pre-set text and video content. Learners trigger preset feedback through physical operations like clicking and dragging. However, this interaction remains at the interface level; the underlying knowledge transfer logic remains linear and rigid. Existing systems push the same standardized content sequence to all learners. This undifferentiated and rigid content supply method is fundamentally contradictory to the educational field's need for a learner group with highly heterogeneous cognitive abilities. This contradiction leads to widespread teaching failure in self-directed learning scenarios, mainly manifested in the following ways: for learners with weak foundations, the rigid content sequence lacks sufficient supporting materials for key knowledge points, easily causing learners to encounter knowledge gaps and subsequent learning failures, resulting in cognitive obstacles; for learners with solid foundations, it forces them to browse a large amount of already mastered redundant content, severely slowing down learning efficiency and causing cognitive redundancy.

[0003] To resolve this contradiction, some technologies attempt to provide adaptive feedback through system diagnosis. For example, Chinese invention patent CN119296397A discloses an interactive teaching method based on contextual environment. For calligraphy teaching, it compares the differences between students' handwriting and standard fonts, analyzes incorrect radicals, and generates targeted training plans for students. While offering a degree of personalization in specific skill training, it is essentially still a diagnostic push model, heavily reliant on the system's accurate diagnosis of learners' specific behaviors, such as writing errors. The system determines subsequent teaching content, failing to fundamentally restructure the knowledge transmission process and thus unable to address the cognitive redundancy or obstacles arising from learners' varying levels of prior knowledge in general knowledge learning. The power to choose the teaching path remains with the learner. Another approach is to introduce complex artificial intelligence models to diagnose learners' cognitive states and dynamically push content. However, this method is not only highly complex and costly, but diagnosing the learner's true cognitive state is itself a difficult and unreliable technical approach, failing to address the rigidity of content supply from the perspective of the teaching process organization.

[0004] Therefore, the technical problem to be solved by this invention is how to avoid complex state diagnosis paths and instead reconstruct the organization and interaction logic of the teaching program itself, so as to break the rigidity of textbook content supply in a simple, low-cost and deterministic way, and enable the system to respond to learners' active needs in real time. Summary of the Invention

[0005] To address the problems mentioned in the background art, the technical solution of the present invention is as follows: An interactive teaching system for digital textbooks, comprising a knowledge node database, content presentation units, and dynamic branching units for teaching paths:

[0006] The knowledge node database is constructed as a storage structure that conforms to the topological characteristics of the subject knowledge graph. It is used to store multiple teaching knowledge nodes as graph entity nodes, and to store the logical associations and teaching dependencies between teaching knowledge nodes as semantic relationship edges of the graph.

[0007] Content presentation units are used to present teaching knowledge nodes;

[0008] The teaching path dynamic branch unit is equipped with an AI recommendation model for decision support, which is used to capture learners' node switching requests for the current teaching knowledge node in the content presentation unit; in response to the node switching request, the knowledge node database is queried, and a mini teaching guidance interface is dynamically generated based on logical association and through the AI ​​recommendation model. The teaching path options included in the mini teaching guidance interface are limited to include: options pointing to supportive teaching knowledge nodes associated with the current teaching knowledge node and options pointing to linear regular teaching knowledge nodes.

[0009] The dynamic branching unit of the teaching path further includes an asymmetric adjustment unit for the guidance interface. As the decision output of the artificial intelligence recommendation model, the asymmetric adjustment unit for the guidance interface is used to: analyze the teaching dependency between the current teaching knowledge node and the next linear teaching knowledge node before the micro-teaching guidance interface is generated; and when the teaching dependency is determined to be strong, apply asymmetric recommendation presentation adjustment to the teaching path options in the micro-teaching guidance interface that point to the supporting teaching knowledge node; the asymmetric recommendation presentation adjustment is limited to visual guidance under artificial intelligence-assisted decision-making, so that learners can make the final active choice. The dynamic branching unit of the teaching path is also used to redirect the content presentation unit to the selected subsequent teaching knowledge node according to the learner's active selection of the teaching path options on the micro-teaching guidance interface.

[0010] Preferably, the system further includes an intervention timing adaptive unit, which is used to monitor the learner's calling behavior of auxiliary tools for the current teaching knowledge node on the content presentation unit; and the teaching path dynamic branching unit is limited to responding to the node switching request and dynamically generating a mini teaching guidance interface only when the intervention timing adaptive unit detects the calling behavior.

[0011] Preferably, the node switching request is a deterministic operation event initiated by the learner to end the learning of the current teaching knowledge node. The deterministic operation event is the operation of clicking the next page button or the next section button.

[0012] Preferably, the teaching attributes include core definitions, concrete examples, prior reviews, or reinforcement exercises.

[0013] Preferably, the asymmetric recommendation presentation adjustment is to set the teaching path options for supportive teaching knowledge nodes as the default highlighted options or preferred recommended options on the interface, and to place the options for linear regular teaching knowledge nodes in a secondary position.

[0014] Preferably, the mini-teaching guidance interface is a dynamically generated overlay or transition page.

[0015] Preferably, the system further includes a path consistency monitoring unit, which monitors and counts the number of times learners select options for supportive instructional knowledge nodes. and the number of times the options for linear conventional teaching knowledge nodes were selected. ; Calculate the path preference ratio ,in And when determining the path preference ratio If the learner consistently exceeds the preset path preference threshold, the learner will be identified as having an inconsistent teaching contract.

[0016] Preferably, the path consistency monitoring unit is also used to intercept the node switching request when the learner selects the option of a linear conventional teaching knowledge node again in a state of non-consistent teaching contract; and to forcibly redirect the content presentation unit to the teaching checkpoint node associated with the previous teaching knowledge node, wherein the teaching checkpoint node is a teaching verification problem that cannot be skipped.

[0017] Preferably, the system further includes a collective path optimization unit, which serves as an online evolutionary learning module for the artificial intelligence recommendation model. The collective path optimization unit is used to asynchronously aggregate path selection data of multiple learners on the same micro-teaching guidance interface for each teaching path option; based on the path selection data, iteratively update the teaching path options and calculate the collective preference weight; and, when dynamically generating the micro-teaching guidance interface, the teaching path dynamic branching unit is also used to sort the teaching path options according to the updated collective preference weight, thereby realizing intelligent recommendation optimization based on collective intelligence.

[0018] Preferably, the teaching knowledge nodes in the knowledge node database are further identified with representational attributes, which are either abstract or concrete representations. The system includes a teaching attribute filtering unit, which is used to obtain the content representation preferences actively set by the learner. Furthermore, when retrieving supportive teaching knowledge nodes in the dynamic branch unit of the teaching path, the content representation preferences are used to perform secondary filtering on the retrieved supportive teaching knowledge nodes.

[0019] Compared with the prior art, the beneficial effects of the present invention are:

[0020] 1. When the dynamic branching unit of the teaching path captures the learner's node switching request, it does not immediately execute linear content jump; based on the pre-set association relationship reflecting the internal logic of teaching in the knowledge node database, it dynamically generates a guide interface containing different teaching path options, transforming the conventional page-turning operation into a teaching path selection event actively participated in by the learner. This changes the system's content supply method from a fixed linear sequence to a dynamic branching structure that can respond to the learner's implicit needs in real time, thereby resolving the inherent contradiction between standardized content supply and learner cognitive heterogeneity.

[0021] 2. The dynamically generated teaching path options include at least options pointing to supportive teaching knowledge nodes such as concrete examples or prior reviews, as well as options pointing to linear conventional knowledge nodes. This option design based on teaching attributes provides differentiated paths for learners with different cognitive foundations using the same set of teaching materials: those with weak foundations actively choose to obtain supportive materials to avoid cognitive obstacles, while those with solid foundations choose conventional paths to skip redundant content, thus avoiding cognitive redundancy and achieving a balance between teaching adaptability and teaching efficiency.

[0022] 3. The core operating mechanism of the system does not rely on complex analysis of learner behavior or diagnosis of cognitive state for triggering and response. Through the structural design of the teaching program, the system transforms personalized technical paths from a high-cost, uncertain approach that attempts to diagnose learners into a low-cost, deterministic approach that empowers learners with guided choices. This design achieves the advanced goal of teaching adaptability through a simple, instructional logic-based engineering implementation, avoiding the technical dependence and implementation difficulty of complex diagnostic models. The dynamic branching units of the teaching path record the learner's path selection history. Based on this selection history, the system automatically adjusts the call weight of reinforcement practice nodes in subsequent teaching processes according to the frequency of learners selecting linear conventional options. It reuses learners' active selection data to construct a simple teaching feedback loop: without the need for complex models, it achieves subsequent adjustments to the teaching program only through the statistics of deterministic choices, providing teaching support for learners who continuously skip options and further enhancing teaching. Attached Figure Description

[0023] Figure 1 This is the core logic flowchart of the dynamic path branching system of the present invention;

[0024] Figure 2 A comparative chart showing the impact of the system configuration of this invention on the learning time of different learners;

[0025] Figure 3 This is a diagram showing the main roles and core functions of the system of this invention. Detailed Implementation

[0026] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be described in detail below. Obviously, the described embodiments are only some embodiments of this invention, not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0027] This invention provides an interactive teaching system for digital textbooks, including a knowledge node database, content presentation units, and dynamic branching units for teaching paths. The knowledge node database is used to structurally store teaching content and its relationships based on a knowledge graph topology. The content presentation units are used to display selected teaching knowledge nodes to learners. The dynamic branching units for teaching paths serve as the system's control logic, responding to deterministic operations initiated by learners, querying the knowledge node database, and dynamically generating a guided interface with pedagogical significance, replacing traditional linear page navigation. Ultimately, based on the learner's active selection, the content presentation units are scheduled to display subsequent teaching content, transforming standardized content supply into a dynamic distribution adapted to the heterogeneity of learners' cognition. Branch paths; a knowledge node database, which can be implemented as a relational database or a graph database. The core data structure is no longer linear chapters, but rather minimized teaching knowledge nodes mapped to graph entities. Each teaching knowledge node is assigned multi-dimensional attribute identifiers in the database. The first attribute is a teaching attribute, used to define the node's role in the teaching methodology. As a specific implementation method, this attribute can be an enumerated field, with a value range including at least core definitions, concrete instances, prior review, or reinforcement exercises. The second attribute is a representation attribute, used to define the presentation style of the node content, with a value range including abstract or concrete representations. The database also stores association data defining the relationships between nodes, specifically forming the edge relationships in the knowledge graph. This includes logical associations, used to define branchable teaching options for nodes, such as core definition nodes being logically associated with multiple concrete instance nodes and prior review nodes, and teaching dependencies, used to identify the necessity of a node for learning its linear successor nodes, such as a strong dependency between a core definition node and its linear next node, the core law; and a dynamic branching unit for the teaching path, which can be implemented by a software module deployed on the server or client. This unit has a built-in rule-based and statistical AI recommendation model, and its core operating mechanism lies in the reconstruction of the teaching logic for learner node switching requests. Node switching requests are limited to deterministic operation events initiated by the learner to end the learning of the current teaching knowledge node. In the specific implementation... This event refers to the learner clicking the "Next Page" or "Next Section" button. When this unit captures this deterministic operation event, it intercepts the default linear page navigation behavior and instead executes the following teaching logic: Based on the identifier of the current teaching knowledge node, the unit queries the logical associations in the knowledge node database and retrieves all supporting teaching options related to the current node. The unit dynamically generates a mini-teaching guidance interface, which can be a dynamically generated overlay or transition page. The presented teaching path options are limited to simultaneously including: options pointing to supporting teaching knowledge nodes associated with the current teaching knowledge node, such as viewing concrete examples or reviewing prior knowledge, and options pointing to linear conventional teaching knowledge nodes, such as directly learning the next knowledge point.Based on the learner's active selection of the teaching path options on the micro-teaching guidance interface, the content presentation unit is redirected to the selected subsequent teaching knowledge node.

[0028] In a preferred embodiment, the dynamic branching unit of the teaching path further includes an asymmetric adjustment unit for the guidance interface, used to address cognitive difficulties caused by learners improperly skipping key prior knowledge during instruction. Specifically, before generating the mini-teaching guidance interface, the knowledge node database is queried to analyze the teaching dependency between the current teaching knowledge node and the next linear teaching knowledge node. When the teaching dependency is determined to be strong, the adjustment unit activates intervention, applying asymmetric recommendation-based presentation adjustments to the teaching path options in the mini-teaching guidance interface that point to supporting teaching knowledge nodes, such as concrete instance nodes or prior review nodes. The specific implementation of this presentation adjustment... The current approach can be as follows: By modifying the display attributes of interface elements, the teaching path options for supportive teaching knowledge nodes can be set as the default highlighted options or preferred recommended options on the interface, while the options for linear, conventional teaching knowledge nodes can be placed in a secondary position. This provides pedagogical guidance without depriving learners of their choice. The system can further include an adaptive intervention timing unit to optimize the accuracy of teaching interventions and avoid teaching fatigue caused by overly frequent guidance interfaces. This unit is used to monitor in real time learners' use of auxiliary tools for the current teaching knowledge node within the content presentation unit. These auxiliary tools can be existing highlighting tools in the digital textbook or... Note-taking tools; from a teaching logic perspective, such invocation behavior is viewed as a simplified representation signal that the learner is encountering high cognitive load at that node; the triggering logic of the dynamic branching unit of the teaching path is constrained by this unit, limited to responding to the node switching request and dynamically generating a mini-teaching guidance interface only when the intervention timing adaptive unit detects the invocation behavior; if no invocation behavior is detected, a regular linear jump is performed, precisely focusing teaching intervention resources on the knowledge nodes most needed by teaching evidence; the intervention timing adaptive unit internally includes an event listening module for monitoring the invocation behavior of auxiliary tools and a log processing module for analyzing the node access sequence; event listening... When the listening module detects that a learner triggers the highlighting or note-taking tools in the content presentation unit, it increments the tool call count associated with the current teaching knowledge node by one. When the log processing module detects that a learner revisits the current teaching knowledge node within a preset time window of 300 seconds after leaving the current teaching knowledge node, it increments the re-entry count of that node by one. When the intervention timing adaptive unit detects that the tool call count exceeds a preset call threshold (e.g., 2 times) or the re-entry count exceeds a preset re-entry threshold (e.g., 1 time), it marks the current teaching knowledge node as an intervention active state within the system, serving as a prerequisite for the dynamic branch unit of the teaching path to respond to node switching requests.

[0029] To address potential learners' malicious avoidance of instructional paths, a path consistency monitoring unit is included; this unit monitors and tracks the number of times learners select options for supportive instructional knowledge nodes in the background. and the number of times the options for linear conventional teaching knowledge nodes were selected. ; Calculate the path preference ratio based on the pre-set teaching logic. This ratio can be based on Calculate, where, This represents the number of times the options for linear, conventional teaching knowledge nodes are selected. The number of times the option for supporting instructional knowledge nodes is selected; a built-in preset path preference threshold is provided, which can be set by teaching experts, such as 0.85; when determining the path preference ratio... If the learner consistently exceeds the preset path preference threshold, for example, exceeding 0.85 on five consecutive guided interfaces, the unit identifies the learner as being in an inconsistent teaching contract state. The path consistency monitoring unit is also used to: intercept the node switching request when the learner is in an inconsistent teaching contract state and selects the option of a linear regular teaching knowledge node on the guided interface again; not execute the selected linear jump, but redirect the content presentation unit to the teaching checkpoint node associated with the previous teaching knowledge node. The teaching checkpoint node is defined in teaching as an unskippable teaching verification question, such as a multiple-choice question related to the core concept of the previous node. The learner must answer the question correctly before being allowed to continue their linear learning path.

[0030] The recommended presentation adjustment of the asymmetric adjustment unit in the guidance interface is set as a progressive intervention. This unit is connected to the path consistency monitoring unit to determine if the learner is in a state of inconsistency between the teaching contract and the learning agreement. When the learner is not in this state, the asymmetric recommended presentation adjustment is implemented by highlighting the options for supportive teaching knowledge nodes by default. When the learner is identified as being in a state of inconsistency between the teaching contract and the learning agreement, the asymmetric adjustment unit automatically upgrades the intervention intensity of the recommended presentation adjustment in subsequent micro-teaching guidance interfaces. This involves retaining the highlighted options for supportive teaching knowledge nodes, placing the options for linear conventional teaching knowledge nodes out of position on the interface, and adding a non-skippable 3-second recommended reading prompt overlay before allowing the learner to make a selection. To achieve self-optimization of the teaching path, a collective path optimization unit is included, used to iteratively optimize the expert-preset teaching path using collective wisdom. This unit essentially constitutes the feedback learning mechanism of the artificial intelligence recommendation model, used to asynchronously aggregate path selection data from multiple learners on the same micro-teaching guidance interface for each teaching path option. Based on the path selection data, a collective preference weight is calculated for each teaching path option and used as a model parameter for updating. The standardized selection frequency statistics are stored and bound to logical associations. When dynamically generating the micro-teaching guidance interface, the dynamic branching unit of the teaching path sorts the teaching path options based on collective preference weights, presenting the option with the highest collective preference weight as the preferred recommendation calculated by the artificial intelligence model to subsequent learners. To achieve adaptability of content representation style based on the adaptability of the teaching path, a teaching attribute filtering unit is included. This scheme requires that in the knowledge node database, teaching knowledge nodes, in addition to identifying teaching attributes, also have representation attributes, which can be abstract or concrete representations. The teaching attribute filtering unit is used to: obtain the content representation preferences actively set by learners, selected once by learners through the settings interface before the course begins; and when the dynamic branching unit of the teaching path retrieves related supporting teaching knowledge nodes based on teaching attributes, such as concrete examples, this filtering unit immediately intervenes, using the learner's actively set content representation preferences to perform a secondary filter on the retrieved list of supporting teaching knowledge nodes, retaining those nodes whose representation attributes match the learner's content representation preferences. This secondary-filtered node list is then used to generate the final micro-teaching guidance interface options.

[0031] Example 1: In this example, in the application scenario of digital physics textbooks in higher education, the textbook content is stored in a knowledge node database according to the aforementioned method. The teaching knowledge node "Newton's Second Law" is marked as a core definition, and the next linear teaching knowledge node, the derivation of the kinetic energy theorem, is marked as a core law. The database explicitly records the strong dependence of the teaching on the derivation from Newton's Second Law to the kinetic energy theorem. Teaching knowledge nodes logically related to Newton's Second Law also include nodes marked as concrete examples (e.g., rocket thrust calculation) and nodes marked as prior review (e.g., inertia concept review). The teaching scenario involves two types of learners with cognitive heterogeneity: Learner A, with a weak foundation in physics and difficulty understanding abstract definitions; and Learner B... With a solid foundation in physics, this learning session was merely a quick review. When learner A and learner B completed the current teaching knowledge node of Newton's Second Law through the content presentation unit and both clicked the next page button, the system captured this deterministic node switching request. The system's dynamic branching unit of the teaching path was triggered, intercepting the default linear jump action of the request to the derivation of the kinetic energy theorem. The asymmetric adjustment unit of the teaching path dynamic branching unit's guidance interface intervened, querying the teaching dependency between the current node Newton's Second Law and the linear next node, the derivation of the kinetic energy theorem, in the knowledge node database based on the node switching request. The unit identified this dependency as a strong dependency and activated the asymmetric intervention logic.

[0032] The dynamic branching unit of the teaching path queries the database to retrieve all teaching path options logically related to Newton's second law, namely, supportive options for rocket thrust calculation, inertia concept review, and linear conventional kinetic energy theorem derivation. Based on the intervention logic activated by the asymmetric adjustment unit of the guidance interface, the system dynamically generates a mini-teaching guidance interface. When presented to learner A and learner B, both perform asymmetric recommendation adjustments: the rocket thrust calculation and inertia concept review options pointing to supportive teaching knowledge nodes are set as the default highlighted or prioritized options, while the kinetic energy theorem derivation option pointing to linear conventional teaching knowledge nodes is placed in a secondary position. Faced with this mini-teaching guidance interface, learner A responds... The system's recommended presentation is adjusted, proactively highlighting the concrete example option of rocket thrust calculation. Responding to this selection, the system redirects the content presentation to the rocket thrust calculation knowledge node. Learner A, before delving into complex derivations, consolidates their concrete understanding of the core definition, avoiding the cognitive obstacles mentioned earlier in education. Learner B, facing the same mini-teaching guidance interface, relies on their solid foundation to determine that reviewing the example is unnecessary, proactively ignoring the system's recommended adjustment and selecting the less important linear conventional option, the derivation of the kinetic energy theorem. The system also responds to this selection, redirecting the content presentation to the kinetic energy theorem derivation knowledge node. Learner B skips content already mastered, avoiding the cognitive redundancy issues mentioned earlier in education.

[0033] Example 2: This example provides experimental data to verify the technical effectiveness of the interactive teaching system of the digital textbook of the present invention in resolving the contradiction between learner cognitive heterogeneity and standardized content supply in a teaching scenario, compared with the traditional linear teaching system; it quantitatively compares the impact of three different teaching content supply methods on two groups of learners with cognitive heterogeneity in terms of learning efficiency and learning mastery; Experimental platform and subjects: 60 first-year university students were recruited as subjects. Through standardized subject pre-tests, the subjects were divided into a low prior knowledge group (30 people, pre-test...). The test was conducted with two groups: a group with scores below 40 and a group with high prior knowledge (30 participants, whose pre-test scores were above 85). The test materials were digital physics textbook modules, with the key assessment point being the derivation of the kinetic energy theorem, and the prerequisite point, Newton's second law, being pre-set as strongly dependent. The test system was deployed in a standardized computer teaching environment. The test group design was as follows: the two groups of participants, the low prior knowledge group and the high prior knowledge group, were randomly divided into three groups of 10 each. Each group used a different configuration of the teaching system to complete the same learning task: Control group A (existing technology): a fixed linear sequence system was used, and learners had to follow the instructions in order to complete the task. The learning sequence is fixed, from Newton's Second Law to concrete examples, from prior review to the derivation of the kinetic energy theorem. Control Group B (partial scheme): uses a system with dynamic branching units in the teaching path. After the Newton's Second Law node, a mini-teaching guidance interface is dynamically generated, but all options in the interface, including supportive and linear options, are presented symmetrically and without bias. The present invention's sample group: uses the complete system of this invention, which simultaneously possesses dynamic branching units in the teaching path and asymmetrical adjustment units in the guidance interface. After the Newton's Second Law node, a mini-teaching guidance interface is dynamically generated. Due to its strong... The system applies asymmetric recommendation adjustments to the options for supporting teaching knowledge nodes, such as concrete examples and prior reviews, by default highlighting them. The experimental process and data collection involved all participants completing the learning task independently, with the system automatically recording the total learning time. Upon completion, all participants immediately took a standardized post-test of learning mastery, consisting of 20 objective questions (out of 100 points). Simultaneously, the system recorded the proportion of learners actively selecting supporting teaching knowledge nodes on the micro-teaching guidance interface in both the control group (B) and the present invention's sample group. The experimental data were summarized and averaged; the results are shown in Table 1.

[0034] Table 1: Comparison of the impact of different teaching system configurations on the teaching effectiveness of two types of learners

[0035]

[0036] Referring to the data in Table 1, for the high prior knowledge group, control group A (fixed linear) had the longest average learning time (31.2 minutes) due to the forced learning of redundant content. Control group B (symmetric branching) and the present invention sample group (asymmetric branching) both allowed learners to skip redundancy (the active selection rate of supportive nodes was only 10%), and the average learning time was shortened to below 16.5 minutes, without affecting the post-test mastery score (both above 91 points), indicating that the dynamic branching mechanism avoids cognitive redundancy. For the low prior knowledge group, although control group A (fixed linear) forced the learning of supportive content, the passive learning method led to a lower post-test mastery score. The score was only 60.5 points; due to the lack of guidance, learners in the control group B (symmetric branch) tended to skip supportive content (active selection rate was only 30%), resulting in a short average learning time (24.1 minutes) and a low post-test mastery score (62.3 points); in contrast, the sample group of this invention (asymmetric branch) improved the recommended presentation of the asymmetric adjustment unit in the guidance interface, increasing the active selection rate of supportive nodes in the low prior knowledge group to 80%, and the average learning time increased accordingly to 33.8 minutes (close to the control group A), with a post-test mastery score of 84.6 points.

[0037] Example 3: This example combines Figures 1 to 3 A description of an interactive teaching system for a digital textbook, such as... Figure 1 As shown, the system includes a knowledge node database, a content presentation unit, a dynamic branching unit for the teaching path, an asymmetric adjustment unit for the guidance interface, a mini teaching guidance interface, and a path consistency monitoring unit. The content presentation unit presents the current teaching knowledge node and receives the learner's node switching request. This request triggers the dynamic branching unit for the teaching path. After capturing the request, the dynamic branching unit for the teaching path queries the logical relationships in the knowledge node database. Triggered by the asymmetric adjustment unit for the guidance interface, this adjustment unit also queries the teaching dependencies in the knowledge node database. Based on this, the dynamic branching unit for the teaching path dynamically generates a mini teaching guidance interface containing supportive options and linear regular options, and applies recommended adjustments. The learner actively selects on this interface. This selection is monitored by the path consistency monitoring unit to determine whether the teaching contract is consistent, and it also guides the dynamic branching unit for the teaching path to redirect the content presentation unit to the selected node. When the path consistency monitoring unit determines that the teaching contract is inconsistent, it can forcibly redirect the content presentation unit to the teaching checkpoint.

[0038] like Figure 2As shown in the chart, the average learning time in minutes is used as the vertical axis to compare the learning performance of learners in the low-prior-knowledge group and the high-prior-knowledge group under three configurations listed on the horizontal axis: fixed linear system, symmetric branching system, and asymmetric branching system. The height comparison of the bars shows that in all three configurations, the average learning time of the low-prior-knowledge group is higher than that of the high-prior-knowledge group. For the high-prior-knowledge group, the average learning time in both the symmetric and asymmetric branching systems is less than 20 minutes, significantly lower than the fixed linear system (more than 30 minutes). For the low-prior-knowledge group, their learning time in both the fixed linear system and the asymmetric branching system is at a high level, exceeding 30 minutes, with the time decreasing to less than 25 minutes only in the symmetric branching system. Figure 3 As shown, the system involves two roles: learners and instructional designers. Learners can perform operations such as requesting node switching, selecting instructional paths, and setting content representation preferences. Instructional designers can perform operations such as managing knowledge nodes and logic and calibrating instructional thresholds. The system then executes a series of internal processes accordingly, including dynamically generating a guided interface in response to learners' requests, applying asymmetric recommendations to the interface, monitoring the consistency of learners' paths, and triggering mandatory instructional checkpoints when necessary.

[0039] Example 4: The internal operation procedures and collaborative logic of the collective path selection unit and the teaching attribute filtering unit in this example are as follows: In the application scenario, the system knowledge node database records teaching knowledge node A, with attributes defined as core attributes; the supporting teaching knowledge nodes logically associated with node A include: node B, which is a concrete instance and marked as a concrete representation; node C, which is a concrete instance and marked as an abstract representation; and node D, which is a priori review and marked as a concrete representation. The collective path selection unit has asynchronously aggregated historical data and recorded it on the mini teaching guidance interface of node A. The path selection data for node B is 120 times, for node C it is 30 times, and for node D it is 85 times. The content representation preference has been actively set to concrete attributes through system settings. A learner with a concrete representation completes learning node A in the content presentation unit and initiates a node switching request. The system's dynamic branching unit for the teaching path captures this request and, before dynamically generating a mini-teaching guidance interface, calls the teaching attribute filtering unit. It obtains the learner's content representation preference as concrete representation and then uses this preference as a filtering condition to perform secondary filtering on the entire list of retrieved supporting teaching knowledge nodes, including nodes B, C, and D. The filtering logic is to retain all nodes B and D with the representation attribute of concrete representation and remove node C with the representation attribute of abstract representation from the list. If the secondary filtering does not find any nodes matching the learner's preference, the fault tolerance logic is triggered, and the original unfiltered node list is returned.

[0040] The system obtains a list of supportive nodes after secondary filtering, including nodes B and D. The dynamic branching unit of the teaching path calls the collective path optimization unit to calculate and sort the collective preference weights for the nodes in this list. The collective path optimization unit assigns preference weights to each node according to a preset weight calculation procedure, such as a weighted frequency algorithm that considers time decay. calculate: ,in for The system calculates the weight of node B as 120 and the weight of node D as 85 based on the cumulative number of path selections made in the previous day. The dynamic branching unit of the teaching path sorts the filtered teaching path options in descending order based on this weight calculation. The order of the teaching path options presented on the dynamically generated mini-teaching guidance interface for the learner is as follows: the first option is node B, i.e., a concrete example; the second option is node D, i.e., a priori review; the third option is the inherent linear conventional teaching knowledge node; node C, i.e., an abstract representation, is not presented on this interface because it did not pass the secondary filtering. This presentation of teaching path options to the learner satisfies both content representation preferences and reflects the optimal ranking based on collective wisdom.

[0041] Example 5: This example illustrates the standardized engineering data filling procedure for the knowledge node database before system deployment. The system provides a teaching content editing backend, where instructional designers deconstruct and import course content into a series of teaching knowledge nodes, assigning teaching attributes and representation attributes to each node. The backend interface presents the current and next teaching knowledge nodes in a linear sequence to the designers in pairs, requiring them to select instructional dependency identifiers (e.g., strong or weak dependencies) for node pairs from a preset enumeration list. Designers also need to define nodes for each core node and actively link one or more supporting nodes through a visual association interface. The teaching knowledge nodes complete the construction of logical connections between nodes, providing a structured data foundation for the subsequent querying and judgment of dynamic branch units of the teaching path. This embodiment describes the standardized calibration procedure for the path preference threshold in the path consistency monitoring unit. Before the system is put into teaching application, it is run in an offline calibration phase, recruiting a group of learners with high mastery as a benchmark group. With the monitoring function turned off, they use this system to complete the learning of the standard teaching module. The system collects the path selection data of all learners in the benchmark group on all micro-teaching guidance interfaces, and calculates the final path preference ratio for each learner in the benchmark group. ,in The calibration procedure processes the baseline group statistically. The value distribution is obtained by adding 1.5 times the standard deviation to the 90th percentile value or the mean value of the distribution. This value is set as the path preference threshold of the teaching module and stored in the system configuration for subsequent determination of the non-consistent state of the teaching contract for regular learners.

[0042] Example 6: This example illustrates the internal decision-making logic and intervention procedures of the path consistency monitoring unit; the path consistency monitoring unit maintains the values ​​of the regular path selection counter for each learner in the background. and the value counted by the continuous deviation counter When the dynamic branch unit of the teaching path captures a node switching request, the logical judgment procedure of this monitoring unit is triggered: check the learner's selection on the micro-teaching guidance interface; if the learner selects a supportive teaching knowledge node, Reset to zero; if the learner selects linear, conventional teaching knowledge nodes, and Add 1 to all; this unit calculates the path preference ratio. ,in The total number of guided interfaces encountered by the learner, while also checking ;when Exceeding a preset path preference threshold, such as 0.85, and If the learner exceeds a preset consecutive deviation count value (e.g., 5 times), the unit will mark the learner as being in an inconsistent state with the instructional contract. When a learner in an inconsistent state selects a linear regular instructional knowledge node again, the intervention procedure of the path consistency monitoring unit is triggered, forcibly redirecting the learner to an instructional checkpoint node. The procedure is set up in the knowledge node database, and an instructional checkpoint question bank is set up. Each verification question in the question bank is forcibly associated with one or more instructional knowledge nodes being assessed by the instructional attributes. When the intervention action is triggered, the node ID of the instructional knowledge node that the learner attempted to skip is obtained. Using this ID as a query index, the unit retrieves the verification questions associated with the node ID that the learner has not yet answered in the instructional checkpoint question bank. These questions are then used as the instructional checkpoint nodes that cannot be skipped and are pushed to the content presentation unit.

[0043] This embodiment further elaborates on the intervention timing adaptive unit, a supplementary monitoring procedure adopted to address abnormal teaching situations where learners feel confused but do not use assistive tools. In addition to monitoring assistive tool usage, the intervention timing adaptive unit monitors learners' node re-entry behavior in the background. It records a time-series log of learners accessing each teaching knowledge node in the background. If it detects that after leaving teaching knowledge node A, within a preset short time window (e.g., 5 minutes), the number of times the learner actively returns to and re-accesses node A exceeds a preset re-entry threshold, then... At this time, the system will re-enter this node behavior, which is regarded in the teaching logic as a simplified representation signal that the learner encountered high cognitive load at node A; the intervention timing adaptive unit is also activated, so that the teaching path dynamic branch unit can dynamically generate a mini teaching guidance interface when the learner initiates a node switching request next time.

[0044] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0045] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. An interactive teaching system for digitized teaching materials, characterized by, The system comprises a knowledge node database, a content presentation unit, and a teaching path dynamic branching unit. The knowledge node database is constructed as a storage structure conforming to the topological characteristics of a subject knowledge graph, and is used to store a plurality of teaching knowledge nodes as graph entity nodes, and store logical associations and teaching dependencies between the teaching knowledge nodes as graph semantic relationship edges. The content presentation unit is used to present the teaching knowledge nodes. The teaching path dynamic branching unit is configured with an artificial intelligence recommendation model of an auxiliary decision type, and is used to capture a node switching request of a learner on the content presentation unit for a current teaching knowledge node; in response to the node switching request, the knowledge node database is queried, and a micro teaching guidance interface is dynamically generated according to the logical associations and through the artificial intelligence recommendation model, wherein the teaching path options contained in the micro teaching guidance interface are limited to simultaneously including options pointing to supporting teaching knowledge nodes associated with the current teaching knowledge node and options pointing to linear regular teaching knowledge nodes. The teaching path dynamic branching unit further comprises a guidance interface asymmetric adjustment unit, which is a decision output end of the artificial intelligence recommendation model, and is used to analyze the teaching dependency between the current teaching knowledge node and the linear next teaching knowledge node before the micro teaching guidance interface is generated; and when it is determined that the teaching dependency is strong dependency, an asymmetric recommendation presentation adjustment is applied to the teaching path options pointing to the supporting teaching knowledge nodes in the micro teaching guidance interface; the asymmetric recommendation presentation adjustment is limited to visual guidance under the artificial intelligence auxiliary decision, for the learner to make a final active selection, and the teaching path dynamic branching unit is also used to redirect the content presentation unit to the selected subsequent teaching knowledge node according to the active selection of the learner on the teaching path options in the micro teaching guidance interface.

2. The interactive teaching system of digitized teaching materials according to claim 1, characterized in that, The system further comprises an intervention timing adaptive unit, which is used to monitor the calling behavior of a learner on the content presentation unit for an auxiliary tool for a current teaching knowledge node. And the teaching path dynamic branching unit is limited to responding to the node switching request and dynamically generating the micro teaching guidance interface only when the intervention timing adaptive unit monitors the occurrence of the calling behavior.

3. The interactive teaching system of digitized teaching materials according to claim 1, wherein, The node switching request is a deterministic operation event initiated by the learner to end the learning of the current teaching knowledge node, and the deterministic operation event is clicking the next page button or the next section button.

4. The interactive teaching system of digitized teaching materials according to claim 1, wherein, The teaching attributes include core definition, concrete instance, pre-review, or reinforcement exercise.

5. The interactive teaching system of digitized teaching materials according to claim 1, wherein, The asymmetric recommendation presentation adjustment sets the teaching path options of the supporting teaching knowledge nodes as the default highlighted options or the preferred recommended options of the interface, and places the options of the linear regular teaching knowledge nodes in a secondary position.

6. The interactive teaching system of digitized teaching materials according to claim 1, wherein, The micro teaching guidance interface is a dynamically generated floating layer or transition page.

7. The interactive teaching system of digitized teaching materials according to claim 1, wherein, The system further comprises a path consistency monitoring unit for monitoring and counting the number of selections of the options of the supportive instructional knowledge nodes by the learner and the number of selections of the options of the linear routine instructional knowledge nodes ; calculating a path preference ratio wherein and identifying the learner as being in a non-consistent state of instructional contract when the path preference ratio continues to exceed a pre-set path preference threshold.

8. The interactive teaching system of digitized teaching materials according to claim 7, characterized in that, The path consistency monitoring unit is further configured to intercept the node switching request when the learner selects the option of the linear regular teaching knowledge node again in the teaching contract inconsistency state, and to redirect the content presentation unit to a teaching checkpoint node associated with the previous teaching knowledge node, the teaching checkpoint node being a non-skippable teaching verification question.

9. The interactive teaching system of digitized teaching materials according to claim 1, wherein, The system further comprises a collective path preference unit as an online evolutionary learning module of the artificial intelligence recommendation model, the collective path preference unit being configured to asynchronously aggregate path selection data of multiple learners on each teaching path option in the same micro teaching guidance interface; Based on the path selection data, iteratively update the collective preference weight of the teaching path option; And the teaching path dynamic branching unit is further configured to sort the teaching path options according to the updated collective preference weight when dynamically generating the micro teaching guidance interface, thereby realizing intelligent recommendation optimization based on collective wisdom.

10. The interactive teaching system of digitized teaching materials according to claim 1, wherein, The teaching knowledge nodes in the knowledge node database are further identified with representation attributes, which are abstract representation or concrete representation teaching attributes filtering unit; the system comprises a teaching attribute filtering unit, which is configured to obtain the content representation preference set by the learner; and when the teaching path dynamic branching unit retrieves the supporting teaching knowledge nodes, the content representation preference is used to filter the retrieved supporting teaching knowledge nodes again.

Images