An adaptive learning and teaching system based on artificial intelligence

Abstract

The application discloses an adaptive learning and training fusion teaching system based on artificial intelligence, and relates to the technical field of learning and training fusion teaching.The system comprises a data sensing and fusion module, a fusion degree regulation module and a teaching coordination regulation module.The application constructs a knowledge graph through multi-source data fusion, dynamically regulates data reuse frequency and fusion trigger threshold based on ecological fusion degree, and optimizes knowledge graph self-update frequency and skill image dynamic update frequency according to teaching coordination efficiency feedback, thereby forming a closed-loop control of data fusion-strategy regulation-teaching execution-effect evaluation, realizing adaptive optimization and digital transformation leap of adaptive learning and training fusion teaching, and improving the reliability of adaptive learning and training fusion teaching.The application solves the problem of low reliability of adaptive learning and training fusion teaching in the prior art.

Description

Technical Field

[0001] This invention relates to the field of learning-training integration technology, and in particular to an adaptive learning-training integration teaching system based on artificial intelligence. Background Technology

[0002] Against the backdrop of digital transformation integrating learning and training in higher vocational colleges, student profiles are constructed by relying on big data analysis, knowledge graphs, and adaptive assessment algorithms. AI forms using integrated OCR (Optical Character Recognition) automatically extract structured information such as students' professional directions. NLP (Natural Language Processing) technology transforms skill pain points into tagged data such as "weak PLC (Programmable Logic Controller) skills." Tests are then distributed for core professional skills (such as a question bank for "motor control circuit design" in mechatronics), with the difficulty dynamically adjusted based on answering speed and accuracy. The system connects to the training management system and enterprise platform via API (Application Programming Interface), integrating historical data such as operation records. The knowledge graph engine matches assessment data with professional skill graphs, marking weaknesses and identifying problems in the skill chain through correlation analysis. Machine learning algorithms generate three-dimensional profiles from behavioral data.

[0003] In the solution generation phase, based on the skill gaps and career goals in the profile, a goal decomposition algorithm is used to break down the phased tasks, plan the progressive path, generate a visual calendar based on the learner's time, match resources according to learning style, and at the same time match corporate mentors, AI assistants and professional tools through recommendation algorithms.

[0004] During the teaching process, online computer vision is used to monitor virtual simulation concentration, triggering prompts when abnormalities occur. An AI Q&A robot responds to professional questions in real time, and complex faults are transferred to enterprise mentors. Before offline training, AI pushes pre-study content and synchronizes data with teachers. Group guidance is carried out based on the "skills gap heat map" generated by AI. After class, the system automatically pushes personalized exercise packages and enterprise case review lessons, realizing personalized training with deep integration of learning and training.

[0005] The above-mentioned technology has at least the following technical problems: In the integrated learning and training teaching of higher vocational colleges, the school's academic education system (such as learning management system LMS), vocational training platform, virtual simulation training environment and enterprise internship management system are all built by different vendors. The data models of each system are heterogeneous, the interfaces are incompatible, and there is a lack of a unified semantic ontology standard, which makes it impossible to effectively converge and integrate, resulting in the inability to form a complete and dynamic understanding of students' abilities and industry needs. The current virtual scenarios have limited physical realism, especially in simulating the tactile sensation of precise operations and the chain reactions of complex systems, where there is a "realism deficit". At the same time, the intelligent educational entities (AI tutors) in the scenarios have shallow cognitive levels, are mostly based on pre-set rules, and are unable to conduct in-depth, heuristic teaching interactions and contextualized feedback. The construction of virtual simulation training scenarios and the ability to coordinate with intelligent education are significantly insufficient, resulting in low quality of core skills data collection and analysis. Even if some data is integrated or individual virtual scenarios are built, existing technologies still struggle to promptly detect changes in external industry technologies, diagnose common internal teaching problems, and automatically trigger an end-to-end closed-loop response from knowledge graph updates and learning path replanning to the reallocation of teaching resources. The lack of an efficient and agile dynamic collaborative response mechanism leads to reduced collaborative efficiency and operational reliability of the overall teaching process, resulting in low reliability of adaptive learning and training integration teaching. Summary of the Invention

[0006] To address the low reliability of existing adaptive learning and training integration teaching technologies, this invention provides an artificial intelligence-based adaptive learning and training integration teaching system. The technical solution is as follows: On the one hand, an AI-based adaptive learning and training integration teaching system is provided. This system includes: a data perception and integration module, an integration degree control module, and a teaching coordination and control module. The data perception and integration module collects learning and training data from different learning scenarios and performs integration processing, obtaining integration parameters during the integration process. Based on these parameters, it obtains the learning and training data ecosystem integration degree, which characterizes the degree of integration of multi-source heterogeneous learning and training teaching data from different learning scenarios. The integration degree control module determines whether to perform ecosystem integration degree control based on the learning and training data ecosystem integration degree. If yes, an adaptive teaching scheduling step is performed after ecosystem integration degree control; otherwise, the adaptive teaching scheduling step is performed directly. Ecosystem integration degree control includes frequency reuse... The adaptive control module includes adaptive control of the self-updating frequency and adaptive control of the fusion trigger threshold. The teaching coordination control module is used to obtain the coordination parameters in the adaptive teaching scheduling process, and obtain the coordination efficiency of the learning and training integration teaching based on the coordination parameters. This efficiency is used to characterize the coordination effectiveness of learning and training integration teaching in the process of learning and training integration digital transformation. Based on the coordination efficiency of learning and training integration teaching, it is determined whether to execute the integration teaching coordination control. If yes, the learning and training integration teaching closed-loop instruction is output after the integration teaching coordination control and fed back to the data perception and fusion module to realize the leap of learning and training integration teaching in digital transformation. If no, the learning and training integration teaching closed-loop instruction is directly output and fed back to the data perception and fusion module. The integration teaching coordination control includes adaptive control of self-updating frequency and adaptive control of dynamic update frequency.

[0007] One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages: 1. By constructing a three-tiered linkage architecture encompassing data perception and fusion, fusion degree regulation, and teaching coordination regulation, a closed-loop optimization of the entire learning and training ecosystem from the data layer to the application layer is achieved. At the data foundation layer, the data ecosystem fusion degree is dynamically calculated by quantifying fusion parameters, and the frequency of data reuse and fusion trigger thresholds are intelligently adjusted based on this indicator to ensure that multi-source heterogeneous learning and training data maintains efficient collaboration in dimensions such as interface connectivity, flow latency, and service call frequency. This adaptive mechanism based on fusion status can significantly improve the quality and stability of data fusion, fundamentally solving problems such as interface response latency and service call anomalies caused by data heterogeneity, ensuring that the knowledge graph construction process is always in an optimal resource allocation state, and improving the reliability of adaptive learning and training integrated teaching.

[0008] 2. At the teaching coordination layer, by tracking the teaching path in real time to dynamically adjust parameters such as the frequency of adjustment and the time-consuming of cross-scene task connection, quantitatively evaluate the teaching coordination efficiency, and trigger the linkage control of the self-update frequency of the knowledge graph and the dynamic update frequency of the skill portrait based on the efficiency deviation. When the trigger interval of the connection task or the cross-scene time-consuming exceeds the tolerance range, the system queries through the mapping table and calculates the gain factor to dynamically compress the update cycle of the knowledge graph or increase the portrait update frequency, enabling the teaching strategy to quickly iterate according to the actual teaching fluency. This feedback mechanism driven by coordination efficiency effectively solves problems such as path planning deviation and scene connection lag caused by model lag, ensures the smoothness and self-adaptability of the cross-scene teaching process, and improves the reliability of adaptive learning and training integrated teaching.

[0009] 3. By feeding back the closed-loop instructions generated by the teaching coordination control to the data perception and fusion module, a complete digital leap closed-loop of "data fusion - strategy control - teaching execution - effect evaluation" is formed. This architecture not only realizes the deep coupling of the learning and training data ecosystem and the teaching coordination strategy, but also continuously optimizes the data transfer efficiency and teaching response agility through the synergistic effect of multiple control loops. The system significantly improves the resource allocation accuracy and teaching intervention timeliness in the digital transformation of learning and training integration, provides complete technical support for building the next-generation adaptive teaching system with self-perception, self-decision-making, and self-optimization capabilities, and improves the reliability of adaptive learning and training integrated teaching. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0011] Figure 1 It is a schematic structural diagram of an adaptive learning and training integrated teaching system based on artificial intelligence provided by an embodiment of the present invention; Figure 2 It is a flowchart of adaptive regulation of the fusion trigger threshold of an adaptive learning and training integrated teaching system based on artificial intelligence provided by an embodiment of the present invention; Figure 3 It is a flowchart of adaptive regulation of the dynamic update frequency of an adaptive learning and training integrated teaching system based on artificial intelligence provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0012] The following will describe the technical solutions in the present invention in conjunction with the drawings.

[0013] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0014] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning.

[0015] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0016] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0017] This invention provides an adaptive learning and training integration teaching system based on artificial intelligence, such as... Figure 1 The diagram shows a structural schematic of an AI-based adaptive learning and training integration teaching system, which includes: a data perception and integration module, an integration degree control module, and a teaching coordination and control module.

[0018] First, the data perception and fusion module is used to collect learning and training data from different learning scenarios and perform fusion processing to obtain fusion parameters during the fusion processing. Based on the fusion parameters, the fusion degree of the learning and training data ecosystem is obtained, which is used to characterize the degree of fusion of multi-source heterogeneous learning and training teaching data from different learning scenarios.

[0019] It's important to understand that by constructing a systematic data fusion quality assessment system, precise perception and quantitative management of the fusion status of multi-source learning and training data are achieved. Specifically, this module collects and analyzes key fusion parameters in real time, such as data interface connectivity, data flow latency, and service call frequency, to comprehensively calculate a quantitative ecological fusion degree index. This transforms the originally abstract data fusion quality into measurable and monitorable numerical values. This fusion degree assessment mechanism based on multi-dimensional parameters effectively overcomes the differences in format, protocol, and timing of multi-source heterogeneous data, significantly improving the integration efficiency and consistency of cross-scenario learning and training data. This provides a high-quality and reliable data foundation for subsequent knowledge graph construction and teaching strategy decisions, fundamentally ensuring the stability and synergy of the learning and training system's data ecosystem.

[0020] To further explain, the data perception and fusion module includes a fusion parameter unit and a fusion parameter quantification unit; the fusion parameter unit is used to collect fusion parameters, including the connectivity rate of the learning and training data ecosystem interface, the learning and training data flow delay, and the calling frequency of the learning and training data service interface. Among them, the connectivity rate of the learning and training data ecosystem interface refers to the list of registered, configured, and "active" APIs that can be directly exported from the gateway's management backend. The source system corresponding to the API is considered to be connected. The ratio of the number of connected source systems to the preset total number of connected source systems is recorded as the connectivity rate of the learning and training data ecosystem interface. The learning and training data flow delay refers to the time difference between the occurrence of a business event (e.g., a student completing an operation on a virtual training platform or submitting an assignment on an LMS) and its reception and availability by the data middle platform or integration platform. The timestamps of the occurrence and completion of business events are recorded in the logs of the data transmission tool. The difference between the timestamp of the completion and the timestamp of the occurrence of the business event is recorded as the learning and training data flow delay. The call frequency of the learning and training data service interface refers to the total number of requests to all learning and training data service APIs within a preset time period, which is queried through the access logs of the API gateway. The ratio of the total number of requests to all learning and training data service APIs to the preset time period is recorded as the call frequency of the learning and training data service interface.

[0021] The fusion parameter quantification unit is used to obtain the integration degree of the learning and training data ecosystem through fusion parameters. Specifically, it multiplies the connectivity compensation factor and connectivity component to obtain a connectivity correction value, where the connectivity component represents the result of the analysis of the ratio of connectivity rate of the learning and training data ecosystem interface to connectivity threshold; it multiplies the flow delay compensation factor and flow delay component to obtain a flow delay correction value, where the flow delay component represents the result of the analysis of the ratio of flow delay threshold to learning and training data flow delay; it multiplies the call frequency compensation factor and call frequency component to obtain a call frequency correction value, where the call frequency component represents the result of the analysis of the ratio of call frequency of the learning and training data service interface to call frequency threshold. The connectivity correction value, flow delay correction value, and call frequency correction value are coupled to obtain the integration degree of the learning and training data ecosystem. Here, multiplication refers to multiplication, coupling refers to addition, and ratio analysis refers to multiplication. The specific constraint expression for the integration degree of the learning and training data ecosystem is as follows: ; In the formula, W represents the integration degree of the learning and training data ecosystem; s1 represents the connectivity compensation factor obtained from the learning and training integrated teaching database; s2 represents the flow delay compensation factor obtained from the learning and training integrated teaching database; s3 represents the call frequency compensation factor obtained from the learning and training integrated teaching database; H represents the connectivity rate of the learning and training data ecosystem interface; L represents the learning and training data flow delay; G represents the call frequency of the learning and training data service interface; H0 represents the connectivity threshold obtained from the learning and training integrated teaching database; L0 represents the flow delay threshold obtained from the learning and training integrated teaching database; and G0 represents the call frequency threshold obtained from the learning and training integrated teaching database.

[0022] It needs to be explained that a higher connectivity rate of the learning and training data ecosystem interface means that more core learning and training systems (LMS, virtual training platforms, enterprise internship management systems, etc.) can achieve stable connection, the data source is more comprehensive, the transmission is more reliable, and the frequency of calling the learning and training data service interface is higher. A higher connectivity rate of the learning and training data ecosystem interface means that data transmission does not need to be achieved through an "intermediate transfer system", reducing the number of routing jumps and the lower the latency of learning and training data flow. The longer the latency of learning and training data flow, the lower the timeliness of data, and even abandoning some high-frequency calling scenarios, resulting in a lower frequency of calling the learning and training data service interface. Meanwhile, there is a positive correlation between the connectivity rate of the learning and training data ecosystem interfaces and the integration degree of the learning and training data ecosystem. The higher the connectivity rate of the learning and training data ecosystem interfaces, the more systems can achieve stable connection, and data can flow freely across systems. This provides a foundation for "unified data format, semantic alignment, and cross-scenario association," directly promoting the integration degree from "shallow connectivity" to "deep integration," resulting in a higher integration degree of the learning and training data ecosystem. Conversely, there is a negative correlation between the learning and training data flow delay and the integration degree of the learning and training data ecosystem. The higher the learning and training data flow delay, the more serious the lag in cross-system data synchronization, causing the integrated data to become "historical data accumulation," unable to support dynamic learning and training needs, resulting in a lower integration degree of the learning and training data ecosystem. Furthermore, there is a positive correlation between the call frequency of the learning and training data service interfaces and the integration degree of the learning and training data ecosystem. The higher the call frequency of the learning and training data service interfaces, the more the integrated data can meet the application needs of multiple scenarios and high frequency, resulting in a higher integration degree of the learning and training data ecosystem.

[0023] Next, the integration degree control module is used to determine whether to perform ecological integration degree control based on the integration degree of the learning and training data ecosystem. If yes, the adaptive teaching scheduling stage is carried out after the ecological integration degree control; otherwise, the adaptive teaching scheduling stage is carried out directly. Ecological integration degree control includes adaptive control of reuse frequency and adaptive control of integration trigger threshold.

[0024] It should be noted that by establishing an intelligent linkage mechanism between data quality and system resources, the optimal allocation of resources for the learning and training system has been achieved. This module makes dynamic decisions based on the real-time data ecosystem integration degree. When the integration degree does not meet the standard, a dual control mechanism is automatically triggered: adaptive adjustment of reuse frequency optimizes data resource utilization efficiency and avoids ineffective computational resource consumption; adaptive adjustment of integration trigger threshold improves data integration accuracy and ensures the quality of knowledge graph construction. This data quality-oriented intelligent control strategy can maintain a stable scheduling state when the system is running well, and can quickly respond and repair when data quality deteriorates. It effectively solves the coordination problem between the data ecosystem and teaching scheduling, and significantly improves the overall resource utilization efficiency and teaching service stability of the system.

[0025] Furthermore, the specific steps for determining whether to implement ecological integration degree regulation are as follows: if the ecological integration degree of learning and training data is greater than or equal to the integration degree benchmark value, then ecological integration degree regulation is not implemented; otherwise, adaptive regulation of reuse frequency and adaptive regulation of integration trigger threshold are implemented based on the integration degree control quantity. The integration degree control quantity represents the negative difference between the ecological integration degree of learning and training data and the integration degree benchmark value.

[0026] It's important to explain that in the existing learning and training data ecosystem, proactive data reuse services are typically based on fixed schedules or simple static rules. This rigid mechanism cannot adapt to dynamic changes in system load: when system resources are idle, it cannot fully utilize computing power to enhance data value; when the system faces pressure, it cannot quickly degrade to ensure core services, resulting in system performance fluctuations, low resource utilization efficiency, and a poor user experience. This solution proposes an intelligent elastic reuse method for learning and training data based on multi-threshold response and bidirectional elasticity. The core of this method lies in dividing the system response threshold into a "lazy working interval" and triggering a closed-loop control strategy with opposite directions based on the interval's state, thereby achieving a globally optimal solution for resource efficiency and system stability. Three system operating zones are defined: "over-limit high-pressure state," "optimal inert state," and "idle promotion state," replacing the traditional "normal / abnormal" binary judgment. For the "idle promotion state" and the "over-limit high-pressure state," distinct but complementary control strategies are designed, achieving bidirectional adaptive behavior of the system. A data value density matrix based on "business criticality" and "computational cost" is proposed, giving the degradation and promotion strategies clear priority objectives. Upon entering the "idle promotion state," not only is the future update frequency increased, but a "compensation update" of the historical data window is also triggered, greatly accelerating the speed at which data converges to the latest state.

[0027] As further explained in detail, the specific process for implementing adaptive multiplexing frequency control is as follows: If the response threshold of the learning and training data ecosystem interface call is less than the lower bound of the response threshold, it is determined to be in an idle enhancement state. The response threshold deviation is input into the response threshold-reuse frequency mapping table for querying, and a reuse frequency compensation update trigger instruction is obtained. Specifically, the result of harmonic averaging of the response threshold deviation and the integration degree comparison is input into the response threshold-reuse frequency mapping table for querying, and a reuse frequency gain factor is obtained. The current active reuse frequency of learning and training data is multiplied by the reuse frequency gain factor to obtain the target active reuse frequency of learning and training data. The response threshold deviation represents the negative difference between the response threshold of the learning and training data ecosystem interface call and the lower bound of the response threshold. This dual-parameter coupled control method can not only make full use of the system's idle resources to enhance data service capabilities, but also ensure that data quality does not degrade due to increased frequency through the integration degree parameter, thus achieving a balance between resource utilization efficiency and data service quality optimization.

[0028] If the response threshold of the learning and training data ecosystem interface call is within the response threshold range, it is determined to be in an optimal inertial state. Adaptive adjustment of the reuse frequency is not executed, and the current active reuse frequency of the learning and training data is maintained. The response threshold range represents the closed interval formed by the lower and upper bounds of the response threshold. Maintaining the existing reuse frequency effectively avoids unnecessary control fluctuations, reduces control overhead by maintaining a stable system operation, and reflects the energy consumption optimization and operational stability of the system under ideal working conditions. This intelligent maintenance mechanism based on state recognition ensures the continuity of data services while avoiding resource waste caused by excessive control.

[0029] Implementing adaptive multiplexing frequency control also includes: If the response threshold of the learning and training data ecosystem interface call exceeds the upper limit of the response threshold, it is determined to be in an overloaded state. The response threshold reference value is then input into the response threshold-reuse frequency mapping table for querying, obtaining a dynamic reuse window shrinkage instruction. Specifically, the harmonic average of the response threshold reference value and the integration degree reference value is input into the response threshold-reuse frequency mapping table for querying, obtaining a reuse window reduction factor. This reuse window reduction factor is multiplied by the current reuse window to obtain the target reuse window. The target reuse window is then multiplied by the current learning and training data active reuse frequency and rounded up to become the target learning and training data active reuse frequency. The response threshold reference value represents the positive difference between the learning and training data ecosystem interface call response threshold and the upper limit of the response threshold. This window and frequency linkage control strategy, under high system pressure, can both reduce the single processing load by shrinking the data window and maintain service capacity through frequency optimization, achieving dual protection of system stress resistance and service availability, effectively solving the system overload protection problem in high-concurrency scenarios.

[0030] In this embodiment, a three-state intelligent control mechanism is established to achieve precise optimization of learning and training data resources under different load conditions. This intelligent control system based on multi-state recognition organically integrates system load and data quality factors through a harmonic averaging algorithm, achieving synergistic optimization of resource utilization efficiency, system stability, and service quality, and providing adaptive resource scheduling guarantees for the learning and training data ecosystem.

[0031] It's important to explain that there's a strong causal relationship between the "triggering timing of cross-scene data fusion" and the "traceability of the final generated profile." When the system performs fusion too early or too frequently (i.e., the threshold is set too low), the data evidence chain used to build the profile is incomplete and fragmented. Traditional data fusion processes are "open-loop," setting a fixed threshold (e.g., triggering once every 100 data points or 30 minutes) and then executing unconditionally. This invention introduces traceability rate as a key quality feedback signal, constructing the entire system as a self-adaptive "closed-loop." The system no longer runs blindly but can adjust its production rhythm (fusion threshold) in real time based on the quality of its output (panoramic profile). This is a fundamental paradigm innovation; this invention does not view data quality (traceability rate) and processing efficiency (controlled by triggering thresholds) in isolation but keenly perceives the inherent, dynamic trade-off between the two. The "trigger threshold" is set as a "control valve" to regulate this. Increasing the threshold results in more complete and comprehensive data fusion in a single instance, improving traceability (but increasing the latency of profile updates); decreasing the threshold results in more timely and agile profile updates, improving real-time performance (but increasing the risk of incomplete data, which may lead to a decrease in traceability). Through this control valve, the system can intelligently and dynamically balance "high quality" and "low latency".

[0032] like Figure 2 The diagram shown is a flowchart of the adaptive control process for the fusion trigger threshold of an AI-based adaptive learning and training integration teaching system provided in this embodiment of the invention. The specific logic is as follows: If the traceability rate of the student-training panoramic portrait is within the traceability rate tolerance range, then the adaptive control of the fusion trigger threshold is not executed. If the traceability rate of the student-training panoramic portrait is less than the lower limit of the traceability rate tolerance, then the traceability rate offset and the fusion degree comparison value are input into the traceability rate-fusion trigger threshold mapping table to obtain the fusion trigger upper limit threshold gain factor. The fusion trigger upper limit threshold gain factor is multiplied by the current fusion trigger upper limit threshold to obtain the target learning and training cross-scene data fusion trigger upper limit threshold. If the traceability rate of the student-training panoramic portrait is greater than the traceability rate tolerance upper limit, then the traceability rate comparison value and the fusion degree comparison value are input into the traceability rate-fusion trigger threshold mapping table to obtain the fusion trigger lower limit threshold reduction factor. The fusion trigger lower limit threshold reduction factor is multiplied by the current fusion trigger upper limit threshold to obtain the target learning and training cross-scene data fusion trigger lower limit threshold.

[0033] As a further explanation, the specific steps for adaptive adjustment of the fusion trigger threshold are as follows: If the traceability rate of the student's panoramic profile is within the traceability tolerance range, then the adaptive adjustment of the fusion trigger threshold will not be executed. The traceability tolerance range represents the closed interval formed by the lower and upper limits of the traceability tolerance. Maintaining the existing fusion trigger threshold avoids unnecessary regulatory interference and ensures the stability of the data fusion process. This steady-state maintenance mechanism effectively reduces system regulation overhead.

[0034] If the traceability rate of the student-training panoramic profile is less than the traceability rate tolerance lower limit, the traceability rate offset and the fusion degree comparison value are input into the traceability rate-fusion trigger threshold mapping table to obtain the fusion trigger upper limit threshold gain factor. The fusion trigger upper limit threshold gain factor is then multiplied by the current fusion trigger upper limit threshold to obtain the target student-training cross-scenario data fusion trigger upper limit threshold. The traceability rate offset represents the negative difference between the student-training panoramic profile traceability rate and the traceability tolerance lower limit. This dual-parameter coupled control strategy automatically raises the fusion trigger threshold when data quality declines, ensuring that only high-quality data can enter the fusion process, thus guaranteeing the traceability of the panoramic profile from the source.

[0035] If the traceability rate of the student training profile exceeds the traceability tolerance limit, the traceability rate reference value and the fusion degree reference value are input into the traceability rate-fusion trigger threshold mapping table to obtain the fusion trigger lower bound threshold reduction factor. This reduction factor is then multiplied by the current fusion trigger upper bound threshold to obtain the target student training cross-scenario data fusion trigger lower limit threshold. The traceability rate reference value represents the positive difference between the student training profile traceability rate and the traceability tolerance limit. This optimization mechanism, while ensuring data quality, appropriately relaxes the fusion trigger conditions, improving the timeliness and coverage of data fusion, enabling the system to fully leverage data value while maintaining high-quality standards. The entire solution, through the collaborative analysis of traceability rate and fusion degree, achieves an organic unity between data quality control and data utilization efficiency, providing intelligent threshold control guarantees for student training cross-scenario data fusion.

[0036] In this embodiment, a three-level intelligent control mechanism based on the traceability rate of the student-training panoramic profile is established to achieve precise optimization and dynamic balance of the data fusion trigger threshold. When the traceability rate is within the normal tolerance range, the system maintains the existing threshold setting, effectively maintaining the stability of the data fusion process, avoiding unnecessary control interference, and significantly reducing system operating overhead. When the traceability rate is lower than the tolerance lower limit, the system generates a fusion trigger upper limit threshold gain factor through collaborative analysis of the traceability rate offset and the fusion degree comparison value, and raises the fusion trigger upper limit threshold accordingly. This mechanism can automatically raise the fusion threshold when data quality declines, ensuring the reliability and traceability of the data involved in fusion from the source. When the traceability rate is higher than the tolerance upper limit, the system obtains a lower limit threshold reduction factor based on the joint query of the traceability rate comparison value and the fusion degree comparison value, and lowers the fusion trigger lower limit threshold accordingly, effectively improving the timeliness and coverage of data fusion while ensuring data quality. This intelligent control system based on dual parameter coupling not only ensures the strictness of data quality control but also maximizes data utilization efficiency, providing a complete adaptive threshold control solution for cross-scenario data fusion in the student-training system.

[0037] Finally, the teaching coordination and control module is used to obtain coordination parameters in the adaptive teaching scheduling process. Based on the coordination parameters, the coordination efficiency of learning and training integration teaching is obtained, which is used to characterize the coordination effectiveness of learning and training integration teaching in the process of digital transformation of learning and training integration. Based on the coordination efficiency of learning and training integration teaching, it is determined whether to execute the integration teaching coordination and control. If so, the learning and training integration teaching closed-loop instruction is output after the integration teaching coordination and control and fed back to the data perception and integration module to realize the leap of learning and training integration teaching in digital transformation. If not, the learning and training integration teaching closed-loop instruction is directly output and fed back to the data perception and integration module. The integration teaching coordination and control includes self-updating frequency adaptive control and dynamic update frequency adaptive control.

[0038] It is worth noting that by constructing a complete intelligent closed-loop control system for teaching coordination, the self-optimization and digital transformation leap of the integrated learning and training system have been achieved. The teaching coordination and control module accurately quantifies and evaluates the coordination efficiency of integrated learning and training by analyzing coordination parameters in the adaptive teaching scheduling process in real time, and determines whether to initiate integrated teaching coordination and control based on this. When control is required, the system optimizes the update rhythm of the knowledge graph through self-updating frequency adaptive adjustment, and simultaneously adjusts the update strategy of the skill profile through dynamic update frequency adaptive adjustment, forming a refined dual control mechanism. Crucially, the control results are fed back to the data perception and fusion module in real time in the form of integrated learning and training closed-loop instructions, establishing a complete feedback loop from the teaching execution layer to the data foundation layer. This innovative design allows the teaching coordination effect to directly drive the optimization of data fusion strategies, not only achieving intelligent collaboration between teaching resources and data resources, but also promoting a spiral leap in the entire system during digital transformation through a continuous teaching data closed loop, ultimately building an intelligent learning and training ecosystem with self-evolving capabilities.

[0039] It should be explained that the teaching coordination and control module includes a coordination parameter unit, a coordination parameter quantification unit, and an integrated teaching coordination and control unit. The coordination parameter unit is used to collect coordination parameters, including the degree of ecological integration to be compared, the frequency of dynamic adjustments to student learning paths, and the time consumption for connecting cross-scenario teaching tasks. Specifically, the degree of ecological integration to be compared refers to the newly acquired ecological integration of learning and training data if ecological integration control has been implemented; otherwise, the current ecological integration of learning and training data is recorded as the degree of ecological integration to be compared. The frequency of dynamic adjustments to student learning paths refers to the total number of times student learning paths (such as course order, practical training projects, resource recommendation directions, assessment methods, etc.) are actively adjusted based on real-time student learning and training data per unit of time. The frequency of dynamic adjustments to student learning paths is obtained through the virtual simulation training platform. The time consumption for connecting cross-scenario teaching tasks refers to the preparation time for students to connect from completing the previous scenario task (such as virtual training) to the next scenario task (such as enterprise practice). The difference between the timestamp of the student completing the previous scenario task and the timestamp of receiving the next scenario task, recorded through transmission logs, is recorded as the time consumption for connecting cross-scenario teaching tasks.

[0040] The coordination parameter quantification unit is used to obtain the coordination efficiency of integrated learning and training teaching through coordination parameters. Specifically, it multiplies the integration degree compensation factor and integration degree component to obtain the integration degree correction value, where the integration degree component represents the result of the analysis of the ratio of the integration degree of the compared ecosystem to the integration degree threshold; it multiplies the adjustment frequency compensation factor and adjustment frequency component to obtain the adjustment frequency correction value, where the adjustment frequency component represents the result of the analysis of the ratio of the dynamic adjustment frequency of the student teaching path to the adjustment frequency threshold; it multiplies the connection time compensation factor and connection time component to obtain the connection time correction value, where the connection time component represents the result of the analysis of the ratio of the connection time threshold to the connection time of cross-scenario teaching task connection. The integration degree correction value, adjustment frequency correction value, and connection time correction value are coupled to obtain the coordination efficiency of integrated learning and training teaching. The specific constraint expression for the coordination efficiency of integrated learning and training teaching is as follows: ; In the formula, E represents the coordination efficiency of learning and training integration teaching; v1 represents the integration degree compensation factor obtained from the learning and training integration teaching database; v2 represents the adjustment frequency compensation factor obtained from the learning and training integration teaching database; v3 represents the connection time compensation factor obtained from the learning and training integration teaching database; T represents the integration degree of the ecosystem to be compared; P represents the frequency of dynamic adjustment of student teaching paths; Q represents the connection time of cross-scenario teaching tasks; T0 represents the integration degree threshold obtained from the learning and training integration teaching database; P0 represents the adjustment frequency threshold obtained from the learning and training integration teaching database; and Q0 represents the connection time threshold obtained from the learning and training integration teaching database.

[0041] It should be noted that a higher degree of integration of the comparison ecosystem means that multi-source data such as LMS, virtual training platforms, and enterprise internship systems can be effectively aggregated. Data such as students' skill gaps, learning progress, and job requirement matching can be comprehensively and in real time, providing sufficient basis for path adjustment and resulting in a higher frequency of dynamic adjustment of students' teaching paths. A higher degree of integration of the comparison ecosystem also means that the interface standards of multiple systems are unified, the data formats are compatible, and the semantics are aligned. The results of the previous scenario task (such as virtual training grades) can be synchronized to the next scenario system (such as the enterprise internship platform) in real time without additional format conversion or manual entry, and the time consumption for connecting cross-scenario teaching tasks is shorter. The longer the time consumption for connecting cross-scenario teaching tasks, the longer students have to wait after completing the previous scenario task before starting the next scenario. The timeliness of path adjustment is lost (e.g., for adjustments to a certain skill gap, the student has entered other unrelated tasks due to the connection delay), and the lower the frequency of dynamic adjustment of students' teaching paths. Meanwhile, there is a positive correlation between the degree of integration of the learning and training ecosystem and the coordination efficiency of learning and training integration. A higher degree of integration means that core systems such as LMS, virtual training platforms, and enterprise internship systems break down "information silos," allowing data to flow freely across systems and be interpreted uniformly (e.g., real-time correlation between virtual training data and enterprise job requirement data). This provides data support for "collaborative teaching decisions, collaborative resource allocation, and collaborative scenario linkage," upgrading coordination efficiency from "high efficiency in fragmented scenarios" to "high efficiency in full-ecosystem collaboration," resulting in greater coordination efficiency in learning and training integration. Furthermore, there is a positive correlation between the frequency of dynamic adjustments to student learning paths and the coordination efficiency of learning and training integration. The higher the frequency of dynamic adjustments to the teaching path, the more the system or teachers can optimize the learning path in a timely manner based on real-time learning and training data (such as students' skill gaps and updates to job requirements). This is reflected in the "efficient collaboration between teaching decisions and student status and industry needs," and the greater the coordination efficiency of integrated learning and training. There is a negative correlation between the time spent connecting cross-scenario teaching tasks and the coordination efficiency of integrated learning and training. The longer the time spent connecting cross-scenario teaching tasks, the more the teaching pace is dragged out, and the more meaningless the previous path adjustments may become (such as students forgetting the key skills in virtual training). The lower the coordination efficiency of integrated learning and training.

[0042] Furthermore, the integrated teaching coordination and control unit is used to determine whether to implement integrated teaching coordination and control. Specifically, if the learning and training integrated teaching coordination efficiency is greater than or equal to the coordination efficiency benchmark value, then integrated teaching coordination and control is not implemented. Otherwise, self-updating frequency adaptive control and dynamic update frequency adaptive control are implemented based on the teaching coordination reference value. The teaching coordination reference value represents the negative difference between the learning and training integrated teaching coordination efficiency and the coordination efficiency benchmark value.

[0043] It's important to explain that by dynamically calibrating the update rhythm of the underlying data model through the effectiveness of "teaching intervention strategies," a cross-level closed-loop optimization system with teaching effectiveness as feedback is constructed. In traditional solutions, knowledge graph updates often rely on fixed schedules or data-driven rules, while this solution creatively uses the "connecting task reminder trigger interval threshold"—a business strategy parameter—as a control signal. This adjustment is possible because of a profound causal chain behind it. In other words, it's an intelligent compensation mechanism that enhances the model's "intrinsic agility" to compensate for the strategy's "external sluggishness," achieving reverse traction and adaptive collaboration from the business layer to the data layer.

[0044] As further explained in detail, the specific process for implementing self-updating frequency adaptive adjustment is as follows: If the trigger interval threshold for the bridging task reminder is less than the lower tolerance limit for the trigger interval, the interval threshold reference value and the teaching coordination reference value are input into the interval threshold-self-update frequency mapping table for querying. This yields the knowledge graph self-update interval reduction factor. The knowledge graph self-update interval reduction factor is then multiplied by the current self-update interval, and the result is rounded down. This result is then multiplied by the current learning-training integration knowledge graph self-update frequency to obtain the target learning-training integration knowledge graph self-update frequency. The interval threshold reference value represents the negative difference between the bridging task reminder trigger interval threshold and the lower tolerance limit for the trigger interval. Through this double multiplication and rounding operation, the target update frequency is accurately calculated. This mechanism can appropriately accelerate the knowledge update pace when the system is overly sensitive, ensuring real-time synchronization of graph data and teaching dynamics, and avoiding delays in teaching decisions due to update delays.

[0045] The self-update frequency adaptive adjustment also includes: If the trigger interval threshold for the connecting task reminder is within the trigger interval tolerance range, the self-update frequency adaptive adjustment will not be executed, and the current self-update frequency of the learning and training integration knowledge graph will be maintained. The trigger interval tolerance range represents the closed interval formed by the lower limit and upper limit of the trigger interval tolerance. This steady-state operation mode ensures the continuity of knowledge graph updates and avoids unnecessary adjustment oscillations, effectively reducing the unnecessary consumption of system computing resources.

[0046] If the trigger interval threshold for the bridging task reminder exceeds the upper limit of the trigger interval tolerance, the interval threshold offset and the teaching coordination comparison amount are input into the interval threshold-self-update frequency mapping table for querying. This yields the self-update frequency multiplication factor. The self-update frequency multiplication factor is multiplied by the current learning-training integration knowledge graph self-update frequency, and the result is rounded up to the nearest integer. This is taken as the target learning-training integration knowledge graph self-update frequency. The interval threshold offset represents the positive difference between the bridging task reminder trigger interval threshold and the upper limit of the trigger interval tolerance. This control strategy effectively solves the problem of sluggish system response by accelerating knowledge iteration to promptly capture changes in teaching status, ensuring that the knowledge graph always reflects the latest learning characteristics.

[0047] In this embodiment, a three-level intelligent control mechanism based on the trigger interval of the connecting task reminder is established to achieve precise optimization and dynamic adaptation of the knowledge graph's self-update frequency. When the trigger interval is less than the tolerance lower limit, the system couples the interval threshold reference quantity with the teaching coordination reference quantity for analysis. It obtains the knowledge graph self-update interval reduction factor through a mapping table query and, through a rigorous multi-step calculation process, refines the update frequency, ensuring the timeliness of teaching decisions is maintained by accelerating the knowledge update pace when the system is overly sensitive. When the trigger interval is within the normal tolerance range, the system maintains a steady-state operation strategy with the existing update frequency, effectively avoiding unnecessary regulatory interference and ensuring the stability of system operation and the efficiency of resource utilization. When the trigger interval exceeds the tolerance upper limit, the system obtains a self-update frequency multiplication factor based on a joint query of the interval threshold offset and the teaching coordination reference quantity. Through multiplication and rounding operations, the knowledge graph update frequency is significantly increased. This mechanism effectively addresses the problem of sluggish system response and captures dynamic changes in teaching in a timely manner by strengthening knowledge iteration. The entire control system achieves intelligent matching between knowledge graph update strategies and actual teaching needs through collaborative analysis of interval thresholds and teaching coordination status. This ensures both the accuracy and timeliness of teaching decisions and the rationality and economy of system resource allocation, providing a complete adaptive optimization solution for knowledge management in the learning and training system.

[0048] It's important to explain that the core principle behind dynamically adjusting the skill profile update frequency based on the time consumption of cross-scenario teaching task transitions lies in establishing a real-time feedback loop from "teaching process experience" to "underlying model evolution." The transition time, as an intuitive indicator of user experience and system efficiency, essentially reveals the discrepancy between the current skill profile and the actual learning situation: when the time consumption increases, it means the profile may not accurately reflect the student's knowledge mastery or cognitive load, leading the system to recommend tasks with weak relevance or inappropriate difficulty, thus causing "stuttering" during the transition process. This mechanism ensures that the skill profile update rhythm intelligently links with the actual smoothness of the teaching process, transforming it from a static reference snapshot into a dynamic cognitive model that can "breathe" and "evolve" in real time with teaching interactions, thereby achieving the optimal balance between precise teaching and system efficiency.

[0049] like Figure 3 The diagram shown is a flowchart of the dynamic update frequency adaptive adjustment process of an AI-based adaptive learning and training integration teaching system provided by an embodiment of the present invention. The specific logic is as follows: if the cross-scenario teaching task connection time is less than or equal to the connection time, then dynamic update frequency adaptive adjustment is not performed; if the cross-scenario teaching task connection time is greater than the connection time, then the connection time reference value and the teaching coordination reference value are input into the connection time-dynamic update frequency mapping table for querying to obtain the dynamic update frequency adjustment factor. It is then determined whether the dynamic update frequency adjustment factor is greater than the frequency adjustment benchmark value. If so, then positive enhancement adjustment is performed based on the frequency adjustment offset. The process is as follows: Input the frequency adjustment offset into the connection time-dynamic update frequency mapping table for querying to obtain the dynamic update frequency gain factor. Multiply the dynamic update frequency gain factor with the current skill profile dynamic update frequency to obtain the adjusted skill profile dynamic update frequency. If not, negative reduction adjustment is performed based on the frequency adjustment correction amount. Specifically, input the frequency adjustment correction amount into the connection time-dynamic update frequency mapping table for querying to obtain the dynamic update frequency reduction factor. Multiply the dynamic update frequency reduction factor with the current skill profile dynamic update frequency to obtain the reduced skill profile dynamic update frequency.

[0050] As further explained in detail, the specific process for dynamic frequency adaptive control is as follows: If the time taken to connect cross-scenario teaching tasks is less than or equal to the connection time, dynamic update frequency adaptive adjustment will not be implemented. This is to avoid unnecessary frequency adjustments that could lead to wasted system resources or fluctuations in teaching strategies caused by excessive updates to skill profiles, thus ensuring the stability of teaching connections and the economy of resource utilization.

[0051] If the time taken to connect cross-scenario teaching tasks is greater than the total time taken to connect, the connection time comparison value and the teaching coordination comparison value are entered into the connection time-dynamic update frequency mapping table for querying. The dynamic update frequency adjustment factor is obtained, and it is determined whether the dynamic update frequency adjustment factor is greater than the frequency adjustment benchmark value. The connection time comparison value represents the positive difference between the connection time and the total time taken to connect cross-scenario teaching tasks.

[0052] Dynamic frequency adaptive adjustment also includes: If so, a positive boost adjustment is performed based on the frequency adjustment offset. Specifically, the frequency adjustment offset is input into the connection time-dynamic update frequency mapping table for querying to obtain the dynamic update frequency gain factor. The dynamic update frequency gain factor is then multiplied by the current skill profile dynamic update frequency to obtain the adjusted skill profile dynamic update frequency. The frequency adjustment offset represents the positive difference between the dynamic update frequency adjustment factor and the frequency adjustment baseline value. By specifically increasing the update frequency, it is ensured that the skill profile can capture the skill changes of students in cross-scenario transitions (such as the ability transfer from theoretical learning scenarios to practical operation scenarios) in real time. This allows subsequent teaching tasks to match resources and paths based on the latest skill status, reducing connection bottlenecks caused by information lag and improving the smoothness and accuracy of cross-scenario teaching.

[0053] If not, a negative reduction adjustment is performed based on the frequency adjustment correction amount. Specifically, the frequency adjustment correction amount is input into the connection time-dynamic update frequency mapping table for querying to obtain the dynamic update frequency reduction factor. The dynamic update frequency reduction factor is then multiplied by the current skill profile dynamic update frequency to obtain the reduced skill profile dynamic update frequency. The frequency adjustment correction amount represents the negative difference between the dynamic update frequency adjustment factor and the frequency adjustment benchmark value. By reasonably reducing the update frequency, unnecessary computational resource consumption and data transmission are reduced while ensuring the timeliness of skill profiles meets basic teaching needs, alleviating system load pressure and reducing resource competition during cross-scene task connection.

[0054] In this embodiment, a smart control mechanism based on the time consumption of cross-scenario teaching task connections is established to achieve precise optimization of the dynamic update frequency of skill profiles. When the connection time is within the normal range, the system maintains the existing update frequency to avoid unnecessary control interference and ensure the stability of the skill profile assessment process. When the connection time exceeds the standard, the system couples the connection time comparison with the teaching coordination comparison, obtains the dynamic update frequency adjustment factor through a mapping table, and initiates differentiated control based on its comparison with the benchmark value: if the adjustment factor exceeds the benchmark value, the system obtains the dynamic update frequency gain factor based on the frequency adjustment offset, and increases the skill profile update frequency through doubling. This mechanism can promptly correct the student ability assessment model by accelerating the profile update speed when there is a significant blockage in the teaching process, thereby quickly improving the matching degree of subsequent teaching task recommendations; if the adjustment factor does not reach the benchmark value, the system obtains the reduction factor based on the frequency adjustment correction and reduces the update frequency accordingly, achieving reasonable saving of computing resources while ensuring teaching effectiveness. This multi-level control strategy based on dual-parameter analysis ensures real-time coordination between skill profile updates and teaching fluency, while maximizing the efficiency of system resource utilization, providing a precise and adaptive dynamic optimization scheme for skill profiles for the learning and training system.

[0055] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. An adaptive learning and training integration teaching system based on artificial intelligence, characterized in that, It includes a data perception and integration module, an integration degree control module, and a teaching coordination and control module: The data perception and fusion module is used to collect learning and training data from different learning scenarios and perform fusion processing, obtain fusion parameters in the fusion processing process, and obtain the fusion degree of learning and training data ecosystem based on the fusion parameters, which is used to characterize the degree of fusion of multi-source heterogeneous learning and training teaching data from different learning scenarios. The integration degree control module is used to determine whether to perform ecological integration degree control based on the integration degree of learning and training data. If yes, an adaptive teaching scheduling step is performed after ecological integration degree control; otherwise, an adaptive teaching scheduling step is performed directly. The ecological integration degree control includes adaptive control of reuse frequency and adaptive control of integration trigger threshold. The teaching coordination and control module is used to obtain coordination parameters in the adaptive teaching scheduling process, obtain the coordination efficiency of learning and training integration teaching based on the coordination parameters, and characterize the coordination effectiveness of learning and training integration teaching in the process of learning and training integration digital transformation. Based on the coordination efficiency of learning and training integration teaching, it is determined whether to execute the integration teaching coordination and control. If yes, the integration teaching coordination and control outputs the learning and training integration teaching closed-loop instruction and feeds it back to the data perception and integration module to realize the leap of learning and training integration teaching in digital transformation. If no, the integration teaching closed-loop instruction is directly output and fed back to the data perception and integration module. The integration teaching coordination and control includes self-updating frequency adaptive control and dynamic update frequency adaptive control. The data perception and fusion module includes a fusion parameter unit and a fusion parameter quantization unit; The fusion parameter unit is used to collect fusion parameters, including the connectivity rate of the learning and training data ecosystem interface, the learning and training data flow delay, and the calling frequency of the learning and training data service interface. The fusion parameter quantification unit is used to obtain the integration degree of the learning and training data ecosystem through fusion parameters. Specifically, it multiplies the connectivity compensation factor and the connectivity component to obtain a connectivity correction value, where the connectivity component represents the result of the analysis of the ratio of the connectivity rate of the learning and training data ecosystem interface to the connectivity threshold; it multiplies the flow delay compensation factor and the flow delay component to obtain a flow delay correction value, where the flow delay component represents the result of the analysis of the ratio of the flow delay threshold to the flow delay of the learning and training data; it multiplies the call frequency compensation factor and the call frequency component to obtain a call frequency correction value, where the call frequency component represents the result of the analysis of the ratio of the call frequency of the learning and training data service interface to the call frequency threshold; and it couples the connectivity correction value, the flow delay correction value, and the call frequency correction value to obtain the integration degree of the learning and training data ecosystem. The teaching coordination and control module includes a coordination parameter unit, a coordination parameter quantification unit, and an integrated teaching coordination and control unit. The coordination parameter unit is used to collect coordination parameters, including the degree of ecological integration to be compared, the frequency of dynamic adjustment of student teaching paths, and the time consumption of connecting cross-scenario teaching tasks. The coordination parameter quantification unit is used to obtain the coordination efficiency of learning-training integration teaching through coordination parameters. Specifically, it multiplies the integration degree compensation factor and integration degree component to obtain an integration degree correction value, where the integration degree component represents the result of the analysis of the ratio of the integration degree of the compared ecology to the integration degree threshold; it multiplies the adjustment frequency compensation factor and adjustment frequency component to obtain an adjustment frequency correction value, where the adjustment frequency component represents the result of the analysis of the ratio of the dynamic adjustment frequency of the student teaching path to the adjustment frequency threshold; it multiplies the connection time compensation factor and connection time component to obtain a connection time correction value, where the connection time component represents the result of the analysis of the ratio of the connection time threshold to the connection time of cross-scenario teaching task connection; and it couples the integration degree correction value, adjustment frequency correction value, and connection time correction value to obtain the coordination efficiency of learning-training integration teaching.

2. The adaptive learning and training integration teaching system based on artificial intelligence according to claim 1, characterized in that, The specific steps for determining whether to perform ecological integration degree regulation are as follows: If the integration degree of the learning and training data ecosystem is greater than or equal to the integration degree benchmark value, then no integration degree regulation will be implemented; otherwise, adaptive regulation of reuse frequency and adaptive regulation of integration trigger threshold will be implemented based on the integration degree reference value. The integration degree reference value represents the negative deviation between the integration degree of the learning and training data ecosystem and the integration degree benchmark value. The specific process for implementing adaptive adjustment of reuse frequency is as follows: If the response threshold of the learning and training data ecosystem interface call is less than the lower bound of the response threshold, it is determined to be in an idle boosting state. The response threshold deviation is input into the response threshold-reuse frequency mapping table for querying, and the reuse frequency compensation update trigger instruction is obtained. Specifically, the result of harmonic averaging of the response threshold deviation and the integration degree comparison is input into the response threshold-reuse frequency mapping table for querying, and the reuse frequency gain factor is obtained. The current learning and training data active reuse frequency and the reuse frequency gain factor are multiplied to obtain the target learning and training data active reuse frequency. The response threshold deviation represents the negative deviation between the learning and training data ecosystem interface call response threshold and the lower bound of the response threshold. If the response threshold of the learning and training data ecosystem interface call is within the response threshold range, it is determined to be in the optimal lazy state, and the adaptive adjustment of the reuse frequency will not be performed. The current active reuse frequency of learning and training data will be maintained. The response threshold range refers to the closed interval formed by the lower bound of the response threshold and the upper bound of the response threshold.

3. The adaptive learning and training integration teaching system based on artificial intelligence according to claim 2, characterized in that, The adaptive adjustment of the multiplexing frequency also includes: If the response threshold of the learning and training data ecosystem interface call is greater than the upper limit of the response threshold, it is determined to be in an over-limit high-pressure state. The response threshold comparison value is input into the response threshold-reuse frequency mapping table for querying to obtain a dynamic reuse window shrinkage instruction. Specifically, the result of harmonic averaging of the response threshold comparison value and the integration degree comparison value is input into the response threshold-reuse frequency mapping table for querying to obtain the reuse window shrinkage factor. The reuse window shrinkage factor is multiplied by the current reuse window to obtain the target reuse window. The result of multiplying the target reuse window by the current learning and training data active reuse frequency and rounding up is taken as the target learning and training data active reuse frequency. The response threshold comparison value represents the positive difference between the learning and training data ecosystem interface call response threshold and the upper limit of the response threshold.

4. The adaptive learning and training integration teaching system based on artificial intelligence according to claim 2, characterized in that, The specific steps for adaptive adjustment of the fusion trigger threshold are as follows: If the traceability rate of the student's panoramic profile is within the traceability rate tolerance range, then the adaptive adjustment of the fusion trigger threshold will not be executed. The traceability rate tolerance range refers to the closed interval formed by the lower limit of the traceability rate tolerance and the upper limit of the traceability rate tolerance. If the traceability rate of the student-training panoramic profile is less than the traceability rate tolerance limit, the traceability rate offset and the fusion degree comparison value are input into the traceability rate-fusion trigger threshold mapping table to obtain the fusion trigger upper limit threshold gain factor. The fusion trigger upper limit threshold gain factor is multiplied by the current fusion trigger upper limit threshold to obtain the target student-training cross-scene data fusion trigger upper limit threshold. The traceability rate offset represents the negative deviation between the student-training panoramic profile traceability rate and the traceability rate tolerance limit. If the traceability rate of the student-training panoramic profile is greater than the traceability rate tolerance limit, the traceability rate comparison quantity and the fusion degree comparison quantity are input into the traceability rate-fusion trigger threshold mapping table to obtain the fusion trigger lower limit threshold reduction factor. The fusion trigger lower limit threshold reduction factor is multiplied by the current fusion trigger upper limit threshold to obtain the target student-training cross-scene data fusion trigger lower limit threshold. The traceability rate comparison quantity represents the positive deviation between the student-training panoramic profile traceability rate and the traceability tolerance limit.

5. The adaptive learning and training integration teaching system based on artificial intelligence according to claim 1, characterized in that, The integrated teaching coordination and control unit is used to determine whether to implement integrated teaching coordination and control. Specifically, if the learning and training integrated teaching coordination efficiency is greater than or equal to the coordination efficiency benchmark value, then the integrated teaching coordination and control is not implemented. Otherwise, based on the teaching coordination reference value, the self-updating frequency adaptive control and the dynamic update frequency adaptive control are implemented. The teaching coordination reference value represents the negative deviation between the learning and training integrated teaching coordination efficiency and the coordination efficiency benchmark value. The specific process for implementing adaptive adjustment of self-update frequency is as follows: If the trigger interval threshold for the bridging task reminder is less than the lower limit of the trigger interval tolerance, the interval threshold comparison value and the teaching coordination comparison value are entered into the interval threshold-self-update frequency mapping table for querying to obtain the knowledge graph self-update interval reduction factor. The knowledge graph self-update interval reduction factor is multiplied by the current self-update interval and then rounded down. The result is then multiplied by the current learning and training integration knowledge graph self-update frequency to obtain the target learning and training integration knowledge graph self-update frequency. The interval threshold comparison value represents the negative deviation between the bridging task reminder trigger interval threshold and the lower limit of the trigger interval tolerance.

6. The adaptive learning and training integration teaching system based on artificial intelligence according to claim 5, characterized in that, The adaptive adjustment of the self-update frequency also includes: If the trigger interval threshold for the task reminder is within the trigger interval tolerance range, the self-update frequency adaptive adjustment will not be performed, and the current self-update frequency of the learning and training integration knowledge graph will be maintained. The trigger interval tolerance range represents the closed interval formed by the lower limit of the trigger interval tolerance and the upper limit of the trigger interval tolerance. If the trigger interval threshold for the bridging task reminder is greater than the upper limit of the trigger interval tolerance, the interval threshold offset and the teaching coordination comparison amount are input into the interval threshold-self-update frequency mapping table for querying to obtain the self-update frequency multiplication factor. The self-update frequency multiplication factor is multiplied by the current learning and training integration knowledge graph self-update frequency and then rounded up to obtain the target learning and training integration knowledge graph self-update frequency. The interval threshold offset represents the positive deviation between the trigger interval threshold for the bridging task reminder and the upper limit of the trigger interval tolerance.

7. The adaptive learning and training integration teaching system based on artificial intelligence according to claim 5, characterized in that, The specific process for the dynamic update frequency adaptive adjustment is as follows: If the time taken to connect cross-scenario teaching tasks is less than or equal to the connection time, then the dynamic update frequency adaptive adjustment will not be executed. If the time taken to connect cross-scenario teaching tasks is greater than the total time taken to connect, the connection time comparison quantity and the teaching coordination comparison quantity are entered into the connection time-dynamic update frequency mapping table for querying to obtain the dynamic update frequency adjustment factor. It is then determined whether the dynamic update frequency adjustment factor is greater than the frequency adjustment benchmark value. The connection time comparison quantity represents the positive deviation between the connection time taken to connect cross-scenario teaching tasks and the total time taken to connect.

8. The adaptive learning and training integration teaching system based on artificial intelligence according to claim 7, characterized in that, The adaptive adjustment of the dynamic update frequency also includes: If so, positive boosting control is performed based on the frequency adjustment offset. Specifically, the frequency adjustment offset is input into the connection time-dynamic update frequency mapping table for querying to obtain the dynamic update frequency gain factor. The dynamic update frequency gain factor is multiplied by the current skill profile dynamic update frequency to obtain the adjusted skill profile dynamic update frequency. The frequency adjustment offset represents the positive deviation between the dynamic update frequency adjustment factor and the frequency adjustment benchmark value. If not, then a negative reduction adjustment is made based on the frequency adjustment correction amount. Specifically, the frequency adjustment correction amount is input into the connection time-dynamic update frequency mapping table for querying to obtain the dynamic update frequency reduction factor. The dynamic update frequency reduction factor is multiplied by the current skill profile dynamic update frequency to obtain the reduced skill profile dynamic update frequency. The frequency adjustment correction amount represents the negative deviation between the dynamic update frequency adjustment factor and the frequency adjustment benchmark value.

Images