AI-driven Education Talent Competency Assessment and Career Path Planning System
By using an AI-driven educational talent competency assessment system, which utilizes multi-source data and dynamic competency modeling, personalized growth paths are generated. This solves the problem of inconsistent competency growth paths in existing educational talent evaluations and enables dynamic reflection and cross-scenario transfer verification of competency assessments.
Patent Information
- Application Number
- CN202611333249.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-08-31
- Publication Date
- 2026-11-17
AI Technical Summary
The existing education talent evaluation system is unable to accurately reflect the ability to apply knowledge, solve problems, design teaching, and transfer knowledge across different scenarios. Furthermore, it lacks a dynamic adjustment mechanism for the ability development path, resulting in inconsistencies between the evaluation results and actual abilities.
The AI-driven education talent competency assessment and growth path planning system establishes a multi-layered competency map through modules such as multi-source data collection, competency map construction, dynamic competency status modeling, competency gap calculation, and dynamic path correction. It performs competency evidence analysis and transfer verification, generates personalized growth paths, and conducts real-time monitoring and deviation location during the execution process.
It enables dynamic reflection of educational talent competency evaluation, improves the authenticity and consistency of evaluation, ensures the relevance and adaptability of growth paths, solves the problem of completing courses but not developing competencies, and enhances the verification of cross-scenario transferability.
Smart Images

Figure CN123117908A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of digital education technology, specifically to an AI-driven system for assessing educational talent capabilities and planning career paths. Background Technology
[0002] With the development of artificial intelligence, big data, and digital education technologies, educational talent cultivation is gradually shifting from traditional, standardized curriculum training to a personalized, continuous, and competency-oriented model. Current educational talent evaluation typically relies on single or limited-dimensional data such as exam scores, course completion, homework grades, teacher evaluations, or questionnaires. The resulting assessments are usually static scores or grades, failing to accurately reflect an educator's ability to apply knowledge, solve problems, design lessons, and transfer knowledge across different learning environments. Meanwhile, existing AI-powered educational competency frameworks are beginning to shift from simple knowledge evaluation to multi-dimensional competency and tiered development. For example, UNESCO's AI competency framework for teachers divides teacher competencies into multiple dimensions and different development levels, explicitly proposing the need to identify strengths and weaknesses in different competencies through diagnostic tools to prioritize personalized training and plan learning paths.
[0003] However, existing technologies such as capability frameworks, knowledge graphs, learning recommendations, or knowledge state tracking typically handle capability definition, knowledge evaluation, or course recommendation functions respectively. There is a lack of a continuous state transfer mechanism between evaluation results and the growth path, making it particularly difficult to handle situations such as "course completion but no actual capability development," "capability performing well in one scenario but unable to transfer to other scenarios," and "capabilities that have been developed deteriorating due to long-term disuse." When the actual capabilities of talent change inconsistent with the pre-set training path, existing systems usually re-recommend courses, failing to further pinpoint the specific capability nodes causing insufficient capability growth and their upstream dependent nodes. Summary of the Invention
[0004] The purpose of this invention is to provide an AI-driven system for assessing educational talent capabilities and planning career paths, in order to solve the technical problems mentioned in the background section.
[0005] To achieve the above objectives, this invention provides the following technical solution: an AI-driven educational talent competency assessment and growth path planning system, comprising a multi-source educational talent data acquisition module, a competency graph construction module, a competency evidence analysis module, a dynamic competency status modeling module, a competency gap calculation module, a growth path generation module, a competency transfer verification module, and a path dynamic correction module. The capability graph construction module establishes a multi-layer capability graph consisting of knowledge nodes, capability nodes, task nodes, job nodes, and learning resource nodes, and establishes prerequisite, dependency, and transfer relationships between capability nodes. The capability evidence analysis module weights each capability piece of evidence based on the reliability of the evidence source, the authenticity of the task, the degree of time decay, and the independence of the evidence, forming a capability evidence set. The dynamic capability state modeling module calculates the current capability value, capability change trend, forgetting risk, and cross-scenario migration value of each capability node based on the capability evidence set, forming a dynamic capability state vector. The capability gap calculation module calculates the capability gap based on the target capability state vector and dynamic capability state vector corresponding to the target position, and forms a capability gap chain by topological sorting the capability gap according to the prerequisites and dependencies. The growth path generation module selects knowledge learning tasks, ability training tasks, practice tasks, and transfer verification tasks from learning resource nodes and task nodes based on the capability gap chain, and generates a growth path that progresses step by step from the current capability state to the target capability state.
[0006] Preferably, the multi-source education talent data collection module collects knowledge test data, course learning data, homework results, teaching practice data, project task data, evaluation feedback data, and target job data of education talents, and adds timestamps, task scenarios, ability tags, evidence sources, and evidence types to each data. The capability transfer verification module generates a transfer task that corresponds to the target capability and is different from the training task after the learning task is executed, and updates the corresponding capability node based on the performance of the transfer task. The path dynamic correction module compares the deviation between the predicted capability trajectory and the actual capability trajectory during the execution of the growth path. When the deviation exceeds a preset threshold, it locates the corresponding capability node and reconstructs only the capability node and its associated growth sub-paths. The updated dynamic capability status is used as the input for the generation of the growth path in the next cycle.
[0007] Preferably, the capability graph construction module establishes a mapping relationship according to the hierarchy of "knowledge-capability-task-position" and establishes directional capability dependency edges between capability nodes; Each capability dependency edge carries a dependency strength parameter, which is used to represent the degree of influence of the preceding capability on the formation of the subsequent capability; The growth path generation module sorts the capability gaps according to the direction of the capability dependency edges, and prioritizes the gap nodes that are located upstream of the target capability and have a high dependency strength as the priority training nodes.
[0008] Preferably, the capability evidence analysis module sets corresponding evidence credibility coefficients for different types of evidence, and calculates the comprehensive evidence weight based on the reliability of the evidence source, the authenticity of the task, the freshness of time, and the degree of repetition between different types of evidence. When multiple sources of evidence point to the same capability node, repetitive evidence with the same source or highly related content is subject to correlation suppression, and the overall weight of capability evidence from different task scenarios and which are independent of each other is increased, so as to reduce the abnormal impact of a single evaluation result on the capability status.
[0009] Preferably, the dynamic capability status modeling module establishes a capability status parameter group for each capability node, consisting of the current capability value, capability growth rate, capability stability, forgetting risk value, and scenario migration value; The forgetting risk value is calculated based on the most recent valid verification time of the capability node, historical usage frequency, number of consecutive valid performances, and number of task scenarios covered. When the forgetting risk value exceeds a preset threshold, a reproduction task and a re-verification task for the capability node are inserted into the growth path.
[0010] Preferably, the capability gap calculation module not only calculates the numerical difference between the target capability value and the current capability value, but also calculates the capability gap priority based on the dependency strength, migration value and capability importance in the target position of the capability node, and forms a capability gap chain with sequential constraints according to the capability gap priority and capability dependency relationship, so that subsequent growth tasks can only enter the next capability stage after meeting the qualification conditions of their corresponding preceding capability nodes.
[0011] Preferably, the growth path generation module configures at least one knowledge learning task, one ability training task, one scenario practice task, and one transfer verification task for each ability gap node, and calculates the path evaluation value based on the expected ability gain, task difficulty, expected learning cost, learning load, and ability transfer benefit corresponding to each task, and selects the growth path whose path evaluation value meets the preset conditions from the candidate paths that meet the ability dependency constraints.
[0012] Preferably, the transfer task generated by the capability transfer verification module has different task parameters, task scenarios, or task data than the corresponding training task. The module calculates the transfer capability value based on the accuracy of the educational personnel in completing the transfer task, the completeness of the problem-solving steps, the rationality of the task resource call, and the stability of the results. When the performance of the training task reaches the set threshold but the transfer capability value does not reach the set threshold, the capability level of the corresponding capability node remains unchanged, and the capability node is marked as a node with insufficient knowledge acquisition but insufficient capability transfer.
[0013] Preferably, the path dynamic correction module generates a predicted capability trajectory based on the current capability status, task sequence, and historical capability change data before the growth path is executed. During the execution process, it acquires the actual capability status at preset time intervals and forms the actual capability trajectory. It performs capability node-level positioning based on the difference between the predicted capability trajectory and the actual capability trajectory, and judges whether the capability growth is insufficient, the capability growth is too fast, or the capability status is degraded based on the direction of the difference. When it is determined that the capability growth is insufficient, the path dynamic correction module further traces the upstream capability node that caused the insufficient growth of the capability node from the capability graph, and determines whether the current capability value of the upstream capability node is lower than the corresponding pre-threshold. When the upstream capability node is lower than the corresponding pre-threshold, the subsequent training tasks in the original growth path are replaced with the compensation training tasks corresponding to the upstream capability node. When the upstream capability node reaches the pre-threshold, the task corresponding to the upstream capability node remains unchanged and the practice task or transfer verification task of the current capability node is adjusted.
[0014] Preferably, it also includes a capability growth profile module. This module records the capability status, capability evidence, task performance, transfer verification results, forgetting risk, and path adjustment records for each capability node of the educational talent according to a time sequence, and generates a capability growth trajectory based on the capability status changes within different time periods. The growth path generation module calls the capability growth trajectory during the next path planning, reduces the repetition training priority of capability nodes that have been completed and meet the transfer capability standards, and increases the training priority of capability nodes that have deteriorated or have insufficient cross-scenario transfer, thereby forming a closed-loop growth path based on continuous updates of historical capability status. Compared with the prior art, the beneficial effects of the present invention are: 1. This invention introduces the reliability of evidence sources, the authenticity of tasks, the freshness of time, and the independence of evidence for different types of ability evidence, forming a credible set of ability evidence. It realizes unified ability mapping and dynamic ability profile of multi-source education talent data, and solves the problems of existing evaluation methods such as single data sources, single assessments that are prone to evaluation bias, and static scores that cannot reflect the development and degradation of abilities. 2. This invention establishes a capability gap chain generation mechanism based on reverse-driven target positions, realizing reverse diagnosis from "target position capability requirements" to "current capability status of talent", and automatically determining the capability training sequence, solving the problems of existing personalized learning systems that are prone to course stacking, insufficient correlation between recommended content and job capability requirements, and inability to determine the training sequence. 3. This invention distinguishes between "completing learning tasks" and "developing practical abilities," verifying whether educational talents truly possess the ability to apply learned knowledge and skills to new tasks and scenarios. It addresses the problem in existing educational talent evaluations of "high course completion rates and high test scores, but insufficient practical task abilities," shifting the assessment of abilities from "whether they have learned" to "whether they can apply and transfer their knowledge," thereby improving the consistency between ability assessments and actual job skills. 4. Before the growth path is executed, this invention generates a predicted ability trajectory based on the talent's current ability status, task sequence, and historical ability change data. During the execution of the growth path, the actual ability status is continuously acquired, and ability node-level deviation analysis is performed on the predicted ability trajectory and the actual ability trajectory. This enables real-time monitoring, deviation location, and local path reconstruction during the execution of the growth path, solving the problems of traditional learning paths lacking dynamic adjustment capabilities once generated and having to completely re-recommend courses when learning effect deviations occur. Attached Figure Description
[0015] Figure 1 This is a block diagram of the overall system structure of the present invention; Figure 2 This is a business closed-loop flowchart of the present invention; Figure 3 This is a structural diagram of the multilayer capability map of the present invention; Figure 4 This is the path dynamic correction logic diagram of the present invention. Detailed Implementation
[0016] Please see Figure 1 , Figure 2 , Figure 3 and Figure 4 : Example 1: Dynamic Competency Assessment of Educational Talents Based on Multi-Source Competency Evidence This embodiment uses an educational professional who needs to grow from an ordinary teacher to a senior digital teaching professional as the evaluation object to illustrate the AI-driven educational professional competence assessment and growth path planning system described in this invention.
[0017] In this embodiment, the system includes a multi-source education talent data acquisition module, a competency map construction module, a competency evidence analysis module, a dynamic competency status modeling module, a competency gap calculation module, a growth path generation module, a competency transfer verification module, a path dynamic correction module, and a competency growth archive module.
[0018] First, the multi-source education talent data collection module acquires multi-source data generated by education talents within a preset evaluation period. This multi-source data includes at least knowledge test data, online course learning data, course assignment data, instructional design data, simulated classroom task data, real teaching task data, student feedback data, peer evaluation data, and target job position data.
[0019] To avoid inconsistencies in time scale and evaluation scale between data from different sources, a unified data structure is generated for each piece of original data: [Ej=(I,t,s,a,r,q)] Wherein, (I) represents the educational talent identifier, (t) represents the time when the data occurred, (s) represents the task scenario, (a) represents the associated capability node, (r) represents the data source, and (q) represents the evidence type.
[0020] For example, if the talent completes the course "AI-based Classroom Instruction Design" on August 1, 2026, and submits a teaching plan on August 3, 2026, then the course completion record and the teaching plan each constitute two pieces of competency evidence. The course completion record is associated with the "AI Teaching Tool Cognition Ability" node, and the teaching plan is associated with the "Digital Instructional Design Ability" node.
[0021] The system further analyzes the teaching objectives, teaching activities, teaching resources, evaluation methods and classroom feedback in the teaching plan in a structured way, and maps them to the corresponding ability nodes.
[0022] In terms of capability graph construction, this embodiment establishes a five-layer heterogeneous capability graph including knowledge nodes, capability nodes, task nodes, job nodes, and learning resource nodes.
[0023] For example: "Basic Principles of Generative AI" belongs to the knowledge node; "Application Ability of AI Teaching Tools" belongs to the ability node; "Designing Classroom Teaching Plans Using AI" belongs to the task node; "Advanced Digital Teaching Talents" belongs to the job node; and "Practical Courses of AI Teaching Tools" belongs to the learning resource node.
[0024] Directed edges are established between the aforementioned nodes, for example: Generative AI Basic Principles → AI Teaching Tool Application Ability; AI Teaching Tool Application Ability → AI-Assisted Instructional Design Ability; AI-Assisted Instructional Design Ability → Smart Classroom Implementation Ability; Smart Classroom Implementation Ability → Advanced Digital Teaching Talent. A dependency strength parameter is set for the dependency relationships between adjacent capability nodes, and the dependency strength is represented by a normalized value of 0 to 1.
[0025] Furthermore, the capability evidence analysis module calculates the credibility of evidence based on the reliability of the source of different evidence, the authenticity of the task, the freshness of time, and the independence of the evidence.
[0026] In one specific implementation, the weight of the comprehensive evidence is set as follows: W j =W {rj} ×W {sj} ×W {tj} ×W {dj} Among them, (W) {rj} (W) indicates the reliability of the source of evidence. {sj} ) indicates the authenticity of the task, (W) {tj} (W) indicates time freshness. {dj} () indicates the independence of evidence.
[0027] For example: the source reliability of knowledge test results is set to 0.80; the source reliability of real teaching task results is set to 0.95; the task authenticity coefficient of simulated teaching tasks is set to 0.85; and the task authenticity coefficient of real classroom tasks is set to 0.95.
[0028] For duplicate data generated by the same test questions, the system reduces its evidence independence coefficient; for independent practice data from different times and different task scenarios, the system increases the evidence independence coefficient.
[0029] Based on this, the dynamic capability state modeling module establishes a capability state parameter set for each capability node: [C i (t)={M i ,T i ,S i ,F i G i}]; where: (M i (T) represents the current ability value; i (S) indicates the trend of change in ability; i ) represents the stability of capability; (F i ) indicates the risk of forgetting ability; (G) i ) represents the cross-scene migration value.
[0030] The current capability value is updated recursively based on recent capability evidence and historical capability status. For example, for the i-th capability node, after acquiring new evidence, the historical capability status is adjusted according to the evidence weight, so that the latest practical evidence has a greater impact on the current capability status.
[0031] The trend of capability change is obtained by differentiating capability values over at least two consecutive evaluation periods.
[0032] Ability stability is calculated based on the dispersion of ability performance across different tasks. When the performance of the same ability varies little across different tasks, the ability stability is considered high.
[0033] The risk of forgetting is calculated based on the most recent valid verification time, historical usage frequency, number of consecutive valid performances, and the number of scenarios covered. When a capability has not been used for a long period of time and there is no new valid verification evidence, the risk of forgetting is increased.
[0034] Ultimately, the dynamic competency status of this educational talent is obtained, for example: Teacher Digital Competency Node Data Table ; As a result, the system no longer only outputs a "comprehensive ability score", but forms a dynamic ability profile that can reflect the talent's ability level, development direction, stability, forgetting status and scenario migration.
[0035] This embodiment solves the problem that traditional talent evaluation relies solely on exam scores or course completion rates, which fails to reflect true teaching ability.
[0036] Example 2: Generation of Competency Gap Chain and Growth Path Based on Back-Calculation of Target Position This embodiment, based on Embodiment 1, further explains how to reverse-generate a career path for educational professionals based on the target position.
[0037] First, the competency requirements for the target position "senior digital teaching talent" are obtained from the target position competency analysis module.
[0038] The target job competency requirements include: a target value of 0.85 for basic digital education competency; a target value of 0.88 for AI tool application competency; a target value of 0.85 for digital instructional design competency; a target value of 0.82 for smart classroom implementation competency; a target value of 0.80 for instructional evaluation optimization competency; and a target value of 0.78 for cross-scenario instructional transfer competency. The system converts these target values into target competency state vectors. Then, the competency gap calculation module calculates the difference between the target competency state vector and the current competency state vector obtained in Example 1. [D i =C i * -C i ]; Among them, (C) i * ) represents the target capability value, (C) i () represents the current ability value.
[0039] For example: The gap in AI tool application capabilities is: [0.88-0.72=0.16]; The digital instructional design capability gap is: [0.85-0.56=0.29]; The capacity gap for implementing smart classrooms is: [0.82 - 0.48 = 0.34]; The gap in teaching evaluation optimization capabilities is: [0.80-0.42=0.38]; However, the system does not directly recommend training content based on the size of the gap, but further reads the dependencies in the capability graph.
[0040] For example, the ability to optimize teaching evaluation depends on the ability to design digital instructional content and implement smart classrooms; the ability to implement smart classrooms, in turn, depends on the ability to apply AI tools and design digital instructional content.
[0041] Therefore, although the numerical gap in teaching evaluation optimization capability is the largest, it is not currently the highest priority training node. The system is built according to dependencies: AI tool application capability → digital instructional design capability → smart classroom implementation capability → teaching evaluation optimization capability.
[0042] Further calculate the priority of capability gaps based on job importance, capability gaps, dependence intensity, and migration value.
[0043] In this embodiment, the capacity gap priority is set as: P i =D i ×I i ×R i ; Among them, (D) i ) represents the capacity gap value, (I i (R) indicates the importance of the position. i () indicates the value of ability transfer.
[0044] The system forms a capability gap chain based on the priorities and capability dependencies.
[0045] After forming the capability gap chain, the growth path generation module matches at least four types of tasks for each capability gap node: The first category is knowledge learning tasks, which are used to fill gaps in basic knowledge. The second category is skills training tasks, which are used to develop specific operational skills; The third category is scenario practice tasks, which are used to verify the practical application of capabilities in target scenarios; The fourth category is migration verification tasks, which are used to determine whether capabilities can be transferred to different task scenarios.
[0046] For example, the following tasks are generated for "digital instructional design ability": Task A: Learn the basics of digital instructional design; Task B: Complete the design of a standard course teaching plan; Task C: Complete AI-assisted instructional design for a real course; Task D: Redesign the teaching plan under different subject conditions and conduct transfer verification.
[0047] The growth path generation module further configures the expected capability gains, task difficulty, expected learning time, learning load, and transfer benefits for each task.
[0048] Let the evaluation value of the candidate growth path be: [V(P)=αG(P)+βR(P)+γE(P)-ΔC(P)-εL(P)]; Where: α-ε are fixed parameters determined by the proportions of each component; (G(P)) is the expected capability gain; (R(P)) is the migration gain; (E(P)) is the expected probability of achieving the goal; (C(P)) is the learning cost; and (L(P)) is the learning load. The system selects the path with the highest evaluation value or that exceeds the preset threshold from the candidate paths that satisfy the capability dependency relationship.
[0049] For example, the final output might be: AI tool basics review → AI tool operation training → AI-assisted instructional design → Real teaching case practice → Interdisciplinary instructional design → Transfer verification → Smart classroom implementation → Integrated teaching project.
[0050] This path is not simply generated according to the course catalog, but is calculated in reverse from the difference between the current ability status and the ability status of the target position.
[0051] Example 3: Verification of Real Capabilities Based on Capability Transfer Validation This embodiment further illustrates how the present invention solves the problem that "completing a course does not equate to developing competence".
[0052] In traditional learning systems, when a student completes an "AI-assisted instructional design" course and passes the course test, their corresponding ability score is usually directly improved.
[0053] This invention does not use the above method, but distinguishes between course completion and competency development.
[0054] For example, education professionals first complete the "AI-assisted instructional design" training task.
[0055] The training tasks adopt fixed teaching themes, requiring talents to use AI tools to generate teaching objectives, teaching activities, teaching materials and evaluation schemes.
[0056] The system calculates the training capability value based on the quality of task completion.
[0057] When the training capability value reaches 0.80 or above, the capability is not immediately considered to have met the standard. Instead, the capability transfer verification module generates a new transfer task.
[0058] The transfer task shall change at least one of the following: teaching theme; target audience; teaching objectives; teaching resources; teaching constraints; evaluation requirements.
[0059] For example, if the training task is "Teaching Design for Middle School Mathematics Functions", the transfer task will automatically generate "Teaching Design for High School Physics Experiment Course".
[0060] Although both tasks fall under the category of "digital instructional design" skills, their specific knowledge content and task parameters differ.
[0061] The system extracts the following aspects of talent in transfer tasks: task completion accuracy; completeness of problem-solving steps; rationality of teaching resource allocation; degree of matching between teaching objectives and teaching activities; completeness of teaching evaluation schemes; and stability of results. It also calculates the transferability value: [G] i =λ1A+λ2B+λ3C+λ4D+λ5E]; Wherein: λ1-λ5 represent weight coefficients, dynamically adjusted according to the proportion; (A) represents the accuracy of task completion; (B) represents the completeness of problem-solving steps; (C) represents the rationality of resource allocation; (D) represents the completeness of the teaching plan; (E) represents the stability of results in different transfer tasks. When the training ability value is 0.86 and the transfer ability value is only 0.55, the system determines that the talent has acquired relevant knowledge and preliminary operational skills, but has not yet formed a stable cross-scenario transfer ability. At this time, the system performs the following processing: First, it does not directly mark the ability node as fully qualified; Second, it marks the ability node as "knowledge acquired but ability transfer insufficient"; Third, it searches for different scenario tasks associated with the ability node from the ability graph; Fourth, it generates a cross-scenario practice task with lower difficulty; Fifth, it regenerates the transfer verification task after the practice task is completed.
[0062] For example, the first transfer failure occurred in "interdisciplinary instructional design", and the system further generated a training task of "transformation of teaching objectives of different disciplines".
[0063] A capability node is marked as "stable migration" only when the migration capability value reaches 0.75 or higher in two consecutive different scenarios.
[0064] This transforms competency assessment from "learning completion" to "learning completion → competency training → real-world tasks → cross-scenario transfer → stable competency confirmation".
[0065] Example 4: Dynamic Path Correction Based on Predicted Capability Trajectory and Actual Capability Trajectory This embodiment illustrates how dynamic correction is performed during the execution of the growth path.
[0066] After the growth path is generated, the path dynamic correction module generates a predicted ability trajectory based on the current ability status of the educational talent, the task sequence, the historical rate of change of ability, and the expected ability gain of the task.
[0067] For example, the predicted digital instructional design capabilities over six consecutive growth cycles were 0.56, 0.61, 0.67, 0.72, 0.79, and 0.85, respectively.
[0068] After the educational talent actually implemented the path, the system obtained the actual ability status respectively: 0.56, 0.60, 0.64, 0.65, 0.66, 0.67.
[0069] The system calculates the deviation between the predicted capability trajectory and the actual capability trajectory.
[0070] When the cumulative deviation of a certain capability node exceeds a preset threshold, such as 0.10, dynamic path correction is triggered.
[0071] However, instead of regenerating the entire learning path, this invention first identifies the specific capability nodes that cause the deviation.
[0072] For example, if it is found that the growth of digital instructional design capabilities is insufficient, then trace upstream along the capability map: digital instructional design capability ← AI tool application capability ← basic digital education capability.
[0073] The system further checks the upstream capability status.
[0074] If the AI tool application capability is found to be 0.83, which exceeds the pre-requisite threshold of 0.80, then the problem is not a lack of AI tool foundation, but rather a lack of practical ability in digital instructional design.
[0075] At this point, the original AI tool learning tasks are retained, and only the practical tasks corresponding to digital instructional design capabilities are replaced.
[0076] For example, the original task was: "Complete a standard course teaching plan." The adjusted task is: "Complete a teaching plan for a real-world scenario with constraints on teaching resources, differences in student foundations, and time limitations." By increasing the complexity of the task scenario, the training task is made closer to the target ability. If the system detects that the AI tool application ability is only 0.68, below the pre-threshold of 0.80, it determines that the insufficient growth of digital instructional design ability stems from upstream capability nodes. In this case, advanced digital instructional design training is not continued; instead, the following is inserted into the current path: AI tool basic compensation training → AI tool specialized operation. →AI tool migration task →Revalidate AI tool application capabilities →Resume digital instructional design training. This forms a local path reconstruction tailored to specific reasons. In this embodiment, the principles of local path reconstruction are: when upstream capabilities are insufficient, supplement upstream capabilities; when upstream capabilities meet the standard but current capabilities are insufficient, adjust the current capability training task; when training capabilities meet the standard but transfer capabilities are insufficient, add cross-scenario practice tasks; when capabilities reach the target but the risk of forgetting increases, add reproduction and revalidation tasks.
[0077] Therefore, the system can adopt different correction strategies for different capability states.
[0078] Example 5: Long-term growth path maintenance based on the risk of forgetting abilities This embodiment further illustrates how the system handles the abilities that educational talents have developed but have not been used for a long time.
[0079] Once an education professional completes digital teaching evaluation skills training and achieves a score of 0.84, the system records that skill level as met.
[0080] If the system does not detect that the talent is performing real tasks related to teaching evaluation within the next 90 consecutive days, the risk of forgetting will be calculated based on the most recent valid verification time, historical usage frequency, number of consecutive valid performances, and number of scenarios covered.
[0081] For example: the most recent valid verification was 90 days ago; the number of uses in the last 30 days was 0; the number of historical valid performances was 8; and the number of task scenarios covered was 2.
[0082] The system calculates the forgetting risk value based on the above data. When the forgetting risk reaches 0.70, the ability is marked as a high forgetting risk ability.
[0083] Instead of rearranging the complete course, the system inserts low-cost reproducibility tasks into the growth path.
[0084] For example, the following tasks can be generated: Teaching Case Evaluation Task → Evaluation Indicator Selection Task → New Teaching Plan Evaluation Task → Competency Revalidation Task. If the performance on the replication task reaches 0.80 or above, the forgetting risk is reduced to 0.35, and the original competency level is maintained. If the performance on the replication task is below 0.60, a competency enhancement task is further generated, and transfer validation is performed again.
[0085] By using the above methods, the system can avoid focusing solely on "new capabilities" in its growth path, while simultaneously maintaining existing capabilities.
[0086] Example 6: Continuous Iteration Based on Comprehensive Ability Growth Portfolio This embodiment further illustrates the implementation method of the ability growth record module.
[0087] The system establishes a time-series competency growth profile for each education professional. Each growth profile includes at least: talent identifier; competency node identifier; competency status; competency evidence; task type; task scenario; task performance; transfer verification results; risk of forgetting; path version; reason for path adjustment; and competency status update time. For example, a professional might have the following records across three stages: Stage 1: AI tool application competency 0.62; digital instructional design competency 0.48; smart classroom implementation competency 0.40. Stage 2: AI tool application competency 0.78; digital instructional design competency 0.65; smart classroom implementation competency 0.52. Stage 3: AI tool application competency 0.86; digital instructional design competency 0.79; smart classroom implementation competency 0.71.
[0088] The system calculates the capability growth trend based on continuous state changes. For capability nodes that have been completed and meet the transfer requirements, the priority of repetitive training is reduced in the next cycle. For capability degradation nodes, the training priority is increased. For nodes with insufficient cross-scenario transfer capabilities, instead of simply adding knowledge learning tasks, scenario practice tasks are added. For nodes with a high risk of forgetting but whose current capabilities still meet the standards, reproduction tasks are prioritized. This ensures that the growth path for the next cycle is not generated from scratch, but rather planned based on the capability status of the previous cycle. Ultimately, this forms a continuous closed loop: capability evidence collection → capability status update → capability gap calculation → growth path generation → learning and training → practice execution → transfer verification → trajectory deviation analysis → local path correction → capability growth profile update → next cycle path generation.
[0089] The above embodiments show that the core of the present invention is not to simply use artificial intelligence to score educational talents, nor to recommend several courses from the course library based on the scoring results. Instead, it transforms the multi-source behavioral data of educational talents into ability evidence with credibility and time attributes, forms a dynamic ability status in the ability graph, and then determines the ability gap chain in reverse according to the target position, and uses ability dependency relationship to determine the growth sequence.
[0090] Furthermore, after the growth path is executed, this invention does not use the course completion rate as the final basis for ability formation. Instead, it examines whether the talent has truly formed the target ability through a transfer task with different scenario parameters than the training task. When there is a significant deviation between the predicted ability trajectory and the actual ability trajectory, the upstream ability is traced through the ability map to determine the specific ability node that caused the deviation. Instead of replanning the entire growth path, a local path reconstruction is performed for the ability node.
[0091] Therefore, this invention can form a dynamic closed-loop technical mechanism that differs from the existing "knowledge mastery assessment + course recommendation" model. Existing research on personalized learning paths using knowledge graphs mainly focuses on path planning based on knowledge concept mastery, knowledge relationships, and the ranking of learning resources, while knowledge tracing is primarily used to predict knowledge status and subsequent learning performance. This invention further expands the evaluation object to job competency status and establishes a continuous data processing and feedback control chain encompassing the credibility of competency evidence, competency transfer, forgetting risk, deviation between predicted and actual competency trajectories, and local path reconstruction. This allows the evaluation results to directly participate in the dynamic adjustment of subsequent growth paths, thereby improving the authenticity of educational talent competency assessment, the relevance of growth paths, and the adaptability of long-term training processes.
[0092] 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 invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.
Claims
1. An AI-driven educational talent competency assessment and growth path planning system, comprising a multi-source educational talent data collection module, a competency graph construction module, a competency evidence analysis module, a dynamic competency status modeling module, a competency gap calculation module, a growth path generation module, a competency transfer verification module, and a path dynamic correction module, characterized in that: The capability graph construction module establishes a multi-layer capability graph consisting of knowledge nodes, capability nodes, task nodes, job nodes, and learning resource nodes, and establishes prerequisite, dependency, and transfer relationships between capability nodes. The capability evidence analysis module weights each capability piece of evidence based on the reliability of the evidence source, the authenticity of the task, the degree of time decay, and the independence of the evidence, forming a capability evidence set. The dynamic capability state modeling module calculates the current capability value, capability change trend, forgetting risk, and cross-scenario migration value of each capability node based on the capability evidence set, forming a dynamic capability state vector. The capability gap calculation module calculates the capability gap based on the target capability state vector and dynamic capability state vector corresponding to the target position, and forms a capability gap chain by topological sorting the capability gap according to the prerequisites and dependencies. The growth path generation module selects knowledge learning tasks, ability training tasks, practice tasks, and transfer verification tasks from learning resource nodes and task nodes based on the capability gap chain, and generates a growth path that progresses step by step from the current capability state to the target capability state.
2. The AI-driven educational talent competency assessment and growth path planning system according to claim 1, characterized in that: The multi-source education talent data collection module collects knowledge test data, course learning data, homework results, teaching practice data, project tasks data, evaluation feedback data, and target job data of education talents, and adds timestamps, task scenarios, ability tags, evidence sources, and evidence types to each data point. The capability transfer verification module generates a transfer task that corresponds to the target capability and is different from the training task after the learning task is executed, and updates the corresponding capability node based on the performance of the transfer task. The path dynamic correction module compares the deviation between the predicted capability trajectory and the actual capability trajectory during the execution of the growth path. When the deviation exceeds a preset threshold, it locates the corresponding capability node and reconstructs only the capability node and its associated growth sub-paths. The updated dynamic capability status is used as the input for the generation of the growth path in the next cycle.
3. The AI-driven educational talent competency assessment and growth path planning system according to claim 1, characterized in that: The capability graph construction module establishes a mapping relationship according to the hierarchy of "knowledge-capability-task-position" and establishes directional capability dependency edges between capability nodes; Each capability dependency edge carries a dependency strength parameter, which is used to represent the degree of influence of the preceding capability on the formation of the subsequent capability; The growth path generation module sorts the capability gaps according to the direction of the capability dependency edges, and prioritizes the gap nodes that are located upstream of the target capability and have a high dependency strength as the priority training nodes.
4. The AI-driven educational talent competency assessment and growth path planning system according to claim 1, characterized in that: The capability evidence analysis module sets corresponding evidence credibility coefficients for different types of evidence and calculates comprehensive evidence weights based on the reliability of the evidence source, the authenticity of the task, the freshness of time, and the degree of repetition between different types of evidence. When multiple sources of evidence point to the same capability node, repetitive evidence with the same source or highly related content is subject to correlation suppression, and the overall weight of capability evidence from different task scenarios and which are independent of each other is increased, so as to reduce the abnormal impact of a single evaluation result on the capability status.
5. The AI-driven educational talent competency assessment and growth path planning system according to claim 1, characterized in that: The dynamic capability status modeling module establishes a capability status parameter group for each capability node, consisting of the current capability value, capability growth rate, capability stability, forgetting risk value, and scenario migration value. The forgetting risk value is calculated based on the most recent valid verification time of the capability node, historical usage frequency, number of consecutive valid performances, and number of task scenarios covered. When the forgetting risk value exceeds a preset threshold, a reproduction task and a re-verification task for the capability node are inserted into the growth path.
6. The AI-driven educational talent competency assessment and growth path planning system according to claim 1, characterized in that: The capability gap calculation module not only calculates the numerical difference between the target capability value and the current capability value, but also calculates the capability gap priority based on the dependency strength, migration value, and capability importance in the target position of the capability node. It then forms a capability gap chain with sequential constraints according to the capability gap priority and capability dependency relationship, so that subsequent growth tasks can only enter the next capability stage after meeting the qualification conditions of their corresponding preceding capability nodes.
7. The AI-driven educational talent competency assessment and growth path planning system according to claim 1, characterized in that: The growth path generation module configures at least one knowledge learning task, one ability training task, one scenario practice task, and one transfer verification task for each ability gap node. It calculates the path evaluation value based on the expected ability gain, task difficulty, expected learning cost, learning load, and ability transfer benefit corresponding to each task, and selects the growth path whose path evaluation value meets the preset conditions from the candidate paths that meet the ability dependency constraints.
8. The AI-driven educational talent competency assessment and growth path planning system according to claim 1, characterized in that: The transfer task generated by the capability transfer verification module has different task parameters, task scenarios or task data than the corresponding training task. The module calculates the transfer capability value based on the accuracy of the educational personnel in completing the transfer task, the completeness of the problem-solving steps, the rationality of the task resource call and the stability of the result. When the performance of the training task reaches the set threshold but the transfer capability value does not reach the set threshold, the capability level of the corresponding capability node remains unchanged and the capability node is marked as a node with insufficient knowledge acquisition but insufficient capability transfer.
9. The AI-driven educational talent competency assessment and growth path planning system according to claim 1, characterized in that: Before the growth path is executed, the path dynamic correction module generates a predicted capability trajectory based on the current capability status, task sequence, and historical capability change data. During the execution process, it acquires the actual capability status at preset time intervals and forms the actual capability trajectory. It performs capability node-level positioning based on the difference between the predicted capability trajectory and the actual capability trajectory, and judges whether the capability growth is insufficient, the capability growth is too fast, or the capability status is degraded based on the direction of the difference. When it is determined that the capacity growth is insufficient, the path dynamic correction module further traces the upstream capacity node that caused the insufficient growth of the capacity node from the capacity map, and determines whether the current capacity value of the upstream capacity node is lower than the corresponding pre-threshold. When the upstream capability node is lower than the corresponding pre-threshold, the subsequent training tasks in the original growth path will be replaced with the compensation training tasks corresponding to the upstream capability node. When an upstream capability node reaches the pre-threshold, the corresponding task of the upstream capability node remains unchanged, and the practice task or migration verification task of the current capability node is adjusted.
10. The AI-driven educational talent competency assessment and growth path planning system according to claim 1, characterized in that: It also includes a capability growth profile module, which records the capability status, capability evidence, task performance, transfer verification results, forgetting risk and path adjustment records of educational talents at each capability node in a time sequence, and generates capability growth trajectory based on the capability status changes in different time periods. The growth path generation module calls the capability growth trajectory during the next path planning, reduces the retraining priority of capability nodes that have been completed and meet the migration capability standards, and increases the training priority of capability nodes that have degraded or have insufficient cross-scenario migration, thereby forming a closed-loop growth path based on continuous updates of historical capability status.
