Self-adaptive path training system and method based on capability evolution prediction

Abstract

The invention discloses an adaptive path training system and method based on capability evolution prediction, and the system comprises a data collection module which is used for collecting the real-time behavior data of a target training object in a training process; the capability state construction module is used for constructing capability state data according to the real-time behavior data; the capability trajectory prediction module is used for performing capability evolution prediction on the capability state data according to a preset time sequence prediction model to obtain a behavior evolution trajectory sequence; the training path generation module is used for constructing a target function on the basis of a preset training task set according to the behavior evolution trajectory sequence and in combination with training time and training cost, and solving the target function on the basis of a preset strengthening algorithm to obtain an optimal training path; and the training execution module is used for sequentially executing the corresponding training tasks based on the optimal training path. The technical problem that in the prior art, a training path is fixed, and a training object cannot be dynamically adjusted is solved.

Description

Technical Field

[0001] This invention relates to the field of adaptive training optimization technology, and in particular to an adaptive path training system and method based on capability evolution prediction. Background Technology

[0002] With the development of artificial intelligence technology, intelligent training systems are increasingly being applied to fields such as education and training, vocational skills training, and complex task training. Traditional training systems typically employ fixed training procedures or standardized training content, meaning all trainees complete training tasks according to a pre-set training path. While this approach is simple to implement, it struggles to adapt to the differences in ability levels, learning speeds, and knowledge structures among different trainees.

[0003] Currently, existing training systems adjust training content using rule-based or simple recommendation algorithms. For example, they might select the next training task based on test scores or make simple recommendations based on knowledge mastery. However, these methods typically make decisions based solely on the current ability state, lacking the ability to predict the development trend of the trainee's abilities, thus making it difficult to globally optimize the training path. While some existing technologies introduce tracking models for simple evolution, current training systems primarily focus on estimating knowledge mastery levels, paying less attention to the evolutionary trajectory of abilities over time. Furthermore, current training path planning often employs static recommendation strategies, lacking the ability to optimize long-term training effects, resulting in low overall training efficiency. Summary of the Invention

[0004] This invention provides an adaptive path training system and method based on capability evolution prediction to solve the technical problem in the prior art where the training path is fixed and cannot be dynamically adjusted for the training object.

[0005] To address the aforementioned technical problems, this invention provides an adaptive path training system based on capability evolution prediction, comprising: a data acquisition module, a capability state construction module, a capability trajectory prediction module, a training path generation module, and a training execution module. The data acquisition module is used to collect real-time behavioral data of the target training object during the training process; The capability status construction module is used to construct capability status data based on the real-time behavior data; The capability trajectory prediction module is used to predict the capability evolution of the capability state data according to a preset time series prediction model, and obtain a behavior evolution trajectory sequence. The training path generation module is used to construct an objective function based on a preset training task set, according to the behavior evolution trajectory sequence, combined with training time and training cost, and solve the objective function based on a preset reinforcement algorithm to obtain the optimal training path; wherein, the preset training task set includes several training tasks, and each training task corresponds to a unique training time and training cost. The training execution module is used to execute the corresponding training tasks sequentially based on the optimal training path, so that the target training object can be trained based on the training tasks.

[0006] As a preferred embodiment, the collection of real-time behavioral data of the target training object during the training process specifically includes: Collect training behavior records of the target training object at each stage of the training process; the training behavior records include: training task completion time, task score, task completion degree, error rate, and task preset level; By integrating the records of each training behavior, real-time behavioral data can be obtained.

[0007] As a preferred embodiment, the step of constructing capability status data based on the real-time behavior data specifically includes: Based on the training task completion time, task score, task completion degree, error rate and task preset level in each training behavior record, and combined with the corresponding preset dimension weights, the capability state vector corresponding to each training behavior record is calculated. Capability state data is constructed based on the capability state vector corresponding to each training session.

[0008] As a preferred embodiment, the step of predicting the capability evolution of the capability state data based on a preset time series prediction model to obtain a behavioral evolution trajectory sequence specifically includes: The capability status data is input into a preset time series prediction model to predict capability evolution, thereby obtaining a sequence of behavioral evolution trajectories. The preset time series prediction model is trained using historical capability status data samples.

[0009] As a preferred embodiment, the step of constructing an objective function based on a preset training task set, according to the behavioral evolution trajectory sequence, combined with training time and training cost, and solving the objective function based on a preset reinforcement algorithm to obtain the optimal training path specifically includes: The training tasks in the preset training task set are constructed into several training combination sequences; wherein, each training combination sequence corresponds to a unique training time, training cost and ability improvement, the ability improvement is calculated based on the behavior evolution trajectory sequence, and the training cost is calculated based on the training time and the resource consumption during the training process; Using the training combination sequence as a candidate solution, and based on the training time, training cost, and capability improvement, and combining the weights corresponding to the training time, training cost, and capability improvement respectively, an objective function is constructed. The objective function is solved based on a preset reinforcement algorithm to obtain the optimal value of the objective function. Based on the training time, training cost and capability improvement corresponding to the optimal value of the objective function, the corresponding training combination sequence is obtained as the optimal training path.

[0010] As a preferred embodiment, the step of sequentially executing the corresponding training tasks based on the optimal training path to enable the target training object to be trained based on the training tasks further includes: Based on the optimal training path, the corresponding training tasks are output and executed sequentially. During the training process of the target training object based on the training task, behavioral data of the target training object is continuously collected.

[0011] As a preferred option, this system also includes: a training path optimization module; The training path optimization module is used to calculate and update capability status data based on the continuously collected behavioral data of the target training object, predict and update the behavioral evolution trajectory sequence based on the updated capability status data, and update the optimal training path based on the updated behavioral evolution trajectory sequence.

[0012] Accordingly, the present invention also provides an adaptive path training method based on capability evolution prediction, comprising: Collect real-time behavioral data of the target training object during the training process; Capability status data is constructed based on the real-time behavioral data; Based on a preset time series prediction model, the capability state data is used to predict capability evolution to obtain a behavioral evolution trajectory sequence. Based on a preset training task set, an objective function is constructed according to the behavioral evolution trajectory sequence, combined with training time and training cost. The objective function is then solved based on a preset reinforcement algorithm to obtain the optimal training path. The preset training task set includes several training tasks, each with a unique training time and training cost. The corresponding training tasks are executed sequentially based on the optimal training path, so that the target training object is trained based on the training tasks.

[0013] As a preferred embodiment, the step of constructing an objective function based on a preset training task set, according to the behavioral evolution trajectory sequence, combined with training time and training cost, and solving the objective function based on a preset reinforcement algorithm to obtain the optimal training path specifically includes: The training tasks in the preset training task set are constructed into several training combination sequences; wherein, each training combination sequence corresponds to a unique training time, training cost and ability improvement, the ability improvement is calculated based on the behavior evolution trajectory sequence, and the training cost is calculated based on the training time and the resource consumption during the training process; Using the training combination sequence as a candidate solution, and based on the training time, training cost, and capability improvement, and combining the weights corresponding to the training time, training cost, and capability improvement respectively, an objective function is constructed. The objective function is solved based on a preset reinforcement algorithm to obtain the optimal value of the objective function. Based on the training time, training cost and capability improvement corresponding to the optimal value of the objective function, the corresponding training combination sequence is obtained as the optimal training path.

[0014] As a preferred option, it also includes: During the training process of the target training object based on the training task, behavioral data of the target training object is continuously collected; Based on the continuously collected behavioral data of the target training object, the capability status data is calculated and updated. Based on the updated capability status data, the behavioral evolution trajectory sequence is predicted and updated. Based on the updated behavioral evolution trajectory sequence, the optimal training path is updated.

[0015] Compared with the prior art, the embodiments of the present invention have the following beneficial effects: The technical solution of this invention collects real-time behavioral data of the target training object during the training process to construct corresponding capability state data, predicts capability evolution over a period of time, obtains a behavioral evolution trajectory sequence, and then constructs a corresponding objective function based on a preset training task set to solve for the optimal value, thus obtaining the final optimal training path. This allows for the pre-planning of the optimal training path, avoiding the problem of training content being too difficult or too easy due to a fixed training path. The target training object is then trained sequentially based on the training tasks corresponding to the optimal training path. At the same time, by comprehensively considering capability evolution, training efficiency, and training cost, the global optimization of the training strategy is achieved, which can significantly improve the intelligence level and training effect of the training system, reduce repetitive and ineffective training, and generate personalized training paths according to the capability level of different training objects. It can be applied to various complex intelligent training scenarios such as education and training, skills training, and simulation training. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the structure of an adaptive path training system based on capability evolution prediction provided in an embodiment of the present invention; Figure 2 This is a flowchart of an adaptive path training method based on capability evolution prediction provided in an embodiment of the present invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] Example 1 Please refer to Figure 1 An adaptive path training system based on capability evolution prediction is provided in this embodiment of the invention, comprising: a data acquisition module 101, a capability state construction module 102, a capability trajectory prediction module 103, a training path generation module 104, and a training execution module 105.

[0019] The data acquisition module 101 is used to collect real-time behavioral data of the target training object during the training process.

[0020] As a preferred embodiment, the collection of real-time behavioral data of the target training object during the training process specifically includes: Collect training behavior records of the target training object at each stage of the training process; the training behavior records include: training task completion time, task score, task completion degree, error rate, and task preset level; By integrating the records of each training behavior, real-time behavioral data can be obtained.

[0021] In this embodiment, refined data collection is performed on the target training object to fully acquire multi-dimensional information such as task completion time, task score, task completion degree, error rate, and task preset level corresponding to each training behavior during the training process. Then, the single behavior records are uniformly integrated and processed to form real-time behavior data that can be used for subsequent analysis, thereby realizing the full-dimensional and fine-grained data quantification and collection of the training process.

[0022] In this embodiment, the behavioral data of the target training object during the training process includes, but is not limited to: training task completion status, training score or rating, task completion time, error type and error frequency, and training task difficulty level. Furthermore, the collected real-time behavioral data set is: D = {d1, d2, … , di}; where di represents the i-th training behavior record; each training data point di is defined as: di = (ti, si, ci, ei, li); where ti represents the completion time of the current training task; si represents the task rating (ranging from 0 to 100), which can be obtained based on the evaluation rating of the training object when performing the training task; ci represents the task completion degree (ranging from 0 to 1, where 0 indicates the task is not completed and 1 indicates the task is completed); ei represents the error rate, i.e., the number of errors that occurred during the execution of the corresponding training task / the total number of operations; li represents the task preset level (ranging from 1 to 5), obtained by presetting the level of the training task, with each training task corresponding to a task preset level.

[0023] Understandably, by collecting key indicators of training behavior one by one and integrating them into real-time behavioral data, it is possible to accurately and objectively reflect the real-time training status and performance of the trainees, avoid the evaluation bias caused by single indicators or scattered data, provide a real and reliable data foundation for subsequent capability modeling and trajectory prediction, and effectively improve the accuracy and pertinence of subsequent training analysis and path optimization.

[0024] The capability status construction module 102 is used to construct capability status data based on the real-time behavior data.

[0025] As a preferred embodiment, the step of constructing capability status data based on the real-time behavioral data specifically includes: Based on the training task completion time, task score, task completion degree, error rate and task preset level in each training behavior record, and combined with the corresponding preset dimension weights, the capability state vector corresponding to each training behavior record is calculated. Capability state data is constructed based on the capability state vector corresponding to each training session.

[0026] In this embodiment, based on multi-dimensional indicators such as task completion time, task score, task completion degree, error rate and task preset level in the single training behavior record of the training object, and combined with the pre-set dimension weights of each indicator, a weighted calculation is performed to calculate the capability state vector corresponding to the single training behavior one by one. Using this vector as the basic unit, complete capability state data is summarized to realize the vectorized and standardized quantitative modeling of training performance.

[0027] In this embodiment, the ability state of the training object is modeled based on the training data, that is, a computational model for the corresponding ability state data is constructed. For example, an ability state vector Ct = (c1, c2, … , ck) can be defined; where: Ct represents the ability state of the target training object at time t; ck represents the k-th ability dimension; wherein, the ability dimension includes knowledge mastery, skill proficiency, task completion efficiency, etc. The computational function model for the ability state vector is: Ct = f(Dt) = f(k_t, s_t, e_t) Where: Dt is the set of behavioral data at time t; k_t = mean(si) / 100, representing the knowledge mastery; s_t = mean(ci), representing the skill proficiency; e_t = 1 / mean(ti), representing the task execution efficiency; mean() means to calculate the average value.

[0028] Furthermore, more capability dimensions can be set, such as determining the error rate by calculating the corresponding error rate ei, and determining the corresponding problem-solving capability by calculating the corresponding task difficulty level li.

[0029] Understandably, by introducing preset dimensional weights to perform differentiated calculations on multiple training indicators, it is possible to more accurately characterize the ability level in a way that better fits the actual training scenario, avoiding the evaluation distortion caused by simply averaging the indicators. At the same time, constructing ability state data with a unified ability state vector can provide structured and computable data support for subsequent ability evolution analysis and training path optimization, thereby improving the scientific nature of ability assessment and the efficiency of subsequent processing.

[0030] The capability trajectory prediction module 103 is used to predict the capability evolution of the capability state data according to a preset time series prediction model, and obtain a behavior evolution trajectory sequence.

[0031] As a preferred embodiment, the step of predicting the capability evolution of the capability state data based on a preset time series prediction model to obtain a behavioral evolution trajectory sequence specifically includes: The capability status data is input into a preset time series prediction model to predict capability evolution, thereby obtaining a sequence of behavioral evolution trajectories. The preset time series prediction model is trained using historical capability status data samples.

[0032] In this embodiment, the constructed capability status data is used as time series input and imported into a pre-constructed time series prediction model trained with historical capability status data samples. The model extracts and extrapolates the temporal features of continuous capability status, thereby predicting the future capability change trend of the training object and ultimately forming a behavioral evolution trajectory sequence that represents the dynamic change law of capability.

[0033] In this embodiment, after obtaining the capability status sequence, a time series prediction model is constructed to predict future capability change trends. The capability status data sequence is represented as follows: C = {C1, C2, …, Ct} Using time series forecasting model F, predict future capability status: Ct+1 = F(C1, C2, … , Ct) Ultimately, a behavioral evolution trajectory sequence T is obtained that can characterize the dynamic changes in ability. C = {Ct+1, Ct+2, …, Ct+m}.

[0034] In this embodiment, the time series prediction model can be a Long Short-Term Memory Network (LSTM), a Transformer time series model, or a time series regression model. It takes the ability state vector of the training object at consecutive time moments as input, learns the inherent laws and trend characteristics of the historical ability state changes over time, and successively predicts the ability state at subsequent time moments, thereby outputting prediction results for multiple consecutive steps, and finally forming a complete ability evolution trajectory sequence.

[0035] Understandably, using a mature time series prediction model trained on historical samples for evolutionary prediction can fully explore the potential patterns of ability state changes over time, and improve the reliability and stability of trajectory prediction. The generated behavioral evolution trajectory sequence can intuitively reflect the trend of ability development, providing a quantitative basis for the dynamic adjustment and optimization of subsequent training paths, and enhancing the foresight and adaptability of training strategies.

[0036] The training path generation module 104 is used to construct an objective function based on a preset training task set, according to the behavior evolution trajectory sequence, combined with training time and training cost, and solve the objective function based on a preset reinforcement algorithm to obtain the optimal training path; wherein, the preset training task set includes several training tasks, and each training task corresponds to a unique training time and training cost.

[0037] In a preferred embodiment, the step of constructing an objective function based on a preset training task set, according to the behavioral evolution trajectory sequence, combined with training time and training cost, and solving the objective function based on a preset reinforcement algorithm to obtain the optimal training path specifically includes: The training tasks in the preset training task set are constructed into several training combination sequences; wherein, each training combination sequence corresponds to a unique training time, training cost and ability improvement, the ability improvement is calculated based on the behavior evolution trajectory sequence, and the training cost is calculated based on the training time and the resource consumption during the training process; Using the training combination sequence as a candidate solution, and based on the training time, training cost, and capability improvement, and combining the weights corresponding to the training time, training cost, and capability improvement respectively, an objective function is constructed. The objective function is solved based on a preset reinforcement algorithm to obtain the optimal value of the objective function. Based on the training time, training cost and capability improvement corresponding to the optimal value of the objective function, the corresponding training combination sequence is obtained as the optimal training path.

[0038] In this embodiment, the most suitable training task sequence is selected from a preset training task set based on the behavioral evolution trajectory sequence. For example, the preset training task set can be set as: S = {s1, s2, … , sn}. The training path is selected according to the following objective function: Maximize: R = α·ΔC + β·(ΔC / T) - γ·Cost Where ΔC = ||Ct+1 - Ct|| represents the ability improvement; T represents the training time; Cost represents the training cost; α, β, and γ are the first weight coefficients, which can be set according to the actual situation. Preferably, α + β + γ = 1, and dynamically adjusted according to different training stages. For example, in the early stage of training, the weight of α is increased to enhance ability improvement; in the middle stage of training, the weight of β is increased to optimize training efficiency; and in the later stage of training, the weight of γ is increased to control training cost. The training cost is defined as: Cost = λ1·T + λ2·Resource Where: Resource represents the computational or system resource consumption during the training process; λ1 and λ2 are the second weight coefficients, which can be set according to the actual situation.

[0039] In this embodiment, the optimal training path is solved using a reinforcement learning algorithm. Preset reinforcement learning algorithms include, but are not limited to, Q-learning or Deep Q-Network (DQN) to obtain the long-term optimal training strategy. Finally, the corresponding optimal training path P is determined. The preset training task set S is the total set of all selectable training tasks. The optimal training path P is an ordered subset formed by selecting from the preset training task set S and sorting it according to the optimal training logic, i.e., P⊆S. All training tasks pi in the optimal training path P originate from the preset training task set S. The process of solving the objective function is directly related to the preset training task set S. By binding the training tasks with the objective function parameters, the maximization of the R-value is precisely correlated with the task sequences in the preset training task set S. Specifically, each training task in the preset training task set S and all possible task combinations (ordered sequences) are taken as candidate solutions. Each candidate solution (i.e., a set of task sequences) corresponds to a unique training time T, training cost Cost, and a unique capability improvement ΔC.

[0040] In this embodiment, all training tasks in the preset training task set are combined and arranged to construct various training combination sequences with different arrangements, serving as all candidate objects for subsequent path optimization. Each training combination sequence is independently matched with its own training time and training resource consumption, thereby determining the unique corresponding training time and training cost parameters. At the same time, based on the behavioral evolution trajectory sequence output above, the capability improvement that each sequence can achieve is accurately calculated, realizing the independent binding and one-to-one correspondence of multi-dimensional quantitative indicators for each candidate sequence.

[0041] In this embodiment, after quantifying the indicators of each training combination sequence, fixed weight coefficients are introduced to adapt to the improvement in capability, training time, and training cost, respectively. A standardized optimization objective function is constructed by combining the three core quantitative indicators. By configuring the weights, the optimization priorities of different indicators are distinguished, and the three constraints of capability growth, training efficiency, and resource consumption are integrated into a unified calculation model. This achieves a quantitative expression of multi-objective optimization requirements and provides a unified calculation and evaluation standard for subsequent optimal solution solving.

[0042] In this embodiment, all training combination sequences are used as the candidate solution set for optimization. A pre-configured reinforcement learning algorithm is used to iteratively calculate and globally solve the objective function. The calculation results of the objective function corresponding to different candidate sequences are continuously compared to select the optimal value of the objective function under extreme conditions. During the solution process, the reinforcement algorithm can autonomously traverse the arrangement logic of various task combinations, taking into account both short-term training gains and long-term training adaptability, ensuring the comprehensiveness and rationality of the solution results.

[0043] Understandably, integrating fragmented training tasks into standardized sequences and quantifying them across multiple dimensions establishes a clear correlation between task combinations and capability gains, time costs, and resource consumption, thus resolving the issues of subjective and unquantifiable selection in traditional training paths. Furthermore, by leveraging weight adaptation and objective function construction, it is possible to flexibly adapt to the optimization focus of different training stages, making training plans more aligned with the capability evolution patterns of the training subjects.

[0044] Furthermore, by utilizing reinforcement learning algorithms to intelligently solve the objective function, the optimal training path can be quickly identified from a massive sequence of task combinations, significantly improving the efficiency and accuracy of path planning. The overall solution forms a complete closed loop from task combination, metric calculation, function construction to intelligent optimization, effectively balancing training effect, training time, and resource consumption, achieving adaptive optimization of the training path, and greatly improving the scientific rigor and adaptability of personalized training.

[0045] The training execution module 105 is used to execute the corresponding training tasks sequentially based on the optimal training path, so that the target training object is trained based on the training tasks.

[0046] As a preferred embodiment, the step of sequentially executing corresponding training tasks based on the optimal training path to enable the target training object to be trained based on the training tasks further includes: During the training process of the target training object based on the training task, behavioral data of the target training object is continuously collected.

[0047] As a preferred embodiment, the system further includes: a training path optimization module; The training path optimization module is used to calculate and update capability status data based on the continuously collected behavioral data of the target training object, predict and update the behavioral evolution trajectory sequence based on the updated capability status data, and update the optimal training path based on the updated behavioral evolution trajectory sequence.

[0048] In this embodiment, throughout the entire process of the target training object performing predetermined training tasks and carrying out routine training, the system maintains a real-time monitoring and dynamic data acquisition mechanism, continuously acquiring various behavioral data generated during the training process to achieve continuous and full-cycle input of training data. Based on the real-time newly added behavioral data, the system continuously iteratively calculates and refreshes the training object's ability status data, replacing and correcting historical ability assessment results to ensure that the ability status data is synchronized with the training object's current actual level in real time.

[0049] In this embodiment, the updated capability status data is used as input to re-drive the time series prediction model to complete the extrapolation calculation. The original behavioral evolution trajectory sequence is iteratively updated, correcting the prediction results of future capability change trends, ensuring that the trajectory evolution pattern aligns with the training performance of the target object at each stage. Combining the updated behavioral evolution trajectory sequence, the indicators of each training task combination sequence are recalculated, and the objective function is solved and candidate sequence selection is completed again. This allows for dynamic adjustment and iterative optimization of the generated optimal training path. For example, during training, the system continuously collects new training data and updates the capability status model and capability evolution trajectory prediction model. When new training data arrives, the capability status is updated, the capability trajectory is re-predicted, and the training path is dynamically adjusted, thus forming a closed-loop optimization mechanism.

[0050] In this embodiment, the closed-loop dynamic update mechanism effectively breaks through the limitations of static training paths, promptly capturing the dynamic fluctuations and growth changes in the training subjects' abilities, and avoiding the disconnect between fixed training schemes and actual ability status. Through the cyclical linkage of data collection, ability updates, trajectory retesting, and path optimization, the training strategy can be adaptively adjusted in real time, continuously balancing training gains, time efficiency, and resource consumption, and significantly improving the personalization, adaptability, and dynamic control capabilities of the training scheme.

[0051] It is understandable that by introducing a capability evolution trajectory prediction mechanism, the training system can predict the future capability development trend of the training object, rather than relying solely on the current capability state for decision-making. By constructing a capability state sequence and using a time series prediction model to model capability changes, the capability development trajectory of the training object in multiple future stages can be obtained. Based on this prediction result, the system can plan the optimal training path in advance, thereby avoiding the problem of training content being too difficult or too easy. Simultaneously, a training path optimization objective function is introduced, achieving global optimization of the training strategy by comprehensively considering capability improvement, training efficiency, and training cost. Therefore, the embodiments of this invention can significantly improve the intelligence level and training effect of the training system.

[0052] Implementing the above embodiments has the following effects: The technical solution of this invention collects real-time behavioral data of the target training object during the training process to construct corresponding capability state data, predicts capability evolution over a period of time, obtains a behavioral evolution trajectory sequence, and then constructs a corresponding objective function based on a preset training task set to solve for the optimal value, thus obtaining the final optimal training path. This allows for the pre-planning of the optimal training path, avoiding the problem of training content being too difficult or too easy due to a fixed training path. The target training object is then trained sequentially based on the training tasks corresponding to the optimal training path. At the same time, by comprehensively considering capability evolution, training efficiency, and training cost, the global optimization of the training strategy is achieved, which can significantly improve the intelligence level and training effect of the training system, reduce repetitive and ineffective training, and generate personalized training paths according to the capability level of different training objects. It can be applied to various complex intelligent training scenarios such as education and training, skills training, and simulation training.

[0053] Example 2 This invention is applied to job competency training scenarios within intelligent recruitment systems. Taking the AI ​​algorithm engineer position as an example, the competency dimensions include: programming skills, algorithm understanding skills, and engineering practice skills.

[0054] In a certain intelligent training system, the historical training data of the target training object is first collected, including information such as training task completion status, task score, and task completion time.

[0055] The system constructs a capability state vector based on this data: Ct = (Knowledge mastery, skill proficiency, task completion efficiency) The system then constructs a time series of historical capability states: C = {C1, C2, …, Ct} The LSTM model is used to train the capability sequence and predict future capability states: Ct+1 = F(C1, C2, … , Ct) Generate future capability trajectories based on the prediction results: T C = {Ct+1, Ct+2, Ct+3} The system then selects a training task sequence suitable for the ability trajectory from the training task library. For example, the preset training task set contains tasks of different difficulty levels: S = {basic tasks, intermediate tasks, advanced tasks}. The system optimizes the objective function as follows: R = α·ΔC + β·(ΔC / T) - γ·Cost Calculate the benefit values ​​of different training paths and select the training path with the highest benefit.

[0056] During the training process, the system continuously collects new training data and updates the capability model, thereby constantly adjusting the training path and achieving dynamic optimization. In this embodiment, each variable has a clear physical meaning and can be obtained through actual collection and calculation by the computer system, thus ensuring the feasibility of this invention.

[0057] In this embodiment, by modeling capability states, personalized training paths can be generated based on the capability levels of different training subjects, achieving personalized training path planning. An capability evolution trajectory prediction model is introduced, enabling the system to predict future capability changes in training subjects, thus possessing capability development trend prediction capabilities. By optimizing training paths, redundant and ineffective training is reduced, improving training efficiency. Simultaneously, the training system can continuously adjust training strategies based on real-time training data, exhibiting dynamic optimization capabilities. Furthermore, this invention can be applied to various intelligent training scenarios such as education and training, skills training, and simulation training, demonstrating a wide range of applicability.

[0058] Example 3 Please see Figure 2 This invention provides an adaptive path training method based on capability evolution prediction, comprising the following steps S201-S205: S201: Collect real-time behavioral data of the target training object during the training process.

[0059] S202: Construct capability status data based on the real-time behavior data.

[0060] S203: Based on a preset time series prediction model, the capability state data is used to predict capability evolution to obtain a behavior evolution trajectory sequence.

[0061] S204: Based on a preset training task set, according to the behavior evolution trajectory sequence, combined with training time and training cost, construct an objective function, and solve the objective function based on a preset reinforcement algorithm to obtain the optimal training path; wherein, the preset training task set includes several training tasks, and each training task corresponds to a unique training time and training cost.

[0062] S205: Execute the corresponding training tasks sequentially based on the optimal training path, so that the target training object is trained based on the training tasks.

[0063] As a preferred embodiment, the collection of real-time behavioral data of the target training object during the training process specifically includes: Collect training behavior records of the target training object at each stage of the training process; the training behavior records include: training task completion time, task score, task completion degree, error rate, and task preset level; By integrating the records of each training behavior, real-time behavioral data can be obtained.

[0064] As a preferred embodiment, the step of constructing capability status data based on the real-time behavior data specifically includes: Based on the training task completion time, task score, task completion degree, error rate and task preset level in each training behavior record, and combined with the corresponding preset dimension weights, the capability state vector corresponding to each training behavior record is calculated. Capability state data is constructed based on the capability state vector corresponding to each training session.

[0065] As a preferred embodiment, the step of predicting the capability evolution of the capability state data based on a preset time series prediction model to obtain a behavioral evolution trajectory sequence specifically includes: The capability status data is input into a preset time series prediction model to predict capability evolution, thereby obtaining a sequence of behavioral evolution trajectories. The preset time series prediction model is trained using historical capability status data samples.

[0066] As a preferred embodiment, the step of constructing an objective function based on a preset training task set, according to the behavioral evolution trajectory sequence, combined with training time and training cost, and solving the objective function based on a preset reinforcement algorithm to obtain the optimal training path specifically includes: The training tasks in the preset training task set are constructed into several training combination sequences; wherein, each training combination sequence corresponds to a unique training time, training cost and ability improvement, the ability improvement is calculated based on the behavior evolution trajectory sequence, and the training cost is calculated based on the training time and the resource consumption during the training process; Using the training combination sequence as a candidate solution, and based on the training time, training cost, and capability improvement, and combining the weights corresponding to the training time, training cost, and capability improvement respectively, an objective function is constructed. The objective function is solved based on a preset reinforcement algorithm to obtain the optimal value of the objective function. Based on the training time, training cost and capability improvement corresponding to the optimal value of the objective function, the corresponding training combination sequence is obtained as the optimal training path.

[0067] As a preferred embodiment, the step of sequentially executing the corresponding training tasks based on the optimal training path to enable the target training object to be trained based on the training tasks further includes: Based on the optimal training path, the corresponding training tasks are output and executed sequentially. During the training process of the target training object based on the training task, behavioral data of the target training object is continuously collected.

[0068] As a preferred embodiment, this method further includes: Based on the continuously collected behavioral data of the target training object, the capability status data is calculated and updated. Based on the updated capability status data, the behavioral evolution trajectory sequence is predicted and updated. Based on the updated behavioral evolution trajectory sequence, the optimal training path is updated.

[0069] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working process of the method described above can be referred to the corresponding principles and processes in the aforementioned system embodiments, and will not be repeated here.

[0070] Implementing the above embodiments has the following effects: The technical solution of this invention collects real-time behavioral data of the target training object during the training process to construct corresponding capability state data, predicts capability evolution over a period of time, obtains a behavioral evolution trajectory sequence, and then constructs a corresponding objective function based on a preset training task set to solve for the optimal value, thus obtaining the final optimal training path. This allows for the pre-planning of the optimal training path, avoiding the problem of training content being too difficult or too easy due to a fixed training path. The target training object is then trained sequentially based on the training tasks corresponding to the optimal training path. At the same time, by comprehensively considering capability evolution, training efficiency, and training cost, the global optimization of the training strategy is achieved, which can significantly improve the intelligence level and training effect of the training system, reduce repetitive and ineffective training, and generate personalized training paths according to the capability level of different training objects. It can be applied to various complex intelligent training scenarios such as education and training, skills training, and simulation training.

[0071] Example 4 Accordingly, the present invention also provides a terminal device, comprising: a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor executes the computer program to implement the adaptive path training method based on capability evolution prediction as described in any of the above embodiments.

[0072] The terminal device in this embodiment includes a processor, a memory, and a computer program and computer instructions stored in the memory and executable on the processor. When the processor executes the computer program, it implements the various steps described in Embodiment 1 above, for example... Figure 2 The steps S201 to S205 are shown. Alternatively, when the processor executes the computer program, it implements the functions of each module / unit in the above-described device embodiment, such as the training path generation module.

[0073] For example, the computer program can be divided into one or more modules / units, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules / units can be a series of computer program instruction segments capable of performing specific functions, which describe the execution process of the computer program in the terminal device. For example, the training path generation module is used to construct an objective function based on a preset training task set, according to the behavioral evolution trajectory sequence, combined with training time and training cost, and solve the objective function based on a preset reinforcement algorithm to obtain the optimal training path.

[0074] The terminal device may be a desktop computer, laptop, handheld computer, or cloud server, etc. The terminal device may include, but is not limited to, a processor and memory. Those skilled in the art will understand that the schematic diagram is merely an example of a terminal device and does not constitute a limitation on the terminal device. It may include more or fewer components than illustrated, or combine certain components, or different components. For example, the terminal device may also include input / output devices, network access devices, buses, etc.

[0075] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the terminal device, connecting all parts of the terminal device via various interfaces and lines.

[0076] The memory can be used to store the computer programs and / or modules. The processor implements various functions of the terminal device by running or executing the computer programs and / or modules stored in the memory and by calling data stored in the memory. The memory may mainly include a program storage area and a data storage area. The program storage area may store the operating system, applications required for at least one function, etc.; the data storage area may store data created based on the use of the mobile terminal, etc. In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as hard disk, RAM, plug-in hard disk, smart media card (SMC), secure digital card (SD card), flash card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0077] Wherein, if the modules / units integrated in the terminal device are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when the computer program is executed by a processor, it can implement the steps of the various method embodiments described above. Wherein, the computer program includes computer program code, which can be in the form of source code, object code, executable file, or some intermediate form, etc. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, portable hard drive, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content contained in the computer-readable medium can be appropriately added or removed according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, the computer-readable medium does not include electrical carrier signals and telecommunication signals.

[0078] Example 5 Accordingly, the present invention also provides a computer-readable storage medium comprising a stored computer program, wherein, when the computer program is executed, it controls the device where the computer-readable storage medium is located to perform the adaptive path training method based on capability evolution prediction as described in any of the above embodiments.

[0079] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. In particular, it should be noted that any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention for those skilled in the art.

Claims

1. An adaptive path training system based on capability evolution prediction, characterized in that, include: The module includes a data acquisition module, a capability status construction module, a capability trajectory prediction module, a training path generation module, and a training execution module. The data acquisition module is used to collect real-time behavioral data of the target training object during the training process; The capability status construction module is used to construct capability status data based on the real-time behavior data; The capability trajectory prediction module is used to predict the capability evolution of the capability state data according to a preset time series prediction model, and obtain a behavior evolution trajectory sequence. The training path generation module is used to construct an objective function based on a preset training task set, according to the behavior evolution trajectory sequence, combined with training time and training cost, and solve the objective function based on a preset reinforcement algorithm to obtain the optimal training path; wherein, the preset training task set includes several training tasks, and each training task corresponds to a unique training time and training cost. The training execution module is used to execute the corresponding training tasks sequentially based on the optimal training path, so that the target training object can be trained based on the training tasks.

2. The adaptive path training system based on capability evolution prediction as described in claim 1, characterized in that, The collection of real-time behavioral data of the target training object during the training process specifically includes: Collect training behavior records of the target training object at each stage of the training process; the training behavior records include: training task completion time, task score, task completion degree, error rate, and task preset level; By integrating the records of each training behavior, real-time behavioral data can be obtained.

3. The adaptive path training system based on capability evolution prediction as described in claim 2, characterized in that, The construction of capability status data based on the real-time behavior data specifically includes: Based on the training task completion time, task score, task completion degree, error rate and task preset level in each training behavior record, and combined with the corresponding preset dimension weights, the capability state vector corresponding to each training behavior record is calculated. Capability state data is constructed based on the capability state vector corresponding to each training session.

4. The adaptive path training system based on capability evolution prediction as described in claim 1, characterized in that, The step of predicting the capability evolution of the capability state data based on a preset time series prediction model to obtain a behavioral evolution trajectory sequence specifically includes: The capability status data is input into a preset time series prediction model to predict capability evolution, thereby obtaining a sequence of behavioral evolution trajectories. The preset time series prediction model is trained using historical capability status data samples.

5. The adaptive path training system based on capability evolution prediction as described in claim 4, characterized in that, The process of constructing an objective function based on a preset training task set, according to the behavioral evolution trajectory sequence, combined with training time and training cost, and solving the objective function based on a preset reinforcement algorithm to obtain the optimal training path specifically includes: The training tasks in the preset training task set are constructed into several training combination sequences; wherein, each training combination sequence corresponds to a unique training time, training cost and ability improvement, the ability improvement is calculated based on the behavior evolution trajectory sequence, and the training cost is calculated based on the training time and the resource consumption during the training process; Using the training combination sequence as a candidate solution, and based on the training time, training cost, and capability improvement, and combining the weights corresponding to the training time, training cost, and capability improvement respectively, an objective function is constructed. The objective function is solved based on a preset reinforcement algorithm to obtain the optimal value of the objective function. Based on the training time, training cost and capability improvement corresponding to the optimal value of the objective function, the corresponding training combination sequence is obtained as the optimal training path.

6. An adaptive path training system based on capability evolution prediction as described in any one of claims 1-5, characterized in that, The step of sequentially executing the corresponding training tasks based on the optimal training path to enable the target training object to be trained based on the training tasks further includes: During the training process of the target training object based on the training task, behavioral data of the target training object is continuously collected.

7. The adaptive path training system based on capability evolution prediction as described in claim 6, characterized in that, Also includes: Training path optimization module; The training path optimization module is used to calculate and update capability status data based on the continuously collected behavioral data of the target training object, predict and update the behavioral evolution trajectory sequence based on the updated capability status data, and update the optimal training path based on the updated behavioral evolution trajectory sequence.

8. An adaptive path training method based on capability evolution prediction, characterized in that, include: Collect real-time behavioral data of the target training object during the training process; Capability status data is constructed based on the real-time behavioral data; Based on a preset time series prediction model, the capability state data is used to predict capability evolution to obtain a behavioral evolution trajectory sequence. Based on a preset training task set, an objective function is constructed according to the behavioral evolution trajectory sequence, combined with training time and training cost. The objective function is then solved based on a preset reinforcement algorithm to obtain the optimal training path. The preset training task set includes several training tasks, each with a unique training time and training cost. The corresponding training tasks are executed sequentially based on the optimal training path, so that the target training object is trained based on the training tasks.

9. The adaptive path training method based on capability evolution prediction as described in claim 8, characterized in that, The process of constructing an objective function based on a preset training task set, according to the behavioral evolution trajectory sequence, combined with training time and training cost, and solving the objective function based on a preset reinforcement algorithm to obtain the optimal training path specifically includes: The training tasks in the preset training task set are constructed into several training combination sequences; wherein, each training combination sequence corresponds to a unique training time, training cost and ability improvement, the ability improvement is calculated based on the behavior evolution trajectory sequence, and the training cost is calculated based on the training time and the resource consumption during the training process; Using the training combination sequence as a candidate solution, and based on the training time, training cost, and capability improvement, and combining the weights corresponding to the training time, training cost, and capability improvement respectively, an objective function is constructed. The objective function is solved based on a preset reinforcement algorithm to obtain the optimal value of the objective function. Based on the training time, training cost and capability improvement corresponding to the optimal value of the objective function, the corresponding training combination sequence is obtained as the optimal training path.

10. The adaptive path training method based on capability evolution prediction as described in claim 8, characterized in that, Also includes: During the training process of the target training object based on the training task, behavioral data of the target training object is continuously collected; Based on the continuously collected behavioral data of the target training object, the capability status data is calculated and updated. Based on the updated capability status data, the behavioral evolution trajectory sequence is predicted and updated. Based on the updated behavioral evolution trajectory sequence, the optimal training path is updated.

Images