Laboratory personnel training closed-loop management method based on standard specification
By constructing an initial knowledge graph and using improved association analysis algorithms, enhanced ant colony optimization algorithms, and reinforcement learning feedback mechanisms, the problems of fixed paths and low intelligence in laboratory training were solved, enabling personalized, dynamically adaptive training management and record traceability.
Patent Information
- Application Number
- CN202511736641.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-11-25
- Publication Date
- 2026-02-24
AI Technical Summary
Existing technologies for training laboratory personnel suffer from a lack of systematic and personalized training systems, a lack of dynamic adaptability in the training process, difficulty in generating personalized learning paths, and challenges in evaluating, recording, and tracing training effectiveness.
An initial knowledge graph is constructed, historical training data is mined using an improved association analysis algorithm, a personnel diagnostic matrix is generated through pre-job testing, personalized learning paths are planned using an enhanced ant colony optimization algorithm, and the path planning is optimized through a reinforcement learning feedback mechanism.
It enables systematic and intelligent management of laboratory personnel training, dynamically adapts to changes in personnel capabilities, generates efficient and personalized learning paths, and provides traceable training records.
Smart Images
Figure CN121563101A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the technical field of laboratory personnel, specifically a closed-loop management method for laboratory personnel training based on standards and specifications. Background Technology
[0002] As technical institutions providing testing, inspection, and calibration services, laboratories are a crucial component of the National Quality Infrastructure (NQI). The competence of their personnel directly impacts the accuracy and reliability of test results and the validity of the laboratory's accreditation. In recent years, with the updating of the "Accreditation Criteria for Testing and Inspection Bodies," the widespread application of "Quality Management System Requirements" (ISO / IEC 17025), and the stringent implementation of "CNAS-CL01:2018 <Accreditation Criteria for Testing and Calibration Laboratories>", unprecedentedly high standards have been set for the competence management of laboratory personnel. The criteria explicitly require laboratories to ensure that "all personnel who affect the accuracy of testing and / or calibration are competent based on appropriate education, training, experience, and / or verifiable skills," and to "maintain records of personnel competence, training, supervision, and authorization." These provisions make personnel training and competence verification the lifeline of laboratory management and a critical area where non-conformities are most likely to occur during external reviews and internal audits.
[0003] Based on the laboratory's current operational status and feedback from previous reviews, the common pain points in personnel training management can be summarized as a lack of three dimensions: systematic approach, personalization, and traceability. Specifically: First, the training system is unsystematic and subjective; second, the training process is "one-size-fits-all" and lacks personalization; third, there are difficulties in evaluating and recording training effectiveness.
[0004] To address these issues, academia and industry have explored various approaches, primarily focusing on intelligent tutoring systems and adaptive learning technologies. While these technologies have achieved some success in general education, significant technological gaps and limitations remain in highly specialized and rule-based laboratory training scenarios. Specifically: First, they fail to adequately consider the temporal nature of learning behaviors, making it difficult to accurately capture the inherent logic and dependence of skill learning; second, they cannot generate an accurate initial profile for laboratory personnel, thus hindering effective initial path recommendations; furthermore, the black-box output of the model makes it difficult to provide clear evidence to reviewers explaining "why this learning path should be recommended to this employee"; finally, existing technologies struggle to comprehensively balance multiple dimensions such as efficiency, difficulty, and logical coherence to plan a globally optimal learning solution for trainees.
[0005] In summary, there is an urgent need for a new technical solution for closed-loop management of laboratory personnel training based on standards and specifications. Summary of the Invention
[0006] The purpose of this application is to provide a closed-loop management method for laboratory personnel training based on standards and specifications, in order to solve the technical problems of fixed laboratory training paths, low level of intelligence, and inability to dynamically adapt to changes in personnel capabilities in the prior art.
[0007] To achieve the above objectives, this application provides a closed-loop management method for laboratory personnel training based on standards and specifications, the method comprising: An initial knowledge graph is constructed based on the standard specifications corresponding to all the knowledge points that laboratory personnel need to master. An improved association analysis algorithm with an introduced time sliding window mechanism was used to mine historical training and assessment data, calculate the confidence and support between knowledge points to update the initial knowledge graph and obtain the training knowledge graph. Based on the baseline of knowledge points corresponding to the job, a pre-job test is conducted on laboratory personnel. Based on the test results and the baseline, the knowledge points that need to be supplemented are identified, and a personnel diagnosis matrix is generated. Based on the personnel diagnosis matrix, a knowledge point mastery status vector is obtained to quantify the knowledge point mastery of laboratory personnel. Based on the knowledge point mastery state vector, an enhanced ant colony optimization algorithm is used to plan the learning path for laboratory personnel training on the training knowledge graph. The learning path is the path with the highest pheromone concentration. After laboratory personnel complete any learning node in the learning path and are assessed, the change in their knowledge mastery is evaluated, and this change is quantified into a reward signal. The reward signal is used to reinforce the planning of the learning path through the enhanced ant colony optimization algorithm; and the reinforcement process is recorded at least.
[0008] Preferably, the construction of the initial knowledge graph includes: The knowledge points in the standard specifications are extracted and atomized, and each independent, indivisible knowledge point is used as a node of the initial knowledge graph. The nodes of the initial knowledge graph are defined hierarchically and in terms of dependencies based on pre-set expert experience; The nodes of the initial knowledge graph are initially associated with each other based on the preset expert experience. The results of this association will be updated based on the calculation of the confidence and support. The edges of the initial knowledge graph are obtained based on the results of defining hierarchy and dependencies and marking initial associations.
[0009] Preferably, the construction of the training knowledge graph includes: The historical training and assessment data of each employee is treated as an event sequence sorted by timestamps. Based on the improved correlation analysis algorithm, the event sequence is divided into multiple subsequences that include specific time periods or a number of events, resulting in a transaction set. The transaction set Used for calculating the confidence level and the support level; The support is calculated, and the support is used to measure the knowledge points or combinations of knowledge points in the transaction set. The prevalence of occurrence in, the support The calculation formula is: in, This represents a learning task within the window. For the total number of transactions, The window also includes knowledge points and knowledge points The number of transactions; Calculate the confidence score, which is used to measure the amount of knowledge learned. After that, learn the knowledge points The probability, the confidence level The calculation formula is: in, Calculated simultaneously including knowledge points and knowledge points The number of transactions accounts for all, including knowledge points. The proportion of transactions; Get the preset minimum support threshold and minimum confidence threshold When an association rule A rule is defined as a strong association rule when both of the following conditions are met: All the strongly associated rules that have been mined are transformed into weighted directed edges in the initial knowledge graph to obtain the training knowledge graph.
[0010] Preferably, the generation of the personnel diagnosis matrix includes: Define a baseline for the comprehensive knowledge points corresponding to a job position. The definition of the baseline includes: constructing a baseline that includes... A collection of individual knowledge points Meanwhile, the pre-employment test includes The question; construct a matrix The matrix For one of Matrix, matrix The mathematical expression is: in, Indicates the first The question tested the first One knowledge point, Indicates the first The question did not test the first one. One knowledge point; Laboratory personnel complete the pre-job test, and based on their responses, a [database name] is generated. The answer score vector and the answer score vector Binarization is performed to obtain the answer score vector. The mathematical expression is: in, Indicates the first The question was answered correctly. Indicates the first The question was answered incorrectly; The personnel diagnosis matrix is generated based on the personnel diagnosis matrix generation formula. The formula for generating the personnel diagnosis matrix is: Where: if the first The question does not test knowledge points. ,Right now Then the corresponding position of the matrix is If the first The question tests the knowledge points ,Right now Then the value at the corresponding position in the matrix is the student's score on that question. ; Based on the personnel diagnosis matrix Generate a The knowledge points mentioned above refer to the mastery of state vectors. Its mathematical expression is: Among them, as long as there is one knowledge point being tested... The title The student did not answer correctly, that is and Then factors will appear in the product. This led to the final Only when all relevant questions are answered correctly will the final result be awarded. The set of knowledge points that need to be supplemented is all those in Median The knowledge points.
[0011] Preferably, the planning of the learning path includes: Based on the knowledge point mastery state vector, construct a set of all unmastered knowledge points. Its mathematical expression is: in, The knowledge point mastery state vector is defined; the enhanced ant colony algorithm, on the training knowledge graph, targets the knowledge point set. Perform optimal path search on the nodes in the array; In each iteration, virtual ants From the current knowledge point node Select the next node to learn. probability Based on pheromone concentration and heuristic information The decision is made jointly, and its mathematical expression is: in, It is a virtual ant Unvisited and belonging to a collection of knowledge points Candidate knowledge points; The heuristic information The formula for calculation is: in, For value inspiration, The edge weights in the training knowledge graph are... The difficulty level of the knowledge points; and: , For nodes The level of mastery of the knowledge points; , For nodes The difficulty level of the knowledge points; Each virtual ant Construct a learning path Then, a comprehensive quality score was given based on multiple dimensions. This includes learning paths to knowledge point sets. The coverage, total learning time, and logical coherence of the training knowledge graph are analyzed; the pheromones on the training knowledge graph are updated based on the scores of all learning paths. in, It refers to the pheromone evaporation rate and the pheromone increment. With path quality Proportional; After sufficient iterations, the learning path with the highest pheromone concentration was identified as the personalized learning plan recommended to laboratory personnel.
[0012] Preferably, the difficulty value The difficulty level was determined based on the first-time pass rate and average study time. The formula for calculation is: in, and These are the weight parameters for the corresponding indicators. This refers to the number of people who passed the knowledge point on their first assessment in historical training and evaluation data. This represents the total number of people who attempted this knowledge point in the historical training and assessment data. For students Master the knowledge points The amount of time spent studying All knowledge points already mastered in historical data The students gathered. and They are calculated for the maximum value and the minimum value, respectively.
[0013] Preferably, the reward signal is used to reinforce the planning of the learning path through the enhanced ant colony optimization algorithm, including: Capturing state changes, the knowledge mastery state of laboratory personnel before learning is the corresponding knowledge mastery state vector. The corresponding update after learning and assessment is as follows The reward signal The number of new knowledge points successfully mastered by laboratory personnel during this learning activity, i.e., the number of new knowledge points in the personnel diagnostic matrix. Become The number of knowledge points, the reward signal The formula for calculation is: in, It represents the total number of knowledge points; Reward signal This feedback is incorporated into the planning of the learning path using the enhanced ant colony optimization algorithm, where reinforcement is achieved by directly strengthening pheromones; this applies to the learning path during a single training session for lab personnel. , for learning path Each edge on Calculate an additional reward-based signal pheromone increment : in, To reinforce the learning rate parameter, which is used to adjust the reward signal. The intensity of the effect on pheromones; the pheromone update formula becomes: in, For pheromone evaporation rate, The reward signal does not exist. The pheromone increment calculated at that time. Based on the reward signal The calculated pheromone increment; The learning path is reinforced based on the updated pheromones, and all learning trajectories, assessment results, and dynamic adjustment processes of the learning path are automatically recorded with user identifiers and timestamps.
[0014] To achieve the above objectives, this application also provides a standard-based closed-loop management device for laboratory personnel training, which applies the standard-based closed-loop management method for laboratory personnel training as described above, including: The knowledge graph construction module is used to build an initial knowledge graph based on standard specifications corresponding to all the knowledge points that laboratory personnel need to master. The intelligent analysis module is used to mine historical training and assessment data using an improved association analysis algorithm that incorporates a time sliding window mechanism, calculate the confidence and support between knowledge points to update the initial knowledge graph and obtain the training knowledge graph. The personalized path planning module is used to set a baseline based on the full range of knowledge points corresponding to the job, conduct pre-job tests on laboratory personnel, identify the knowledge points that need to be supplemented based on the test results and the baseline, and generate a personnel diagnosis matrix; based on the personnel diagnosis matrix, a knowledge point mastery status vector is obtained to quantify the knowledge point mastery of laboratory personnel. The dynamic learning and assessment module is used to plan the learning path for laboratory personnel training on the training knowledge graph based on the knowledge point mastery state vector and through the enhanced ant colony optimization algorithm. The learning path is the path with the highest pheromone concentration. The traceable recording and output module is used to assess the change in the knowledge mastery status of laboratory personnel after they complete any learning node in the learning path and are assessed, and to quantify the change into a reward signal. The reward signal is used to reinforce the planning of the learning path through the enhanced ant colony optimization algorithm; and at least the reinforcement process is recorded.
[0015] To achieve the above objectives, this application also provides a computer device, including at least one processor, at least one memory, and a data bus; The processor and the memory communicate with each other via the data bus; The memory stores program instructions that can be executed by the processor, which invokes the program instructions to execute the standard-based closed-loop management method for laboratory personnel training as described above.
[0016] To achieve the above objectives, this application also provides a storage medium storing a computer program thereon, which, when executed by a processor, implements the standard-based closed-loop management method for laboratory personnel training as described above.
[0017] Beneficial effects: The standard-based closed-loop management method for laboratory personnel training proposed in this application systematically solves the technical problems of existing laboratory training paths being fixed, having low intelligence, and being unable to dynamically adapt to changes in personnel capabilities by introducing a pre-job test diagnostic matrix, an enhanced ant colony optimization algorithm, and a reinforcement learning feedback mechanism based on real results. Attached Figure Description
[0018] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0019] Figure 1 A flowchart illustrating a closed-loop management method for laboratory personnel training based on standards and specifications, provided for embodiments of this application; Figure 2 A flowchart illustrating the construction of an initial knowledge graph provided in this application embodiment; Figure 3 A flowchart illustrating the construction of a training knowledge graph provided in this application embodiment; Figure 4 A flowchart illustrating the generation of the personnel diagnosis matrix provided in this application embodiment; Figure 5 A flowchart illustrating the planning of a learning path provided in an embodiment of this application; Figure 6 A flowchart illustrating the enhanced planning of learning paths provided in this application embodiment; Figure 7 A structural block diagram of a standard-compliant closed-loop management device for laboratory personnel training provided in this application embodiment.
[0020] The implementation, functional features, and advantages of this invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation
[0021] The technical solutions in the embodiments of this application will be clearly and completely described below. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.
[0022] In this document, the term "comprising" is intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.
[0023] To address the technical challenges of existing laboratory training paths being fixed, lacking intelligence, and unable to dynamically adapt to changes in personnel capabilities, this embodiment systematically solves all of the above challenges by introducing a pre-job testing and diagnostic matrix, an enhanced ant colony optimization algorithm, and a reinforcement learning feedback mechanism based on real-world results.
[0024] Reference Figure 1 , Figure 1 This is a flowchart illustrating the closed-loop management method for laboratory personnel training based on standards and specifications in this embodiment.
[0025] like Figure 1 As shown in the figure, this embodiment discloses a closed-loop management method for laboratory personnel training based on standards and specifications. The method includes: S10: Construct an initial knowledge graph based on the standard specifications corresponding to all the knowledge points that laboratory personnel need to master.
[0026] In this specific application, a weighted domain knowledge graph based on standard specifications is constructed. The knowledge points that the laboratory needs to master are atomized and broken down as the basic nodes of the knowledge graph. Subsequently, the relationships between the nodes are defined.
[0027] Reference Figure 2 , Figure 2 This is a flowchart illustrating the construction of the initial knowledge graph in this embodiment.
[0028] Specifically, such as Figure 2 As shown, the construction of the initial knowledge graph includes: S11: Knowledge Point Extraction and Atomization: The knowledge points in the aforementioned standard specifications are extracted and atomized, with each independent, indivisible knowledge point serving as a node in the initial knowledge graph. In this specific application, firstly, all relevant knowledge points are systematically extracted from authoritative documents such as CNAS-CL01:2018, various testing standards, instrument operating procedures, and safety manuals. Subsequently, these knowledge points are atomized and decomposed to ensure that each node represents an independent, indivisible knowledge unit, serving as a basic node in the knowledge graph.
[0029] S12: Definition of Hierarchy and Dependencies: Based on preset expert experience, the nodes of the initial knowledge graph are defined in terms of hierarchy and dependencies. In the specific application of this embodiment, based on the existing internal logic of laboratory knowledge and expert experience, the hierarchy and prerequisite dependencies between knowledge nodes are defined. Subdivided knowledge points collectively constitute a more macroscopic knowledge domain. The skeleton of the knowledge graph is constructed to ensure the logical correctness of subsequent learning path planning.
[0030] S13: Initial Association Labeling: Based on preset expert experience, initial association relationships are labeled for the nodes of the initial knowledge graph. The labeling results will be updated based on the calculated confidence and support levels. In this specific application, non-dependent, non-hierarchical association relationships are defined between different knowledge nodes based on the inherent correlation between historical data and expert experience or knowledge. The association relationships established in this step are initial settings and will be dynamically optimized and weighted through historical data mining in subsequent steps.
[0031] The edges of the initial knowledge graph are obtained based on the results of defining hierarchy and dependencies and marking initial associations.
[0032] S20: Using an improved association analysis algorithm that incorporates a time sliding window mechanism, historical training and assessment data are mined to calculate the confidence and support between knowledge points in order to update the initial knowledge graph and obtain the training knowledge graph.
[0033] In the specific application of this embodiment, for the mining of knowledge point correlations in historical data, an improved correlation analysis algorithm incorporating a time sliding window mechanism is used to mine historical training and assessment data, calculating the confidence and support between knowledge points. Finally, the mined strong correlation rules are used as weighted directed edges to form a weighted knowledge network that reflects the logical and dependency strength between knowledge points.
[0034] Reference Figure 3 , Figure 3 This is a flowchart illustrating the construction of the training knowledge graph in this embodiment.
[0035] Specifically, such as Figure 3 As shown, the construction of the training knowledge graph includes: S21: Training data serialization and window truncation: Treat each employee's historical training and assessment data as an event sequence sorted by timestamp, and based on the improved association analysis algorithm, divide the event sequence into multiple subsequences including specific time periods or event numbers to obtain a transaction set. The transaction set Used for calculating the confidence and support; in the specific application of this embodiment, subsequent association analysis will only be performed within these windowed transaction sets.
[0036] S22: Support calculation, calculating the support, the support being used to measure the knowledge points or combinations of knowledge points in the transaction set. The prevalence of occurrence in, the support The calculation formula is: in, This represents a learning task within the window. For the total number of transactions, The window also includes knowledge points and knowledge points The number of transactions; S23: Confidence score calculation, the confidence score is used to measure the learning of knowledge points. After that, learn the knowledge points The probability, the confidence level The calculation formula is: in, Calculated simultaneously including knowledge points and knowledge points The number of transactions accounts for all, including knowledge points. The proportion of transactions; S24: Strong association rule filtering and weighting to obtain the preset minimum support threshold. and minimum confidence threshold When an association rule A rule is defined as a strong association rule when both of the following conditions are met: All the mined strong association rules are transformed into weighted directed edges in the initial knowledge graph to obtain the training knowledge graph. In one application of this embodiment, for example, for strong association rules... This will create a path from node in the knowledge graph. Pointing to node The weight of a directed edge. Its confidence level can be assigned a value: .
[0037] S30: Based on the knowledge points of the entire domain corresponding to the job, a baseline is set, and a pre-job test is conducted on laboratory personnel. Based on the test results and the baseline, the knowledge points that need to be supplemented are identified, and a personnel diagnosis matrix is generated. Based on the personnel diagnosis matrix, a knowledge point mastery state vector is obtained to quantify the knowledge point mastery of laboratory personnel.
[0038] In the specific application of this embodiment, we determine the trainee profile based on the test completion, set a baseline based on the knowledge points of the entire job domain, and conduct a pre-job test on the trainees. This test includes the knowledge points required for the job. Based on the baseline, we compare it with the diagnostic matrix to find the knowledge points that need to be supplemented and generate a trainee diagnostic matrix, where rows represent knowledge points and columns represent questions.
[0039] Reference Figure 4 , Figure 4 This is a flowchart illustrating the generation of the personnel diagnosis matrix in this embodiment.
[0040] Specifically, such as Figure 4 As shown, the generation of the personnel diagnosis matrix includes: S31: Baseline and The matrix definition defines the baseline corresponding to the full-domain knowledge points for a given job position. The definition of the baseline includes: constructing a matrix that includes... A collection of individual knowledge points Meanwhile, the pre-employment test includes The question; construct a matrix The matrix For one of Matrix, matrix The mathematical expression is: in, Indicates the first The question tested the first One knowledge point, Indicates the first The question did not test the first one. One knowledge point.
[0041] S32: Laboratory personnel answer and score. Laboratory personnel complete the pre-job test, and a score is generated based on their answers. The answer score vector and the answer score vector Binarization is performed to obtain the answer score vector. The mathematical expression is: in, Indicates the first The question was answered correctly. Indicates the first The question was answered incorrectly; S33: Diagnosis matrix generation, generating the personnel diagnosis matrix based on the personnel diagnosis matrix generation formula. The output of this step is the final trainee diagnostic matrix. The dimension of this matrix is also 1. The rows represent knowledge points, and the columns represent questions. The elements in the matrix precisely reflect the trainee's mastery of the relevant knowledge points for each specific question. The formula for generating the personnel diagnosis matrix is: Where: if the first The question does not test knowledge points. ,Right now Then the corresponding position of the matrix is If the first The question tests the knowledge points ,Right now Then the value at the corresponding position in the matrix is the student's score on that question. ,in, for or This formula clearly shows which questions students have failed to master and which knowledge points they have not mastered.
[0042] S34: Knowledge Gap Identification. To determine the knowledge points that need to be supplemented from a macro perspective, the system needs to use a diagnostic matrix. Generate a Mastering the knowledge points of state vectors A knowledge point is considered "mastered" only if the student answers all the questions testing that knowledge point correctly. The mathematical expression for this is: Among them, as long as there is one knowledge point being tested... The title The student did not answer correctly, that is and Then factors will appear in the product. This led to the final Only when all relevant questions are answered correctly will the final result be awarded. The set of knowledge points that need to be supplemented is all those in Median The knowledge points.
[0043] S40: Based on the knowledge point mastery state vector, an enhanced ant colony optimization algorithm is used to plan the learning path for laboratory personnel training on the training knowledge graph. The learning path is the path with the highest pheromone concentration.
[0044] In the specific application of this embodiment, the optimal learning path planning based on the diagnostic matrix involves using an enhanced ant colony optimization algorithm based on the aforementioned generated personnel diagnostic matrix. The aim is to plan the most efficient learning path to compensate for the diagnosed knowledge gaps. This is guided by a heuristic function that parses the diagnostic matrix, prioritizing knowledge gaps. By iteratively scoring potential paths and reinforcing efficient routes with pheromones, the system ultimately determines the path with the highest pheromone concentration as the personalized learning plan recommended to the trainees.
[0045] Reference Figure 5 , Figure 5 This is a flowchart illustrating the planning of the learning path in this embodiment.
[0046] Specifically, such as Figure 5 As shown, the planning of the learning path includes: S41: Definition of learning objectives and search space. The input to this step is the knowledge point mastery state vector generated earlier. The system first determines the core learning objective based on this vector, which is the set of all knowledge points that have not yet been mastered. Its mathematical expression is: in, The knowledge point mastery state vector is defined; the enhanced ant colony algorithm, on the training knowledge graph, targets the knowledge point set. Perform optimal path search on the nodes in the array; S42: Ant colony path construction and probabilistic transition model, in each iteration, virtual ants... From the current knowledge point node Select the next node to learn. probability Based on pheromone concentration and heuristic information The decision is made jointly, and its mathematical expression is: in, It is a virtual ant Unvisited and belonging to a collection of knowledge points Candidate knowledge points; S43: Heuristic function based on diagnostic results, heuristic information This is the core of the algorithm's intelligence; it directly analyzes the diagnostic results of personnel training and sets knowledge gaps as the highest priority. Its calculation formula consists of multiple heuristic factors, among which value heuristics... The heuristic information plays a leading role. The formula for calculation is: in: For inspiration of value; The edge weights in the training knowledge graph represent the degree of correlation between knowledge points; The learning difficulty of the knowledge points.
[0047] In the specific application of this embodiment, It is a state vector of knowledge points mastery This directly determines the necessity of evaluating learning node j, ensuring that only unmastered knowledge points are considered. Only those with the highest inspirational value (value) have the greatest inspirational value. The value of the knowledge points already acquired is Therefore, it is automatically ignored in the search. The formula is: in, For nodes The level of mastery of the knowledge points; , For nodes The difficulty level of the knowledge points.
[0048] As a preferred embodiment of this example, the difficulty value Defined based on historical user data, it includes two metrics: the first-time pass rate and the average learning time, and the difficulty level. The formula for calculation is: in, and These are the weight parameters for the corresponding indicators. This refers to the number of people who passed the knowledge point on their first assessment in historical training and evaluation data. This represents the total number of people who attempted this knowledge point in the historical training and assessment data. For students Master the knowledge points The amount of time spent studying All knowledge points already mastered in historical data The students gathered. and They are calculated for the maximum value and the minimum value, respectively.
[0049] S44: Path quality score and pheromone update for each virtual ant Construct a learning path Then, a comprehensive quality score was given based on multiple dimensions. This includes learning paths to knowledge point sets. The coverage, total learning time, and logical coherence of the training knowledge graph are analyzed; the pheromones on the training knowledge graph are updated based on the scores of all learning paths. in, It refers to the pheromone evaporation rate and the pheromone increment. With path quality Proportional; edges traversed by high-quality paths will acquire more pheromones, thus guiding subsequent virtual ants. They tend to explore more efficient paths.
[0050] S45: Optimal path determination: After sufficient iteration, pheromones accumulate along the optimal or near-optimal path. Ultimately, the path with the highest pheromone concentration is determined as the most efficient personalized learning plan recommended by the system to trainees.
[0051] S50: After laboratory personnel complete any learning node in the learning path and are assessed, evaluate the change in their knowledge mastery status and quantify the change into a reward signal. The reward signal is used to reinforce the planning of the learning path through the enhanced ant colony optimization algorithm; and at least record the reinforcement process.
[0052] In this specific application, we employ a closed-loop dynamic optimization based on reinforcement learning. To achieve continuous adaptive adjustment of the path, a closed-loop optimization framework based on reinforcement learning is constructed. After a learner completes a learning node on the recommended path and passes the assessment, the knowledge tracker evaluates the change in their knowledge mastery status (e.g., a "0" in the diagnostic matrix successfully becomes a "1") and quantifies this positive change as a positive reward signal. This reward signal is used to directly reinforce the ant colony optimization algorithm used in the path planning. Specifically, learning paths that generate high rewards (i.e., significant learning effects) receive additional pheromone rewards for the edges they traverse on the knowledge network.
[0053] Reference Figure 6 , Figure 6 This is a flowchart illustrating the enhanced planning of the learning path in this embodiment.
[0054] Specifically, such as Figure 6 As shown, the reward signal is used to reinforce the planning of the learning path through the enhanced ant colony optimization algorithm, including: S51: Status Assessment and Reward Signal Generation. This step is triggered after the learner completes the recommended learning path and passes the assessment. The system assesses changes in the learner's knowledge mastery status through knowledge tracking and quantifies them as reward signals. It captures these status changes; the learner's prior knowledge mastery status is defined as the corresponding knowledge mastery status vector. The corresponding update after learning and assessment is as follows The reward signal The number of new knowledge points successfully mastered by laboratory personnel during this learning activity, i.e., the number of new knowledge points in the personnel diagnostic matrix. Become The number of knowledge points, the reward signal The formula for calculation is: in, It represents the total number of knowledge points; since the state value is... or This formula can accurately calculate the number of newly acquired knowledge points, and its function is to generate a large positive reward value.
[0055] S52: Reward-based reinforcement learning update rule, this step updates the learning effect in the real world (reward signal). The reward signal is fed back into the ant colony optimization algorithm, where it is dynamically optimized by directly reinforcing the ant colony's memory, i.e., through pheromones. This feedback is incorporated into the planning of the learning path using the enhanced ant colony optimization algorithm, where reinforcement is achieved by directly strengthening pheromones; this applies to the learning path during a single training session for lab personnel. , for learning path Each edge on Calculate an additional reward-based signal pheromone increment : in, To reinforce the learning rate parameter, which is used to adjust the reward signal. The intensity of the impact on pheromones; the final pheromone update, with additional pheromone rewards, will be added on top of the original pheromone update rules. Therefore, the complete update formula becomes: in: For pheromone evaporation rate; The reward signal does not exist. The pheromone increment calculated in time, that is, the pheromone increment calculated based on path prediction quality; Based on the reward signal The calculated pheromone increment is the pheromone increment calculated based on the actual learning effect.
[0056] S53: Dynamic optimization closed loop, through the above update rules, can efficiently help students master knowledge (i.e., obtain high...). The pheromone concentration is doubly enhanced (from internal scoring and external rewards) along the learning path, forming a complete closed loop of recommendation, learning, assessment, reward, and reinforcement. All learning trajectories, assessment results, and dynamic adjustments to the path are automatically recorded by the system with user identifiers and timestamps, and can be exported with one click. This provides logically complete and intelligently generated training evidence for external review, thereby completely resolving qualification audit risks.
[0057] Reference Figure 7 , Figure 7 This is a structural block diagram of the standard-based closed-loop management device for laboratory personnel training in this embodiment.
[0058] like Figure 7 As shown, this embodiment discloses a standard-based closed-loop management device for laboratory personnel training, applying the standard-based closed-loop management method for laboratory personnel training as described above, including: The knowledge graph construction module is used to build an initial knowledge graph based on the standard specifications corresponding to all the knowledge points that laboratory personnel need to master. In the specific application of this implementation, the knowledge graph construction module atomically decomposes the standard specifications such as laboratory safety requirements, technical capability requirements and quality assurance requirements to construct a domain knowledge graph containing knowledge nodes and relationships.
[0059] The intelligent analysis module utilizes an improved association analysis algorithm incorporating a time-sliding window mechanism to mine historical training and assessment data, calculate the confidence and support between knowledge points to update the initial knowledge graph, and obtain the training knowledge graph. In this specific application, the intelligent analysis module stores the knowledge graph and historical training and assessment data; and uses an association analysis algorithm to mine historical data to dynamically update and quantify the relationships between knowledge points in the knowledge graph.
[0060] The personalized learning path planning module is used to set a baseline based on the comprehensive knowledge points corresponding to the job, conduct pre-job tests for laboratory personnel, identify the knowledge points that need to be supplemented based on the test results and the baseline, and generate a personnel diagnosis matrix. Based on the personnel diagnosis matrix, a knowledge point mastery state vector is obtained to quantify the knowledge point mastery status of laboratory personnel. In the specific application of this implementation, the personalized learning path planning module builds a user profile based on the user's pre-job diagnosis or historical data, identifies their knowledge gaps, and uses an improved ant colony algorithm to search in the knowledge graph with the knowledge nodes to be mastered as the target, to plan a personalized learning path for the user that meets the preconditions and has the best learning efficiency.
[0061] The dynamic learning and assessment module, based on the knowledge point mastery state vector, uses an enhanced ant colony optimization algorithm to plan the learning path for laboratory personnel on the training knowledge graph. The learning path is the one with the highest pheromone concentration. In this specific application, the dynamic learning and assessment module displays the training content on the user's personalized learning path, monitors the learning progress, and assesses the user after completing a learning node. The module updates the user's state in the knowledge graph in real time based on the assessment results and triggers the path planning module to dynamically adjust the subsequent learning path, forming a closed loop of "diagnosis-learning-assessment-re-diagnosis".
[0062] The traceable recording and output module is used to assess changes in the knowledge mastery of laboratory personnel after they complete any learning node in the learning path and pass the assessment. This change is quantified into a reward signal, which is used to reinforce the planning of the learning path using an enhanced ant colony optimization algorithm. The module also records at least the reinforcement process. In this specific application, the traceable recording and output module records all data associated with the user, including not only learning and assessment results but also the basis for generating the personalized learning path, its content, and the trajectory of each dynamic adjustment. All records are appended with user identifiers and timestamps. This module also supports exporting these complete electronic records, which reflect the training logic, to standardized formats (such as Excel and PDF) for external review or internal audit.
[0063] This embodiment also discloses a computer device, including at least one processor, at least one memory, and a data bus; The processor and the memory communicate with each other via the data bus; The memory stores program instructions that can be executed by the processor, which invokes the program instructions to execute the standard-based closed-loop management method for laboratory personnel training as described above.
[0064] This embodiment also discloses a storage medium on which a computer program is stored. When the computer program is executed by a processor, it implements the closed-loop management method for laboratory personnel training based on standards and specifications as described above.
[0065] It should be noted that the standard-based laboratory personnel training closed-loop management device, computer equipment, and storage medium in this embodiment correspond to the aforementioned standard-based laboratory personnel training closed-loop management method. Therefore, any content not specifically described in the standard-based laboratory personnel training closed-loop management device, computer equipment, and storage medium in this embodiment, including but not limited to functional definitions, working principles, and technical effects, can be referred to the description in the aforementioned standard-based laboratory personnel training closed-loop management method, and will not be repeated here.
[0066] In the embodiments provided in this application, it should be understood that the embodiments described herein can be implemented in hardware, software, firmware, middleware, code, or any suitable combination thereof. For hardware implementation, the processor may be implemented in one or more of the following: application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), processors, controllers, microcontrollers, microprocessors, other electronic units designed to implement the functions described herein, or combinations thereof. For software implementation, some or all of the processes of the embodiments may be performed by a computer program instructing the associated hardware. During implementation, the program may be stored in a computer-readable storage medium or transmitted as one or more instructions or code on a computer-readable storage medium. Computer-readable storage media include computer storage media and communication media, wherein communication media include any medium that facilitates the transmission of a computer program from one place to another. Storage media may be any available medium accessible to a computer. Computer-readable storage media may include, but are not limited to, RAM, ROM, EEPROM, CD-ROM or other optical disk storage, magnetic disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code having the form of instructions or data structures and accessible to a computer.
[0067] Finally, it should be noted that the above description is only a preferred embodiment of this application and is not intended to limit this application. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.
Claims
1. A closed-loop management method for laboratory personnel training based on standards and specifications, characterized in that, The method includes: An initial knowledge graph is constructed based on the standard specifications corresponding to all the knowledge points that laboratory personnel need to master. An improved association analysis algorithm with an introduced time sliding window mechanism was used to mine historical training and assessment data, calculate the confidence and support between knowledge points to update the initial knowledge graph and obtain the training knowledge graph. Based on the baseline of knowledge points corresponding to the job, a pre-job test is conducted on laboratory personnel. Based on the test results and the baseline, the knowledge points that need to be supplemented are identified, and a personnel diagnosis matrix is generated. Based on the personnel diagnosis matrix, a knowledge point mastery status vector is obtained to quantify the knowledge point mastery of laboratory personnel. Based on the knowledge point mastery state vector, an enhanced ant colony optimization algorithm is used to plan the learning path for laboratory personnel training on the training knowledge graph. The learning path is the path with the highest pheromone concentration. After laboratory personnel complete any learning node in the learning path and are assessed, the change in their knowledge mastery is evaluated, and this change is quantified into a reward signal. The reward signal is used to reinforce the planning of the learning path through the enhanced ant colony optimization algorithm; and the reinforcement process is recorded at least.
2. The closed-loop management method for laboratory personnel training based on standards and specifications according to claim 1, characterized in that, The construction of the initial knowledge graph includes: The knowledge points in the standard specifications are extracted and atomized, and each independent, indivisible knowledge point is used as a node of the initial knowledge graph. The nodes of the initial knowledge graph are defined hierarchically and in terms of dependencies based on pre-set expert experience; The nodes of the initial knowledge graph are initially associated with each other based on the preset expert experience. The results of this association will be updated based on the calculation of the confidence and support. The edges of the initial knowledge graph are obtained based on the results of defining hierarchy and dependencies and marking initial associations.
3. The closed-loop management method for laboratory personnel training based on standards and specifications according to claim 1, characterized in that, The construction of the training knowledge graph includes: The historical training and assessment data of each employee is treated as an event sequence sorted by timestamps. Based on the improved correlation analysis algorithm, the event sequence is divided into multiple subsequences that include specific time periods or a number of events, resulting in a transaction set. The transaction set Used for calculating the confidence level and the support level; The support is calculated, and the support is used to measure the knowledge points or combinations of knowledge points in the transaction set. The prevalence of occurrence in, the support The calculation formula is: in, This represents a learning task within the window. For the total number of transactions, The window also includes knowledge points and knowledge points The number of transactions; Calculate the confidence score, which is used to measure the amount of knowledge learned. After that, learn the knowledge points The probability, the confidence level The calculation formula is: in, Calculated simultaneously including knowledge points and knowledge points The number of transactions accounts for all, including knowledge points. The proportion of transactions; Get the preset minimum support threshold and minimum confidence threshold When an association rule A rule is defined as a strong association rule when both of the following conditions are met: All the strongly associated rules that have been mined are transformed into weighted directed edges in the initial knowledge graph to obtain the training knowledge graph.
4. The closed-loop management method for laboratory personnel training based on standards and specifications according to claim 1, characterized in that, The generation of the personnel diagnosis matrix includes: Define a baseline for the comprehensive knowledge points corresponding to a job position. The definition of the baseline includes: constructing a baseline that includes... A collection of individual knowledge points Meanwhile, the pre-employment test includes The question; construct a matrix The matrix For one of Matrix, matrix The mathematical expression is: in, Indicates the first The question tested the first One knowledge point, Indicates the first The question did not test the first one. One knowledge point; Laboratory personnel complete the pre-job test, and based on their responses, a [database name] is generated. The answer score vector and the answer score vector Binarization is performed to obtain the answer score vector. The mathematical expression is: in, Indicates the first The question was answered correctly. Indicates the first The question was answered incorrectly; The personnel diagnosis matrix is generated based on the personnel diagnosis matrix generation formula. The formula for generating the personnel diagnosis matrix is: Where: if the first The question does not test knowledge points. ,Right now Then the corresponding position of the matrix is If the first The question tests the knowledge points ,Right now Then the value at the corresponding position in the matrix is the student's score on that question. ; Based on the personnel diagnosis matrix Generate a The knowledge points mentioned above refer to the mastery of state vectors. Its mathematical expression is: Among them, as long as there is one knowledge point being tested... The title The student did not answer correctly, that is and Then factors will appear in the product. This led to the final Only when all relevant questions are answered correctly will the final result be awarded. The set of knowledge points that need to be supplemented is all those in Median The knowledge points.
5. The closed-loop management method for laboratory personnel training based on standards and specifications according to claim 1, characterized in that, The planning of the learning path includes: Based on the knowledge point mastery state vector, construct a set of all unmastered knowledge points. Its mathematical expression is: in, The knowledge point mastery state vector is defined; the enhanced ant colony algorithm, on the training knowledge graph, targets the knowledge point set. Perform optimal path search on the nodes in the array; In each iteration, virtual ants From the current knowledge point node Select the next node to learn. probability Based on pheromone concentration and heuristic information The decision is made jointly, and its mathematical expression is: in, It is a virtual ant Unvisited and belonging to a collection of knowledge points Candidate knowledge points; The heuristic information The formula for calculation is: in, For value inspiration, The edge weights in the training knowledge graph are... The learning difficulty of the knowledge points; and: , For nodes The level of mastery of the knowledge points; , For nodes The difficulty level of the knowledge points; Each virtual ant Construct a learning path Then, a comprehensive quality score was given based on multiple dimensions. This includes learning paths to knowledge point sets. The coverage, total learning time, and logical coherence of the training knowledge graph are analyzed; the pheromones on the training knowledge graph are updated based on the scores of all learning paths. in, It refers to the pheromone evaporation rate and the pheromone increment. With path quality Proportional; After sufficient iterations, the learning path with the highest pheromone concentration was identified as the personalized learning plan recommended to laboratory personnel.
6. The closed-loop management method for laboratory personnel training based on standards and specifications according to claim 5, characterized in that, The difficulty value The difficulty level was determined based on the first-time pass rate and average study time. The formula for calculation is: in, and These are the weight parameters for the corresponding indicators. This refers to the number of people who passed the knowledge point on their first assessment in historical training and evaluation data. This refers to the total number of people who attempted this knowledge point in the historical training and assessment data. For students Master the knowledge points The amount of time spent studying All knowledge points already mastered in historical data The students gathered. and The calculations are for the maximum value and the minimum value, respectively.
7. The closed-loop management method for laboratory personnel training based on standards and specifications according to claim 1, characterized in that, The reward signal is used to reinforce the planning of the learning path through the enhanced ant colony optimization algorithm, including: Capturing state changes, the knowledge mastery state of laboratory personnel before learning is the corresponding knowledge mastery state vector. The corresponding update after learning and assessment is as follows The reward signal The number of new knowledge points successfully mastered by laboratory personnel during this learning activity, i.e., the number of new knowledge points in the personnel diagnostic matrix. Become The number of knowledge points, the reward signal The formula for calculation is: in, It represents the total number of knowledge points; Reward signal This feedback is incorporated into the planning of the learning path using the enhanced ant colony optimization algorithm, where reinforcement is achieved by directly strengthening pheromones; this applies to the learning path during a single training session for lab personnel. , for learning path Each edge on Calculate an additional reward-based signal pheromone increment : in, To reinforce the learning rate parameter, which is used to adjust the reward signal. The intensity of the effect on pheromones; the pheromone update formula becomes: in, For pheromone evaporation rate, The reward signal does not exist. The pheromone increment calculated at that time. Based on the reward signal The calculated pheromone increment; The learning path is reinforced based on the updated pheromones, and all learning trajectories, assessment results, and dynamic adjustment processes of the learning path are automatically recorded with user identifiers and timestamps.
8. A closed-loop management device for laboratory personnel training based on standards and specifications, employing the closed-loop management method for laboratory personnel training based on standards and specifications as described in any one of claims 1 to 7, characterized in that, include: The knowledge graph construction module is used to build an initial knowledge graph based on standard specifications corresponding to all the knowledge points that laboratory personnel need to master. The intelligent analysis module is used to mine historical training and assessment data using an improved association analysis algorithm that incorporates a time sliding window mechanism, calculate the confidence and support between knowledge points to update the initial knowledge graph and obtain the training knowledge graph. The personalized path planning module is used to set a baseline based on the full range of knowledge points corresponding to the job, conduct pre-job tests on laboratory personnel, identify the knowledge points that need to be supplemented based on the test results and the baseline, and generate a personnel diagnosis matrix. Based on the personnel diagnosis matrix, a knowledge point mastery state vector is obtained to quantify the knowledge point mastery of laboratory personnel. The dynamic learning and assessment module is used to plan the learning path for laboratory personnel training on the training knowledge graph based on the knowledge point mastery state vector and through the enhanced ant colony optimization algorithm. The learning path is the path with the highest pheromone concentration. The traceable recording and output module is used to assess the change in the knowledge mastery status of laboratory personnel after they complete any learning node in the learning path and are assessed, and to quantify the change into a reward signal. The reward signal is used to reinforce the planning of the learning path through the enhanced ant colony optimization algorithm; and at least the reinforcement process is recorded.
9. A computer device, characterized in that, Includes at least one processor, at least one memory, and a data bus; The processor and the memory communicate with each other via the data bus; The memory stores program instructions that can be executed by the processor, which invokes the program instructions to execute the standard-based closed-loop management method for laboratory personnel training as described in any one of claims 1 to 7.
10. A storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the closed-loop management method for laboratory personnel training based on standards and specifications as described in any one of claims 1 to 7.