Balance training data management method and system

Abstract

The invention discloses a balance training data management method and system, and relates to the technical field of balance training data management, and the method comprises the following steps: constructing a complete abnormal behavior data set which comprises all data used for training an abnormal behavior detection model, and on the basis, carrying out the detailed division of the data set, dividing the abnormal behavior data into two parts: synthetic abnormal behavior data and actual abnormal behavior data; according to the method, the dependence of the abnormal behavior detection model on the synthetic data is monitored in real time, the pre-training machine learning model is utilized to predict and optimize the training strategy, the model is ensured to learn more actual abnormal behavior characteristics, and the generalization ability and the detection precision are improved. Through hierarchical training and dynamic adjustment of the data proportion, false alarms are reduced, system response efficiency is improved, public safety is guaranteed, and through real-time monitoring and optimization, it is ensured that the model adapts to data changes, robustness is enhanced, stable operation in a complex environment is achieved, and long-term maintenance and upgrading are supported.

Description

Technical Field

[0001] The present invention relates to the technical field of balanced training data management, and in particular to a balanced training data management method and system. Background Art

[0002] The balanced training data management system is a system used to manage and optimize data distribution during the training process of machine learning models. It analyzes the imbalanced distribution of categories or features in the training data set, and uses data enhancement, undersampling, oversampling and other technologies to adjust the proportion of the data set to ensure that the model can fully learn different categories or features and avoid the model being biased towards a certain category, thereby improving the generalization ability and prediction accuracy of the model. The system can also monitor the changes in data balance during training in real time and dynamically adjust strategies to adapt to different training needs.

[0003] In video surveillance, the amount of normal behavior data is much greater than that of abnormal behavior samples, which will cause the model to tend to predict normal behavior, thereby increasing the risk of missed reports (failure to detect abnormalities) and false reports (misjudging normal behavior as abnormal). The balanced training data management system expands the number and diversity of abnormal samples by synthesizing abnormal behavior data (such as using generative adversarial networks (GANs) or data enhancement technology), ensuring that the abnormal behavior detection model has a more comprehensive understanding of the characteristics of abnormal behavior during training. In addition, the system can also dynamically adjust the data sampling strategy based on real-time feedback, optimize the detection threshold of the abnormal behavior monitoring model, and further reduce false reports and missed reports, thereby enhancing the monitoring system's recognition accuracy and response capabilities to abnormal events.

[0004] When the balanced training data management system enhances the detection capability of the abnormal behavior detection model by synthesizing abnormal behavior data, if the abnormal behavior detection model over-relies on the features of the synthetic abnormal behavior data for learning and ignores the diversity of actual abnormal behaviors, the existing technology is usually unable to detect this problem in time. When this problem occurs, the abnormal behavior detection model may misjudge and identify normal behavior as abnormal behavior. For example, in banks, shopping malls or transportation hubs, ordinary customer activities are misjudged as suspicious behavior. Frequent false alarms will cause the system to trigger wrong emergency responses (such as repeated dispatch of security personnel and equipment shutdown). Emergency resources are exhausted, and real events cannot be handled in time, further amplifying security vulnerabilities and social panic.

[0005] The above information disclosed in this Background section is only for enhancement of understanding of the background of the present disclosure and therefore it may contain information that does not constitute the prior art that is already known to one of ordinary skill in the art. Summary of the invention

[0006] The purpose of the present invention is to provide a balanced training data management method and system, which ensures that the abnormal behavior detection model does not rely too much on synthetic data by real-time monitoring of the dependence of the abnormal behavior detection model on synthetic data, and uses a pre-trained machine learning model to predict and optimize the training strategy, so as to learn more about the actual abnormal behavior characteristics and capture the diversity of real scenes. A hierarchical training method is adopted, which is first preliminarily trained with actual abnormal behavior data, and then fine-tuned with synthetic data to improve the generalization ability and detection accuracy of the model, effectively reduce false alarms and erroneous responses, save emergency resources, improve the response efficiency of the system in real events, and ensure public safety. The system dynamically adjusts the data ratio and training strategy to maintain healthy learning of the model, and corrects problems in time through real-time monitoring to enhance its robustness, so that the model can adapt to data changes and new abnormal behaviors, ensure stable operation in complex environments, and support long-term maintenance and upgrades to solve the problems in the above-mentioned background technology.

[0007] In order to achieve the above object, the present invention provides the following technical solution: a balanced training data management method, comprising the following steps:

[0008] Build a complete abnormal behavior dataset, which contains all the data used to train the abnormal behavior detection model. On this basis, divide the dataset into two parts: synthetic abnormal behavior data and actual abnormal behavior data.

[0009] During the training of the abnormal behavior detection model, the training status of the synthetic abnormal behavior data is obtained in real time, and key information related to the training of the synthetic abnormal behavior data is extracted from the real-time monitored data;

[0010] In the monitoring window, the extracted key information is analyzed and processed, and the feature vector after analysis and processing is input into the pre-trained machine learning model, and the pre-trained machine learning model is used to predict the usage of the synthetic abnormal behavior data in the current training stage;

[0011] According to the prediction results of the machine learning model, the usage of synthetic abnormal behavior data is classified into normal use and over-reliance;

[0012] If the prediction result is normal use, continue to maintain the current training strategy and data usage method. At this time, the abnormal behavior detection model is healthy in learning the synthetic abnormal behavior data and the actual abnormal behavior data, which can effectively improve the generalization ability and detection accuracy of the model.

[0013] For cases that are judged to be over-reliant, the composition ratio of the training data set is readjusted to increase the weight of the actual abnormal behavior data and reduce the proportion of abnormal behavior data, so as to encourage the abnormal behavior detection model to learn more about the characteristics of the actual abnormal behavior;

[0014] During the training process, a hierarchical training method is used to allow the abnormal behavior detection model to be initially trained on actual abnormal behavior data to learn the characteristics of real scenarios. Then, fine-tuning is performed on synthetic abnormal behavior data to enrich the abnormal behavior detection model's understanding of abnormal behavior and prevent the abnormal behavior detection model from being dominated by the characteristics of synthetic abnormal behavior data.

[0015] Based on the prediction results of the machine learning model, the usage of synthetic abnormal behavior data in the current training phase is judged in real time until the abnormal behavior detection model no longer relies too much on synthetic abnormal behavior data for training.

[0016] Preferably, the data set is divided into synthetic abnormal behavior data and actual abnormal behavior data in detail, and the specific steps are as follows:

[0017] Collect abnormal behavior data from multiple sources to ensure the breadth and coverage of the data;

[0018] Label the samples in the data set to identify which are synthetic abnormal behavior data and which are actual abnormal behavior data;

[0019] After the data is labeled, in-depth feature analysis is performed to evaluate the differences and quality of the characteristics of the synthetic abnormal behavior data and the actual abnormal behavior data;

[0020] After completing data annotation and analysis, the dataset is divided into synthetic abnormal behavior dataset and actual abnormal behavior dataset according to training requirements and data distribution.

[0021] Preferably, the use of the synthetic abnormal behavior data in the current training phase is predicted using a pre-trained machine learning model, and the specific steps are as follows:

[0022] During the training process of the abnormal behavior detection model, indicators related to the training status of synthetic abnormal behavior data are obtained in real time;

[0023] Extract key information related to synthetic abnormal behavior data training from real-time monitoring data, including weight distribution, gradient information, training loss trend, and feature activation distribution;

[0024] After extracting key information, it is input into the monitoring window for analysis and processing, which includes feature vectorization, time series analysis, and anomaly detection;

[0025] Input the analyzed feature vector into a pre-trained machine learning model;

[0026] The pre-trained machine learning model generates a usage dependency coefficient based on the input feature vector, which is used to quantify the degree of dependence of the current training stage on the synthetic abnormal behavior data.

[0027] Preferably, within the monitoring window, the usage dependency coefficient generated when the machine learning model predicts the usage of the synthetic abnormal behavior data in the current training phase is compared and analyzed with a preset usage dependency coefficient reference threshold, and the usage of the synthetic abnormal behavior data is classified. The specific classification process is as follows:

[0028] If the usage dependency coefficient is greater than or equal to the usage dependency coefficient reference threshold, the usage of the synthetic abnormal behavior data in the monitoring window is classified as excessive dependency usage;

[0029] If the usage dependency coefficient is less than the usage dependency coefficient reference threshold, the usage of the synthetic abnormal behavior data in the monitoring window is classified as normal usage.

[0030] Preferably, an adaptive weighting algorithm and a sample reweighting algorithm are used to readjust the composition ratio of the training data set. The specific steps are as follows:

[0031] Initialize the adaptive weighting mechanism and assign initial weights to each type of data;

[0032] During the training process, the performance of the model on the synthetic abnormal behavior data and the actual abnormal behavior data is tracked in real time through the monitoring window, and the key indicators of the abnormal behavior detection model on each type of data are collected. If the monitoring results show that the loss of the abnormal behavior detection model on the synthetic abnormal behavior data is rapidly reduced, but the error or omission rate on the actual abnormal behavior data remains high, the abnormal behavior data is input into a feedback loop to analyze the contribution of each type of data to the training results of the abnormal behavior detection model and determine whether the data usage ratio needs to be adjusted;

[0033] According to the feedback data collected during the training of the abnormal behavior detection model, the weights of the synthetic abnormal behavior data and the actual abnormal behavior data are dynamically adjusted.

[0034] Preferably, after applying the adaptive weighting and sample reweighting strategies, the training process continues to run, and iterative optimization is performed based on the performance of the abnormal behavior detection model. After each round of training, the performance of the abnormal behavior detection model on the actual abnormal behavior data and the synthetic abnormal behavior data is evaluated again to check whether the abnormal behavior detection model has reduced its dependence on the synthetic abnormal behavior data and improved its detection capability for the actual abnormal behavior data. If the dependence problem is alleviated, the system gradually maintains the current weights without further adjustment. If the problem still exists, a new round of weight optimization and sample reweighting is continued.

[0035] Preferably, through the staged learning rate adjustment and weight freezing algorithm, a hierarchical training method is adopted to allow the abnormal behavior detection model to be preliminarily trained on the actual abnormal behavior data, and then fine-tuned on the synthetic abnormal behavior data. The specific steps are as follows:

[0036] Before training begins, the abnormal behavior detection model is initialized and a phased learning rate adjustment strategy is formulated;

[0037] In the first stage, the abnormal behavior detection model is trained on actual abnormal behavior data to enable it to learn key features in real scenarios. Once the abnormal behavior detection model reaches the preset performance on the actual abnormal behavior data, it will enter the next stage of fine-tuning.

[0038] When entering the second stage of training, the weight freezing technology is used, that is, most of the trained parameters in the abnormal behavior detection model are locked, and only some weights related to anomaly detection are released;

[0039] After fine-tuning is completed, a comprehensive performance evaluation of the abnormal behavior detection model is performed to ensure that the abnormal behavior detection model performs balanced on actual abnormal behavior data and synthetic abnormal behavior data.

[0040] Preferably, before training begins, the abnormal behavior detection model is initialized and a phased learning rate adjustment strategy is formulated. The specific steps are as follows:

[0041] First, a basic abnormal behavior detection model is selected and the initial parameters of the abnormal behavior detection model are set. On this basis, two stages of training are defined: the preliminary training stage and the fine-tuning stage. In each stage, different learning rates are used to control the convergence speed of the abnormal behavior detection model to ensure that the abnormal behavior detection model is gradually optimized during the training process and reduce the risk of overfitting.

[0042] Preferably, the usage of the synthetic abnormal behavior data in the current training phase is determined in real time based on the prediction results of the machine learning model. The specific steps are as follows:

[0043] During the training process, multiple usage dependency coefficients are generated in real time to establish an analysis set. The usage dependency coefficients in the analysis set are used to quantify the degree of dependence of the model on the synthetic abnormal behavior data.

[0044] In the constructed analysis set, the mean and standard deviation of several usage dependency coefficients are calculated. The mean of the usage dependency coefficient reflects the average degree of dependency of the abnormal behavior detection model in multiple training cycles, and the standard deviation of the usage dependency coefficient is used to evaluate the volatility of the usage dependency coefficient.

[0045] The mean usage dependency coefficient is compared with the preset usage dependency coefficient reference threshold to determine whether the abnormal behavior detection model exceeds the allowed degree of dependency. If the mean usage dependency coefficient is greater than or equal to the usage dependency coefficient reference threshold or the standard deviation of the usage dependency coefficient is greater than or equal to the standard deviation reference threshold, the training strategy is dynamically adjusted and the composition of the training data set is re-optimized until the mean usage dependency coefficient is less than the usage dependency coefficient reference threshold and the standard deviation of the usage dependency coefficient is less than the standard deviation reference threshold.

[0046] Balanced training data management system, including data set construction and division module, real-time data monitoring module, key information analysis and prediction module, usage classification module, training strategy maintenance module, data ratio adjustment module, hierarchical training module and real-time usage judgment module;

[0047] The dataset construction and division module builds a complete abnormal behavior dataset, which contains all the data used to train the abnormal behavior detection model. On this basis, the dataset is carefully divided into two parts: synthetic abnormal behavior data and actual abnormal behavior data.

[0048] The real-time data monitoring module obtains the training status of the synthetic abnormal behavior data in real time during the training of the abnormal behavior monitoring model, and extracts key information related to the training of the synthetic abnormal behavior data from the real-time monitoring data;

[0049] The key information analysis and prediction module analyzes and processes the extracted key information in the monitoring window, and inputs the analyzed and processed feature vectors into the pre-trained machine learning model, and uses the pre-trained machine learning model to predict the usage of the synthetic abnormal behavior data in the current training stage;

[0050] The usage classification module classifies the usage of the synthetic abnormal behavior data into normal usage and over-reliance usage according to the prediction results of the machine learning model;

[0051] The training strategy maintenance module, if the prediction result is normal use, continues to maintain the current training strategy and data usage. At this time, the abnormal behavior detection model is healthy in learning the synthetic abnormal behavior data and the actual abnormal behavior data, which can effectively improve the generalization ability and detection accuracy of the model;

[0052] The data ratio adjustment module readjusts the composition ratio of the training data set for situations that are judged to be over-reliant, increases the weight of actual abnormal behavior data, and reduces the proportion of abnormal behavior data, so as to encourage the abnormal behavior detection model to learn more about the characteristics of actual abnormal behavior;

[0053] Hierarchical training module: During the training process, a hierarchical training method is used to allow the abnormal behavior detection model to be initially trained on actual abnormal behavior data to learn the basic characteristics of real scenarios. Then, fine-tuning is performed on synthetic abnormal behavior data to enrich the abnormal behavior detection model's understanding of abnormal behavior and prevent the abnormal behavior detection model from being dominated by the characteristics of synthetic abnormal behavior data.

[0054] The real-time usage judgment module judges the usage of the synthetic abnormal behavior data in the current training phase in real time based on the prediction results of the machine learning model, until the abnormal behavior detection model no longer relies too much on the synthetic abnormal behavior data for training.

[0055] In the above technical solution, the technical effects and advantages provided by the present invention are:

[0056] The present invention monitors the degree of dependence of the abnormal behavior detection model on synthetic abnormal behavior data in real time, uses a pre-trained machine learning model to predict and optimize the training strategy, ensures that the abnormal behavior detection model does not overly rely on synthetic abnormal behavior data, and more fully learns the actual abnormal behavior characteristics to capture the diversity of real scenes. A hierarchical training method is adopted, first using actual abnormal behavior data for preliminary training, and then using synthetic abnormal behavior data for fine-tuning, to enhance the abnormal behavior detection model's understanding of abnormal behavior, improve its generalization ability and detection accuracy. It effectively reduces false alarms, reduces the risk of the system triggering an erroneous emergency response, saves emergency resources, and improves the system's response efficiency in real events, ensuring public safety and stability.

[0057] The present invention extracts key information related to the synthetic abnormal behavior data and uses the pre-trained machine learning model for prediction, dynamically adjusts the data ratio and training strategy, ensures that the abnormal behavior detection model learns in a healthy state, and prevents performance degradation. Real-time monitoring and optimization of the system make the abnormal behavior detection model training process transparent and controllable, correct potential problems in a timely manner, and ensure that the abnormal behavior detection model can adapt to data changes and new abnormal behaviors. Through iterative training and optimization, the robustness of the abnormal behavior detection model is improved, and it can run stably in complex environments, which not only improves the reliability and effectiveness of the system, but also provides support for future maintenance and upgrades. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0059] Figure 1 The figure is a flow chart of the method for managing balanced training data of the present invention.

[0060] Figure 2 It is a module schematic diagram of the balanced training data management system of the present invention. DETAILED DESCRIPTION

[0061] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as limited to the examples set forth herein; rather, these example embodiments are provided so that the description of the present disclosure will be more comprehensive and complete, and the concept of the example embodiments will be fully conveyed to those skilled in the art.

[0062] The present invention provides Figure 1 The balanced training data management method shown comprises the following steps:

[0063] Build a complete abnormal behavior dataset, which contains all the data used to train the abnormal behavior detection model. On this basis, divide the dataset into two parts: synthetic abnormal behavior data and actual abnormal behavior data.

[0064] The data set is divided into synthetic abnormal behavior data and actual abnormal behavior data in detail. The specific steps are as follows:

[0065] Collect abnormal behavior data from multiple sources to ensure the breadth and coverage of the data;

[0066] These data may come from actual surveillance videos, IoT sensors, historical records of smart monitoring platforms, etc. Actual abnormal behavior data may involve scenarios such as intrusion, theft, violent incidents, and abnormal operations. At the same time, data enhancement technology or generative models (such as GAN) are used to generate synthetic abnormal behavior data to make up for the lack of actual abnormal behavior data. After the collection is completed, all data are cleaned and standardized to remove noise data, incomplete samples, and duplicate data to ensure data consistency and reliability. The purpose of this step is to lay a high-quality data foundation for subsequent division and analysis.

[0067] Label the samples in the data set to identify which are synthetic abnormal behavior data and which are actual abnormal behavior data;

[0068] This can be done through the source information, metadata analysis, and timestamps of the data. For example, actual abnormal behavior data usually has a clear scene source (such as bank monitoring records), while synthetic abnormal behavior data will appear as samples from data augmentation or GAN generation. Data features can also be automatically analyzed, such as synthetic abnormal behavior data may have specific patterns or generate noise. After the labeling is completed, each sample is bound to its label metadata to provide a basis for subsequent classification and processing.

[0069] After the data is labeled, in-depth feature analysis is performed to evaluate the differences and quality of the characteristics of the synthetic abnormal behavior data and the actual abnormal behavior data;

[0070] Specifically, feature engineering can be used to extract key features of each sample, such as behavior type, time dimension, spatial position, and dynamic change information of video frames. By analyzing the similarities and differences between different features in synthetic and actual abnormal behavior data, it can be ensured that the deviation between synthetic abnormal behavior data and actual abnormal behavior data is controlled within a reasonable range. In addition, the quality, distribution, and representativeness of the data must be evaluated to ensure the diversity and authenticity of actual abnormal behavior data, as well as the coverage and rationality of synthetic abnormal behavior data. This step lays the foundation for subsequent training data combination.

[0071] After completing data annotation and analysis, the dataset is divided into synthetic abnormal behavior dataset and actual abnormal behavior dataset according to training requirements and data distribution;

[0072] A fixed ratio (such as 7:3) or a dynamic ratio based on data quality can be used to ensure that actual abnormal behavior data has sufficient weight to prevent the model from over-relying on synthetic abnormal behavior data. After the division is completed, the two parts of data are stored in different data warehouses or partitions, and a complete data management mechanism is established, such as data version control and metadata records. This management mechanism ensures that the model can flexibly call synthetic and actual abnormal behavior data during training, and supports real-time updating and adjustment of the division ratio, providing support for model optimization and strategy adjustment.

[0073] This division helps clarify the model's dependence on different types of data during training, providing a basis for subsequent monitoring and adjustment.

[0074] During the training of the abnormal behavior detection model, the training status of the synthetic abnormal behavior data is obtained in real time, and key information related to the training of the synthetic abnormal behavior data is extracted from the real-time monitored data;

[0075] In the monitoring window, the extracted key information is analyzed and processed, and the feature vector after analysis and processing is input into the pre-trained machine learning model, and the pre-trained machine learning model is used to predict the usage of the synthetic abnormal behavior data in the current training stage;

[0076] Use the pre-trained machine learning model to predict the usage of synthetic abnormal behavior data in the current training phase. The specific steps are as follows:

[0077] During the training process of the abnormal behavior detection model, indicators related to the training status of synthetic abnormal behavior data are obtained in real time;

[0078] The specific monitoring content includes: key performance indicators such as the change in the loss value of the model in each training cycle, accuracy, recall rate and precision. At the same time, it is also necessary to monitor the gradient update and training time of the model in different batches of training to ensure the transparency of the training process. By continuously tracking these indicators under the monitoring window, the system can promptly detect whether the model is overly focused on synthetic abnormal behavior data. The purpose of this step is to lay the foundation for subsequent key information extraction and dependency analysis.

[0079] Extract key information related to synthetic abnormal behavior data training from real-time monitoring data, including weight distribution, gradient information, training loss trend, and feature activation distribution;

[0080] Weight distribution: Analyze whether the anomaly detection model assigns abnormally high weights to certain features in the synthetic anomaly behavior data during training.

[0081] Gradient information: Extract the gradient changes of the abnormal behavior detection model when processing synthetic abnormal behavior data to observe whether the gradient explodes or disappears.

[0082] Training loss trend: Tracks whether the loss function of the anomaly detection model on synthetic anomaly data converges quickly, suggesting that the model may be biased towards synthetic anomaly data.

[0083] Feature activation distribution: Analyze the activation of the neural network when processing synthetic abnormal behavior data to determine whether the abnormal behavior detection model is highly dependent on certain specific features.

[0084] These key information provide the model's learning status of synthetic abnormal behavior data, ensuring that the system can accurately grasp the risk points in the training process.

[0085] After extracting key information, it is input into the monitoring window for analysis and processing, which includes feature vectorization, time series analysis, and anomaly detection;

[0086] Feature vectorization: Convert the extracted weights, gradients, loss trends and other data into vector format for input into subsequent machine learning models.

[0087] Time series analysis: Based on the time series data of different training batches, determine whether the learning state of the abnormal behavior detection model gradually tilts towards the synthetic abnormal behavior data over time.

[0088] Anomaly detection: Analyze whether the fluctuation of loss value and gradient exceeds the preset threshold, and promptly detect signs that the abnormal behavior detection model may be overly dependent on synthetic abnormal behavior data.

[0089] This process ensures that the system can accurately capture potential anomalies in the training process and lay the foundation for the machine learning model's predictions.

[0090] Input the processed feature vector into a pre-trained machine learning model (such as random forest, XGBoost, or neural network);

[0091] This machine learning model is specifically used to determine the model's reliance on synthetic abnormal behavior data. By analyzing key information in training, the machine learning model can predict the reliance at the current stage. This step enables the system to identify problems in the training process with data-driven intelligent means, rather than relying solely on manual experience.

[0092] The pre-trained machine learning model generates a usage dependency coefficient based on the input feature vector, which is used to quantify the degree of dependence of the current training stage on the synthetic abnormal behavior data;

[0093] After receiving the input feature vector, the pre-trained machine learning model (such as XGBoost, random forest or neural network) generates a comprehensive score by calculating the importance and contribution of each feature to the model decision based on the key information in the feature vector (such as weight distribution, gradient change, loss trend, feature activation pattern, etc.). These scores will go through a normalization process to map the results to an interval between 0 and 1, thereby forming a usage dependency coefficient. The model uses multiple algorithms such as feature importance analysis, time series trend and abnormal pattern matching to determine the degree of model dependence on synthetic abnormal behavior data. If the model detects that the proportion of synthetic abnormal behavior data in feature learning or model weight is too high, the usage dependency coefficient will be close to 1; if the model shows that the use of synthetic and actual abnormal behavior data is relatively balanced, the usage dependency coefficient will approach 0. The generation of the usage dependency coefficient can quantitatively reflect the dependency state of the model in the current training stage and provide an accurate basis for the dynamic adjustment of the training strategy.

[0094] A pre-trained machine learning model refers to a model that has been trained in advance and is specifically used to make predictions or judgments on new tasks or data. The key to this model is that it has completed preliminary training on a dataset that is highly relevant to the target task, and has learned how to extract key features from the input data and perform effective classification or regression analysis. In the task of abnormal behavior detection, pre-trained machine learning models (such as random forests, XGBoost, support vector machines, or deep neural networks) are designed to analyze specific feature vectors and predict the model's dependence on synthetic abnormal behavior data during training. By pre-training on historical training data and feature data, the model can quickly generate effective results when new data is input without the need for retraining, thereby accelerating the decision-making process and improving the efficiency of the system.

[0095] The advantage of this model is that it can use historical data and experience to learn to identify abnormal patterns and trends in model training in advance. For example, when monitoring the usage dependency of synthetic abnormal behavior data, the model will pay attention to details such as weight deviation, gradient anomalies, loss convergence rate, and compare this information with the learned patterns. If it is found that the current training state matches the typical "over-dependence" pattern, the model will give a higher usage dependency coefficient. This pre-training mechanism ensures the real-time monitoring capability of the system without having to wait until the training is completed to perform performance evaluation, thereby reducing potential risks and ensuring that the model training process is always within the controllable range.

[0096] When building pre-trained machine learning models, supervised learning methods are usually used. The model is initially trained on a large amount of historical data (such as past training records of the monitoring system) to learn how to identify the correlation and importance between different features. For example, when building a random forest model for dependency analysis, a large amount of training data with known labels (labeled as over-dependency or normal use) is first input. Through multiple training and optimization of decision trees, the model can learn how to make accurate predictions based on the input features. For the XGBoost model, a step-by-step boosting method is used to gradually reduce the prediction error so that the performance of the model is fully optimized during the training phase.

[0097] In specific applications, when the feature vector extracted by the real-time system is input into the pre-trained model, the model will quickly complete the calculation and generate a usage dependency coefficient. The usage dependency coefficient reflects the degree of dependence on the synthetic abnormal behavior data in the current model training. The value ranges from 0 to 1. A value close to 1 indicates a high degree of dependence, and a value close to 0 indicates a low degree of dependence. By using the usage dependency coefficient, the system can quantify and monitor potential problems in the training process, and take measures such as mixed sampling or strategy adjustment in a timely manner to prevent the abnormal behavior monitoring model from being unbalanced. The use of pre-trained models not only improves the efficiency of monitoring and prediction, but also significantly reduces resource consumption and enhances the system's automated processing capabilities and model robustness.

[0098] According to the prediction results of the machine learning model, the usage of synthetic abnormal behavior data is classified into normal use and over-reliance;

[0099] In the monitoring window, the usage dependency coefficient generated when the machine learning model predicts the usage of the synthetic abnormal behavior data in the current training phase is compared and analyzed with the pre-set usage dependency coefficient reference threshold, and the usage of the synthetic abnormal behavior data is classified. The specific classification process is as follows:

[0100] If the usage dependency coefficient is greater than or equal to the usage dependency coefficient reference threshold, the usage of the synthetic abnormal behavior data in the monitoring window is classified as excessive dependency usage;

[0101] If the usage dependency coefficient is less than the usage dependency coefficient reference threshold, the usage of the synthetic abnormal behavior data in the monitoring window is classified as normal usage.

[0102] Normal use specifically means that the abnormal behavior detection model's learning of synthetic abnormal behavior data and actual abnormal behavior data is balanced, and no adjustment is required during the training process.

[0103] Over-reliance on usage: The anomaly detection model pays too much attention to the synthetic anomaly data and may ignore the characteristics of the actual anomaly data.

[0104] If the prediction result is normal use, continue to maintain the current training strategy and data usage method. At this time, the abnormal behavior detection model is healthy in learning the synthetic abnormal behavior data and the actual abnormal behavior data, which can effectively improve the generalization ability and detection accuracy of the model.

[0105] Maintaining the existing strategy can ensure stable model training, avoid disturbances caused by unnecessary adjustments, and save resources and time costs.

[0106] For cases that are judged to be over-reliant, the composition ratio of the training data set is readjusted to increase the weight of the actual abnormal behavior data and reduce the proportion of abnormal behavior data, so as to encourage the abnormal behavior detection model to learn more about the characteristics of the actual abnormal behavior;

[0107] Adopt adaptive weighting algorithm and sample reweighting algorithm to readjust the composition ratio of training data set. The specific steps are as follows:

[0108] Initialize the adaptive weighting mechanism to assign initial weights to each type of data (synthetic abnormal behavior data and actual abnormal behavior data);

[0109] For example, the initial weight of the actual abnormal behavior data is set to W_actual = 0.5, and the initial weight of the synthetic abnormal behavior data is set to W_synthetic = 0.5. These weights are dynamically adjusted with feedback during the training process. In addition, a weight adjustment strategy needs to be defined, such as the performance of the loss function or the accuracy of the abnormal behavior detection model on the actual abnormal behavior data. Through this initialization, the system establishes a benchmark point for weight adjustment so that the influence of the actual abnormal behavior data can be gradually strengthened according to the learning of the abnormal behavior detection model.

[0110] During the training process, the monitoring window is used to track the performance of the model on synthetic abnormal behavior data and actual abnormal behavior data in real time, and key indicators such as loss, precision, and recall rate of the abnormal behavior monitoring model on each type of data are collected. If the monitoring results show that the loss of the abnormal behavior detection model on synthetic abnormal behavior data decreases rapidly, but the error or false negative rate on the actual abnormal behavior data remains high, it indicates that the abnormal behavior detection model may be overly dependent on the characteristics of synthetic abnormal behavior data. The system inputs these monitoring data into a feedback loop to analyze the contribution of each type of data to the training results of the abnormal behavior detection model and determine whether the proportion of data usage needs to be adjusted.

[0111] Dynamically adjust the weights of synthetic abnormal behavior data and actual abnormal behavior data based on the feedback data collected during the training of the abnormal behavior detection model;

[0112] For example, if the system detects that the abnormal behavior detection model has increased its reliance on synthetic abnormal behavior data, the weight W_actual of the actual abnormal behavior data will be increased (e.g., from 0.5 to 0.7), while the weight W_synthetic of the synthetic abnormal behavior data will be reduced (e.g., from 0.5 to 0.3). This adaptive weighting mechanism ensures that the abnormal behavior detection model gradually pays more attention to the characteristics of the actual abnormal behavior data. In addition, the system will reweight the samples according to the latest weight ratio, that is, dynamically select data in the training batch according to the adjusted weight ratio. This adjustment can be achieved by random sampling in each batch, so that the abnormal behavior detection model is always consistent with the latest weight distribution during training.

[0113] After applying the adaptive weighting and sample reweighting strategies, continue to run the training process and iteratively optimize according to the performance of the abnormal behavior detection model. After each round of training, the system re-evaluates the performance of the abnormal behavior detection model on the actual abnormal behavior data and synthetic abnormal behavior data to check whether the abnormal behavior detection model has reduced its dependence on synthetic abnormal behavior data and improved its detection ability on the actual abnormal behavior data. If the dependence problem is alleviated, the system will gradually maintain the current weight without further adjustment. If the problem still exists, continue with a new round of weight optimization and sample reweighting.

[0114] This iterative process ensures that the abnormal behavior detection model will not be unbalanced due to short-term adjustments, and gradually improves its ability to learn actual abnormal behavior characteristics during long-term training, ensuring the generalization and detection capabilities of the abnormal behavior detection model.

[0115] During the training process, a hierarchical training method is used to allow the abnormal behavior detection model to be initially trained on actual abnormal behavior data to learn the basic characteristics of real scenarios. Then, fine-tuning is performed on synthetic abnormal behavior data to enrich the abnormal behavior detection model's understanding of abnormal behavior and prevent the abnormal behavior detection model from being dominated by the characteristics of synthetic abnormal behavior data.

[0116] Through the phased learning rate adjustment and weight freezing algorithm, a layered training method is adopted to allow the abnormal behavior detection model to be initially trained on the actual abnormal behavior data, and then fine-tuned on the synthetic abnormal behavior data. The specific steps are as follows:

[0117] Before training begins, the abnormal behavior detection model is initialized and a phased learning rate adjustment strategy is formulated. First, an appropriate basic abnormal behavior detection model (such as a convolutional neural network or a Transformer abnormal behavior detection model) is selected and the initial parameters of the abnormal behavior detection model are set. On this basis, two stages of training are defined: the preliminary training stage (training on actual abnormal data) and the fine-tuning stage (fine-tuning on synthetic abnormal behavior data). In each stage, a different learning rate is used to control the convergence speed of the abnormal behavior detection model - a higher learning rate (such as 0.01 to 0.001) is used in the preliminary training stage to ensure that the abnormal behavior detection model can quickly learn the key features in the actual abnormal behavior data; the learning rate is reduced in the fine-tuning stage (such as 0.001 to 0.0001) to avoid the abnormal behavior detection model losing the basic features that have been learned during adjustment. This strategy ensures that the abnormal behavior detection model is gradually optimized during the training process, reduces the risk of overfitting, and improves the adaptability of the abnormal behavior detection model.

[0118] In the first stage, the abnormal behavior detection model is trained on actual abnormal behavior data. The goal of this stage is to enable the abnormal behavior detection model to learn key features in real scenarios, such as abnormal behavior patterns within a specific time, spatial position changes, or behavior trajectories in surveillance videos. Through a high learning rate, the abnormal behavior detection model can quickly capture the core patterns in the actual abnormal behavior data and achieve better optimization results on the loss function. At the same time, the system needs to monitor the performance of the abnormal behavior detection model in real time, such as the loss decline trend, precision, and recall rate, to ensure that the training process of the abnormal behavior detection model on the actual abnormal behavior data is stable. Once the abnormal behavior detection model reaches the preset performance on the actual abnormal behavior data (such as the accuracy and loss value of the validation set reach a certain threshold), it can enter the next stage of fine-tuning.

[0119] When entering the second stage of training, the weight freezing technology is used, that is, most of the trained parameters in the abnormal behavior detection model (such as the parameters of the convolutional layer or the Transformer encoder) are locked, and only some weights related to anomaly detection (such as the last few layers or the classifier layer) are released. At this time, the learning rate is reduced to a small value (such as 0.0001) to ensure that the parameters of the abnormal behavior detection model will not be greatly offset due to fine-tuning. In this stage, the abnormal behavior detection model is fine-tuned on the synthetic abnormal behavior data, and the diversity of the synthetic abnormal behavior data is further enriched to further enrich the understanding of new or rare abnormal behaviors. This training strategy ensures that the abnormal behavior detection model will not be dominated by the synthetic abnormal behavior data, but through the supplement of the synthetic abnormal behavior data, a more comprehensive grasp of various abnormal patterns.

[0120] After fine-tuning is completed, a comprehensive performance evaluation of the abnormal behavior detection model is required to ensure that it performs evenly on actual abnormal behavior data and synthetic abnormal behavior data. This includes evaluating the accuracy, recall rate, F1 score of the abnormal behavior detection model on the test set, as well as the abnormal behavior detection model's ability to identify different types of abnormal behaviors. If the evaluation results show that the abnormal behavior detection model performs well on synthetic abnormal behavior data, but false positives or negatives occur on actual abnormal behavior data, the abnormal behavior detection model needs to be optimized through further fine-tuning or adjustment of data weights. If the abnormal behavior detection model performs as expected, the current abnormal behavior detection model can be frozen and deployed in actual scenarios. In addition, the system can periodically retrain and fine-tune the abnormal behavior detection model to cope with new abnormal behaviors that appear in actual scenarios, ensuring the robustness and generalization of the abnormal behavior detection model in long-term use.

[0121] Through the above four steps, the hierarchical training method can ensure that the abnormal behavior detection model can not only learn the key features in the actual abnormal behavior data, but also expand the understanding of diverse abnormal behaviors from the synthetic abnormal behavior data. This combination of phased learning rate adjustment and weight freezing not only improves the detection ability and stability of the abnormal behavior detection model, but also effectively prevents the abnormal behavior detection model from over-reliance on a certain type of data, ensuring its long-term adaptability and reliability in complex environments.

[0122] Based on the prediction results of the machine learning model, the use of synthetic abnormal behavior data in the current training phase is judged in real time until the abnormal behavior detection model no longer overly relies on synthetic abnormal behavior data for training;

[0123] Based on the prediction results of the machine learning model, the usage of synthetic abnormal behavior data in the current training phase is judged in real time. The specific steps are as follows:

[0124] During the training process, multiple usage dependency coefficients are generated in real time to establish an analysis set. The usage dependency coefficients in the analysis set are used to quantify the degree of dependence of the model on the synthetic abnormal behavior data.

[0125] After each batch of training is completed, the usage dependency coefficient of the current training stage is calculated based on the prediction results of the pre-trained machine learning model (for example, close to 1 indicates high dependency, and close to 0 indicates low dependency). These usage dependency coefficients are collected into an analysis set and continuously updated. The construction of the analysis set ensures that the system can accurately grasp the data dependency trend of the abnormal behavior detection model during the training process based on the usage dependency coefficients of multiple rounds of training data. This data accumulation provides rich samples for subsequent analysis, avoiding misjudgments due to the contingency of a single training.

[0126] In the constructed analysis set, the mean and standard deviation of several usage dependency coefficients are calculated. The mean of the usage dependency coefficient reflects the average degree of dependence of the abnormal behavior detection model in multiple training cycles, which can indicate whether the abnormal behavior detection model as a whole is overly dependent on the synthetic abnormal behavior data; and the standard deviation of the usage dependency coefficient is used to evaluate the volatility of the usage dependency coefficient, indicating the stability of the abnormal behavior detection model's dependence on the synthetic abnormal behavior data in different training batches. If the standard deviation of the usage dependency coefficient is large, it means that the abnormal behavior detection model is too dependent on the synthetic abnormal behavior data at certain stages and is unstable, and the training strategy needs to be further adjusted; if the standard deviation of the usage dependency coefficient is small, it means that the abnormal behavior detection model has a relatively consistent dependence trend on the data. This step of calculation ensures that the system can accurately grasp the dependence status of the abnormal behavior detection model and discover potential problems through volatility analysis.

[0127] The mean of the usage dependency coefficient is compared with the preset usage dependency coefficient reference threshold to determine whether the abnormal behavior detection model exceeds the allowed degree of dependence. At the same time, the standard deviation of the usage dependency coefficient is compared with the standard deviation reference threshold to evaluate the stability of the model dependence. If the mean of the usage dependency coefficient is greater than or equal to the usage dependency coefficient reference threshold, or the standard deviation of the usage dependency coefficient is greater than or equal to the standard deviation reference threshold, it indicates that the abnormal behavior detection model still has excessive dependence on synthetic abnormal behavior data. The system will dynamically adjust the training strategy, such as reducing the proportion of synthetic abnormal behavior data, increasing the weight of actual abnormal behavior data, or switching to mixed sampling mode to re-optimize the composition of the training data set. The system will repeatedly perform this judgment and adjustment process until the mean of the usage dependency coefficient is less than the usage dependency coefficient reference threshold, and the standard deviation of the usage dependency coefficient is less than the standard deviation reference threshold, indicating that the abnormal behavior detection model no longer relies on synthetic abnormal behavior data and the training enters the ideal state. At this point, the abnormal behavior detection model has successfully achieved full learning of the actual abnormal behavior characteristics and has good robustness and generalization capabilities.

[0128] The present invention monitors the degree of dependence of the abnormal behavior detection model on the synthetic abnormal behavior data in real time, uses the pre-trained machine learning model to predict and adjust the training strategy, ensures that the abnormal behavior detection model does not over-rely on the synthetic abnormal behavior data, enables the model to more fully learn the characteristics of the actual abnormal behavior data, and captures the diversity and complexity in the real scene. By increasing the weight of the actual abnormal behavior data, the model's understanding of the real abnormal behavior is strengthened during the training process, avoiding the deviation caused by over-reliance on the synthetic abnormal behavior data. In the training, a hierarchical training method is adopted, firstly performing preliminary training on the actual abnormal behavior data, and then fine-tuning on the synthetic abnormal behavior data, further enriching the model's understanding of abnormal behavior. This strategy improves the generalization ability and detection accuracy of the model, and effectively reduces the risk of misjudging normal behavior as abnormal behavior. Reducing false alarms not only reduces the possibility of the system triggering an erroneous emergency response, saves emergency resources, but also improves the response efficiency of the system when a real event occurs, maintaining public safety and social stability.

[0129] The present invention extracts key information related to synthetic abnormal behavior data during the training process, inputs it into the pre-trained machine learning model for analysis and prediction, and the system can determine the degree of dependence of the model on the synthetic abnormal behavior data in real time. According to the prediction results, the composition ratio and training strategy of the training data set are dynamically adjusted to ensure that the model is always learning in a healthy state. This dynamic monitoring and adjustment mechanism makes the model training process more transparent and controllable, and timely discovers and corrects potential problems to prevent model performance from degrading. Continuous monitoring and optimization enable the abnormal behavior detection model to quickly adapt and maintain a high level of detection performance when facing changes in data distribution or new abnormal behaviors. Through iterative training and optimization, the robustness of the model is enhanced, and it can operate stably in complex and changeable actual environments. This not only improves the reliability and effectiveness of the system, but also provides strong support for long-term model maintenance and upgrades, ensuring that the system continues to play a role in future development.

[0130] The present invention provides Figure 2 The balanced training data management system shown includes a data set construction and division module, a real-time data monitoring module, a key information analysis and prediction module, a usage classification module, a training strategy maintenance module, a data ratio adjustment module, a hierarchical training module, and a real-time usage judgment module;

[0131] The dataset construction and division module builds a complete abnormal behavior dataset, which contains all the data used to train the abnormal behavior detection model. On this basis, the dataset is carefully divided into two parts: synthetic abnormal behavior data and actual abnormal behavior data.

[0132] The real-time data monitoring module obtains the training status of the synthetic abnormal behavior data in real time during the training of the abnormal behavior monitoring model, and extracts key information related to the training of the synthetic abnormal behavior data from the real-time monitoring data;

[0133] The key information analysis and prediction module analyzes and processes the extracted key information in the monitoring window, and inputs the analyzed and processed feature vectors into the pre-trained machine learning model, and uses the pre-trained machine learning model to predict the usage of the synthetic abnormal behavior data in the current training stage;

[0134] The usage classification module classifies the usage of the synthetic abnormal behavior data into normal usage and over-reliance usage according to the prediction results of the machine learning model;

[0135] The training strategy maintenance module, if the prediction result is normal use, continues to maintain the current training strategy and data usage. At this time, the abnormal behavior detection model is healthy in learning the synthetic abnormal behavior data and the actual abnormal behavior data, which can effectively improve the generalization ability and detection accuracy of the model;

[0136] The data ratio adjustment module readjusts the composition ratio of the training data set for situations that are judged to be over-reliant, increases the weight of actual abnormal behavior data, and reduces the proportion of abnormal behavior data, so as to encourage the abnormal behavior detection model to learn more about the characteristics of actual abnormal behavior;

[0137] Hierarchical training module: During the training process, a hierarchical training method is used to allow the abnormal behavior detection model to be initially trained on actual abnormal behavior data to learn the basic characteristics of real scenarios. Then, fine-tuning is performed on synthetic abnormal behavior data to enrich the abnormal behavior detection model's understanding of abnormal behavior and prevent the abnormal behavior detection model from being dominated by the characteristics of synthetic abnormal behavior data.

[0138] A real-time usage judgment module judges the usage of synthetic abnormal behavior data in the current training phase in real time based on the prediction results of the machine learning model, until the abnormal behavior detection model no longer overly relies on synthetic abnormal behavior data for training;

[0139] The balanced training data management method provided in the embodiment of the present invention is implemented by the above-mentioned balanced training data management system. The specific methods and processes of the balanced training data management system are detailed in the embodiment of the above-mentioned balanced training data management method, which will not be repeated here.

[0140] The above description is only by way of illustration of certain exemplary embodiments of the present invention. It is undoubted that those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A balanced training data management method, characterized in that: The following steps are involved: Build a complete abnormal behavior dataset, which contains all the data used to train the abnormal behavior detection model. On this basis, divide the dataset into two parts: synthetic abnormal behavior data and actual abnormal behavior data. During the training of the abnormal behavior detection model, the training status of the synthetic abnormal behavior data is obtained in real time, and key information related to the training of the synthetic abnormal behavior data is extracted from the real-time monitored data; In the monitoring window, the extracted key information is analyzed and processed, and the feature vector after analysis and processing is input into the pre-trained machine learning model, and the pre-trained machine learning model is used to predict the usage of the synthetic abnormal behavior data in the current training stage; According to the prediction results of the machine learning model, the usage of synthetic abnormal behavior data is classified into normal use and over-reliance; If the prediction result is normal use, continue to maintain the current training strategy and data usage method. At this time, the abnormal behavior detection model is healthy in learning the synthetic abnormal behavior data and the actual abnormal behavior data, which can effectively improve the generalization ability and detection accuracy of the model. For cases that are judged to be over-reliant, the composition ratio of the training data set is readjusted to increase the weight of the actual abnormal behavior data and reduce the proportion of abnormal behavior data, so as to encourage the abnormal behavior detection model to learn more about the characteristics of the actual abnormal behavior; During the training process, a hierarchical training method is used to allow the abnormal behavior detection model to be initially trained on actual abnormal behavior data to learn the characteristics of real scenarios. Then, fine-tuning is performed on synthetic abnormal behavior data to enrich the abnormal behavior detection model's understanding of abnormal behavior and prevent the abnormal behavior detection model from being dominated by the characteristics of synthetic abnormal behavior data. Based on the prediction results of the machine learning model, the usage of synthetic abnormal behavior data in the current training phase is judged in real time until the abnormal behavior detection model no longer relies too much on synthetic abnormal behavior data for training.

2. The method for managing balanced training data according to claim 1, characterized in that: The data set is divided into synthetic abnormal behavior data and actual abnormal behavior data in detail. The specific steps are as follows: Collect abnormal behavior data from multiple sources to ensure the breadth and coverage of the data; Label the samples in the data set to identify which are synthetic abnormal behavior data and which are actual abnormal behavior data; After the data is labeled, in-depth feature analysis is performed to evaluate the differences and quality of the characteristics of the synthetic abnormal behavior data and the actual abnormal behavior data; After completing data annotation and analysis, the dataset is divided into synthetic abnormal behavior dataset and actual abnormal behavior dataset according to training requirements and data distribution.

3. The method for managing balanced training data according to claim 1, characterized in that: Use the pre-trained machine learning model to predict the usage of synthetic abnormal behavior data in the current training phase. The specific steps are as follows: During the training process of the abnormal behavior detection model, indicators related to the training status of synthetic abnormal behavior data are obtained in real time; Extract key information related to synthetic abnormal behavior data training from real-time monitoring data, including weight distribution, gradient information, training loss trend, and feature activation distribution; After extracting key information, it is input into the monitoring window for analysis and processing, which includes feature vectorization, time series analysis, and anomaly detection; Input the analyzed feature vector into a pre-trained machine learning model; The pre-trained machine learning model generates a usage dependency coefficient based on the input feature vector, which is used to quantify the degree of dependence of the current training stage on the synthetic abnormal behavior data.

4. The method for managing balanced training data according to claim 3, characterized in that: In the monitoring window, the usage dependency coefficient generated when the machine learning model predicts the usage of the synthetic abnormal behavior data in the current training phase is compared and analyzed with the pre-set usage dependency coefficient reference threshold, and the usage of the synthetic abnormal behavior data is classified. The specific classification process is as follows: If the usage dependency coefficient is greater than or equal to the usage dependency coefficient reference threshold, the usage of the synthetic abnormal behavior data in the monitoring window is classified as excessive dependency usage; If the usage dependency coefficient is less than the usage dependency coefficient reference threshold, the usage of the synthetic abnormal behavior data in the monitoring window is classified as normal usage.

5. The method for managing balanced training data according to claim 1, characterized in that: Adopt adaptive weighting algorithm and sample reweighting algorithm to readjust the composition ratio of training data set. The specific steps are as follows: Initialize the adaptive weighting mechanism and assign initial weights to each type of data; During the training process, the performance of the model on the synthetic abnormal behavior data and the actual abnormal behavior data is tracked in real time through the monitoring window, and the key indicators of the abnormal behavior detection model on each type of data are collected. If the monitoring results show that the loss of the abnormal behavior detection model on the synthetic abnormal behavior data is rapidly reduced, but the error or omission rate on the actual abnormal behavior data remains high, the abnormal behavior data is input into a feedback loop to analyze the contribution of each type of data to the training results of the abnormal behavior detection model and determine whether the data usage ratio needs to be adjusted; According to the feedback data collected during the training of the abnormal behavior detection model, the weights of the synthetic abnormal behavior data and the actual abnormal behavior data are dynamically adjusted.

6. The method for managing balanced training data according to claim 5, characterized in that: After applying the adaptive weighting and sample reweighting strategies, continue to run the training process and perform iterative optimization based on the performance of the abnormal behavior detection model. After each round of training, evaluate the performance of the abnormal behavior detection model on the actual abnormal behavior data and synthetic abnormal behavior data again to check whether the abnormal behavior detection model has reduced its dependence on synthetic abnormal behavior data and improved its detection capability for actual abnormal behavior data. If the dependency problem is alleviated, the system gradually maintains the current weights without further adjustment. If the problem still exists, continue with a new round of weight optimization and sample reweighting.

7. The method for managing balanced training data according to claim 1, characterized in that: Through the phased learning rate adjustment and weight freezing algorithm, a layered training method is adopted to allow the abnormal behavior detection model to be initially trained on the actual abnormal behavior data, and then fine-tuned on the synthetic abnormal behavior data. The specific steps are as follows: Before training begins, the abnormal behavior detection model is initialized and a phased learning rate adjustment strategy is formulated; In the first stage, the abnormal behavior detection model is trained on actual abnormal behavior data to enable it to learn key features in real scenarios. Once the abnormal behavior detection model reaches the preset performance on the actual abnormal behavior data, it will enter the next stage of fine-tuning. When entering the second stage of training, the weight freezing technology is used, that is, most of the trained parameters in the abnormal behavior detection model are locked, and only some weights related to anomaly detection are released; After fine-tuning is completed, a comprehensive performance evaluation of the abnormal behavior detection model is performed to ensure that the abnormal behavior detection model performs balanced on actual abnormal behavior data and synthetic abnormal behavior data.

8. The method for managing balanced training data according to claim 7, characterized in that: Before training begins, initialize the abnormal behavior detection model and formulate a phased learning rate adjustment strategy. The specific steps are as follows: First, a basic abnormal behavior detection model is selected and the initial parameters of the abnormal behavior detection model are set. On this basis, two stages of training are defined: the preliminary training stage and the fine-tuning stage. In each stage, different learning rates are used to control the convergence speed of the abnormal behavior detection model to ensure that the abnormal behavior detection model is gradually optimized during the training process and reduce the risk of overfitting.

9. The method for managing balanced training data according to claim 4, characterized in that: Based on the prediction results of the machine learning model, the usage of synthetic abnormal behavior data in the current training phase is judged in real time. The specific steps are as follows: During the training process, multiple usage dependency coefficients are generated in real time to establish an analysis set. The usage dependency coefficients in the analysis set are used to quantify the degree of dependence of the model on the synthetic abnormal behavior data. In the constructed analysis set, the mean and standard deviation of several usage dependency coefficients are calculated. The mean of the usage dependency coefficient reflects the average degree of dependency of the abnormal behavior detection model in multiple training cycles, and the standard deviation of the usage dependency coefficient is used to evaluate the volatility of the usage dependency coefficient. The mean usage dependency coefficient is compared with the preset usage dependency coefficient reference threshold to determine whether the abnormal behavior detection model exceeds the allowed degree of dependency. If the mean usage dependency coefficient is greater than or equal to the usage dependency coefficient reference threshold or the standard deviation of the usage dependency coefficient is greater than or equal to the standard deviation reference threshold, the training strategy is dynamically adjusted and the composition of the training data set is re-optimized until the mean usage dependency coefficient is less than the usage dependency coefficient reference threshold and the standard deviation of the usage dependency coefficient is less than the standard deviation reference threshold.

10. A balanced training data management system, used to implement the balanced training data management method according to any one of claims 1 to 9, characterized in that: It includes data set construction and division module, real-time data monitoring module, key information analysis and prediction module, usage classification module, training strategy maintenance module, data ratio adjustment module, hierarchical training module and real-time usage judgment module; The dataset construction and division module builds a complete abnormal behavior dataset, which contains all the data used to train the abnormal behavior detection model. On this basis, the dataset is carefully divided into two parts: synthetic abnormal behavior data and actual abnormal behavior data. The real-time data monitoring module obtains the training status of the synthetic abnormal behavior data in real time during the training of the abnormal behavior monitoring model, and extracts key information related to the training of the synthetic abnormal behavior data from the real-time monitoring data; The key information analysis and prediction module analyzes and processes the extracted key information in the monitoring window, inputs the analyzed and processed feature vectors into the pre-trained machine learning model, and uses the pre-trained machine learning model to predict the usage of the synthetic abnormal behavior data in the current training phase; The usage classification module classifies the usage of the synthetic abnormal behavior data into normal usage and over-reliance usage according to the prediction results of the machine learning model; The training strategy maintenance module, if the prediction result is normal use, continues to maintain the current training strategy and data usage. At this time, the abnormal behavior detection model is healthy in learning the synthetic abnormal behavior data and the actual abnormal behavior data, which can effectively improve the generalization ability and detection accuracy of the model; The data ratio adjustment module readjusts the composition ratio of the training data set for situations that are judged to be over-reliant, increases the weight of actual abnormal behavior data, reduces the proportion of abnormal behavior data, and prompts the abnormal behavior detection model to learn more about the characteristics of actual abnormal behavior; Hierarchical training module: During the training process, a hierarchical training method is used to allow the abnormal behavior detection model to perform preliminary training on actual abnormal behavior data and learn the basic characteristics of real scenarios; Then, fine-tune the model on the synthetic abnormal behavior data to enrich the abnormal behavior detection model’s understanding of abnormal behavior and prevent the abnormal behavior detection model from being dominated by the characteristics of the synthetic abnormal behavior data. The real-time usage judgment module judges the usage of the synthetic abnormal behavior data in the current training phase in real time based on the prediction results of the machine learning model, until the abnormal behavior detection model no longer overly relies on the synthetic abnormal behavior data for training.

Images