69 results about "Insufficient Sample" patented technology

The invention discloses an activated carbon forming defect online detection method and device, and belongs to the technical field of material defect detection.The method comprises the steps that technological parameters are collected in real time and preprocessed, a parameter sequence is generated, self-sensing recognition of a new technology is conducted, lagged starting of a quality control vacuum period is avoided, and once the new technology is recognized, the quality control vacuum period is started. An unknown defect pre-recognition mechanism is triggered immediately, novel anomalies which do not conform to known defects are recognized, and an unknown defect candidate set is generated; constructing a multi-stage hierarchical detection network, performing adaptive detection on the unknown defect candidate set, establishing a reference model, screening samples, updating the reference model, and forming a defect identification model adaptive to a new process, so that the influence of insufficient samples on the detection accuracy is effectively relieved; and setting three-dimensional quantitative indexes, judging a transition period, automatically switching to a full-model detection mode when the indexes meet preset requirements, and objectively evaluating the recognition capability of the detection system on new process defects and the process suitability through the quantitative indexes.

The invention relates to a small-sample contrast enhancement fine tuning method and system based on a large language model, which are used for identifying named entities in recruitment texts. The method comprises the steps of performing cleaning and format conversion on an original recruitment text, and generating an input sample conforming to a natural language instruction format; under the condition that the labeled samples are insufficient, positive and negative sample pairs are constructed to enhance the recognition capability of the model on entity categories and boundaries; carrying out low-rank parameter updating on the pre-trained large language model by adopting a LoRA fine tuning technology, and reducing computing resource consumption in combination with 4-bit quantitative training; in a pre-training large language model reasoning process, through a multi-dimensional joint confidence evaluation mechanism, confidence of four dimensions of entity levels, lengths, types and contexts is synthesized, and low-confidence identification results are filtered after dynamic weighted normalization processing. The method is suitable for recruitment recommendation, talent matching and other downstream tasks, and has the advantages of high recognition accuracy, low training cost, high system robustness and the like.

The invention relates to a load resource data anomaly detection system based on a novel power system, which comprises a historical data management module, a real scene management module and a simulation generation module, the historical data management module is used for marking historical load data in a historical data load information base to generate a historical load sample. The whole system effectively breaks through the technical bottlenecks of insufficient samples and weak model generalization ability through multi-dimensional sample generation, dynamic compensation correction and adversarial training, and provides an efficient and reliable solution for load anomaly detection of a novel power system.

The invention provides a part defect detection method and system based on yov5, and the method comprises the following steps: collecting a defect photo set of a defective part, and obtaining a target defect region image; marking the target defect area image to obtain target defect data; training the initial yolov5 model based on the target defect data, obtaining a yolov5 training model and evaluating the performance of the model until a preset convergence condition is reached, and obtaining a yolov5 model; and inputting the defect picture of the defect part to be detected into the yolov5 model to obtain a defect detection result. According to the part defect detection method and system based on the yolov5 provided by the invention, the problems of insufficient sample quantity, low pertinence, low marking quality and the like are avoided, the rapid classification of different part defect types and sizes is realized, and the occurrence of false detection and missing detection conditions is effectively prevented.

The invention provides a negative sample enhanced APT attack detection method based on graph structure learning, and mainly solves the problem of poor detection effect caused by insufficient data set compatibility, side information utilization and data set negative samples in the prior art. According to the scheme, the method comprises the following steps: 1) acquiring a heterogeneous data set, and constructing a visual traceability graph; 2) preprocessing the traceability graph, dividing the traceability graph into snapshots according to timestamps, and constructing a training set and a test set; 3) designing an encoder and a decoder of the drawing neural network, taking the time window, the node features, the adjacency matrix and the edge features as encoder input, and generating a reconstructed adjacency matrix through the decoder; 4) designing a loss function, and performing RNN training on the graph neural network; and 5) inputting to-be-detected data into the trained model, and identifying abnormal attack traffic to complete detection. According to the method, the accuracy and robustness of APT attack detection can be effectively improved, and the method can be used for development and deployment of APT attack detection and defense systems in the field of network security.

The present application relates to the field of data processing technology, and in particular to a method, system, and device for optimizing a copper foil preparation process based on machine learning, comprising collecting historical preparation data, historical quality data, and probability values ​​of all copper foil preparation equipment as first sample data, and historical real-time preparation data and historical real-time quality data of a target object as second sample data, respectively training a first model and a second model, periodically collecting the real-time preparation data of the target object, inputting the first model to obtain a first prediction result and probability value, inputting the second model to obtain a second prediction result, and then combining the third model to obtain a target prediction result. When the standard deviation between the target prediction result and the copper foil preparation is greater than a preset value, adjusting the preparation data to obtain the optimal preparation parameters. The present invention solves the problem of low precision of process optimization parameters due to insufficient sample data when the copper foil preparation equipment is not put into use for a long enough time, thereby achieving the effect of improving the precision of process optimization parameters during the production process.

The invention discloses a composite material reflectivity spectral information classification method and system based on a PCA-SVM algorithm. The method comprises the following steps: collecting reflectivity spectral data of a composite material sample; preprocessing data to eliminate measurement deviation and unify numerical scale; carrying out dimensionality reduction on the preprocessed high-dimensional spectral data through principal component analysis, and extracting feature components retaining main variance information; based on dimension reduction features, a support vector machine is adopted to construct a multi-label classification model according to a'one-to-other 'strategy; predicting the test sample, and generating a multi-label classification result through probability output and threshold processing; and analyzing the classification performance by using the multi-label evaluation index. The data processing module of the system executes preprocessing, dimension reduction, modeling, prediction and evaluation operations. The method is suitable for lossless identification of multi-component composite samples, both interpretability and identification precision are considered, the performance bottleneck of a traditional method under the conditions of feature overlapping, insufficient samples and the like is effectively overcome, and efficient and accurate classification of the composite materials is achieved.

The invention discloses a hydroelectric generating set fault diagnosis data enhancement method based on a diffusion model and generative adversarial training, and belongs to the field of industrial equipment fault diagnosis. According to the method, the data enhancement model combining the diffusion model and the generative adversarial network is constructed, a multi-modal non-Gaussian distribution mechanism and a latent variable control generation process are introduced, the limitation of a single Gaussian hypothesis of a traditional diffusion model is broken through, and richer and more real sample generation is realized. And a condition discriminator and a time sequence modeling mechanism are introduced, so that the training of the model under different noise levels is more stable, and the authenticity judgment capability of the discriminator on the generated sample is enhanced. The problems that an existing hydroelectric generating set fault diagnosis system is insufficient in sample, low in generated sample quality and unreal in sample distribution are solved, and the diversity and quality of data are remarkably improved.

The invention relates to the technical field of data processing, in particular to a copper foil preparation process optimization method, system and equipment based on machine learning, and the method comprises the steps: collecting historical preparation data, historical quality data and probability values of all copper foil preparation equipment as first sample data; taking historical real-time preparation data and historical real-time quality data of the target object as second sample data, respectively training the first model and the second model, periodically collecting the real-time preparation data of the target object, inputting the real-time preparation data into the first model to obtain a first prediction result and a probability value, and inputting the real-time preparation data into the second model to obtain a second prediction result. And a target prediction result is obtained in combination with the third model, and when the target prediction result and the copper foil preparation standard difference value are larger than a preset value, the preparation data are adjusted to obtain optimal preparation parameters. According to the method, the problem that the process optimization parameter precision is low due to insufficient sample data when the investment time of copper foil preparation equipment is not long enough is solved, and the effect of improving the process optimization parameter precision in the production process is achieved.

The invention discloses a three-phase inverter open-circuit fault diagnosis method based on deep transfer learning. The three-phase inverter open-circuit fault diagnosis method comprises the following steps: step 100, collecting a sample set 1 with a label in a source domain; collecting a sample set 2 with labels in the target domain and a sample set 3 without labels in the target domain; step 200, constructing a convolutional neural network model for fault diagnosis, and training the model by using a sample set 1 of a source domain to obtain a pre-training model alpha; step 300, constructing a fault diagnosis model based on transfer learning on the basis of the pre-training model alpha, and training the model by using a sample set 2 of a target domain to obtain a transfer model beta; step 400, using the migration model beta to predict unlabeled samples in the unlabeled sample set 3 to generate a pseudo label, and using a consistency regularization technology to smooth a pseudo label prediction result for subsequent training of the migration model beta to obtain a final optimization model gamma; and step 500, inputting to-be-detected three-phase inverter data into the model gamma, and outputting a predicted health state. Through the method, the problem of insufficient target domain samples is solved, and the diagnosis accuracy is improved.

The invention discloses a rolling bearing small sample cross-working-condition fault diagnosis method and system, and relates to the technical field of rolling bearing fault diagnosis, and the method comprises the steps: segmenting a rolling bearing vibration signal; extracting time domain and frequency domain data from the subsamples; performing feature extraction on the time-domain data and the frequency-domain data, and performing cross connection on the time-domain data and the frequency-domain data based on features extracted by the time-domain data and the frequency-domain data to obtain to-be-classified features; adopting the meta-learning model after meta-training to the to-be-classified features to obtain a bearing fault classification result; wherein the meta-training process comprises the steps that model parameters are trained among different working condition tasks through inner circulation and outer circulation, a dynamic attenuation learning rate strategy is adopted in the inner circulation, and the learning rate is attenuated according to a preset strategy along with the increase of training adaptation steps. Few vibration signal samples are effectively utilized for feature extraction, the precision of classification results is improved, and the problems that early samples are insufficient and difficult to diagnose when rotating machinery breaks down in industrial production are solved.

The invention relates to a load model parameter identification method and system based on sample enhancement and reinforcement learning and a medium, and provides the load model parameter identification method based on sample enhancement and reinforcement learning for the problems in the prior art, and the method comprises the following steps: preprocessing transient data of a power system, and obtaining a preprocessed sample; inputting the preprocessed sample into a generative adversarial network for adversarial training to generate a new sample set; constructing a load parameter identification model based on a TD3 reinforcement learning algorithm, and importing the new sample set into the load parameter identification model for training to obtain a trained data model; and inputting online real-time data in the power system into the trained data model to output model parameters, and taking an output result as a final result of parameter identification. According to the invention, data enhancement is carried out by using the GAN, high-quality load data is generated, and the problem of insufficient samples is relieved. And the parameter identification process is optimized through reinforcement learning, so that the model can be dynamically adjusted, local optimum is avoided, and the identification precision is improved.

This application relates to the field of equipment performance testing and evaluation technology, and provides a method and system for equivalent reduction of equipment test sample size. By constructing a data fusion evaluation model based on actual and numerical test data, the fidelity of actual test response estimation under small sample conditions is improved. An equivalent measure is designed and calculated by fusing the mean square error of the evaluation model from a single data source with a preset test confidence level, and a box plot of the equivalent measure from a single data source is obtained. By connecting the lower quartile and upper quartile of each obtained box plot, equivalent composite confidence intervals for actual and numerical simulation tests are plotted. Based on these equivalent composite confidence intervals, the theoretically required actual test sample size is equivalently reduced to the numerical simulation test sample size, achieving an equivalent reduction of the actual test sample size by the numerical simulation test, thus solving the problem of equipment performance evaluation under insufficient actual test samples.

The invention discloses a small sample encrypted traffic classification method and system, and belongs to the technical field of network security. In order to solve the problem that accurate classification is difficult under the condition that encrypted traffic samples are insufficient, a flow representation mode based on a time sequence convolutional neural network is mainly adopted, a supervised and unsupervised fused flow difference enhancement mechanism is combined, and multi-task cooperation meta-learning is introduced to optimize model parameters, so that the classification accuracy is improved. The generalization ability of the model under the small sample condition is improved. According to the method, high-precision classification of the encrypted traffic under the condition of a small number of labeled samples can be realized, and the method is suitable for cross-version and cross-domain application scenes.

The application discloses a semi-supervised network flow anomaly detection method and device based on a stacked autoencoder, and the method comprises the following steps: selecting a training data set and randomly sampling two-by-two to form a sample pair set; training a two-stage stacked autoencoder by using the training data set to obtain a feature extractor; extracting and fusing the features of the sample pairs in the sample pair set to train an anomaly score generator; randomly extracting samples from the training data set and the data to be detected to form a sample pair to obtain an anomaly score; and determining whether the data is abnormal according to an anomaly reference value and the anomaly score. The application can effectively reduce the influence of insufficient labeled samples on the model performance and generalization ability, and improve the precision of the model in network flow data anomaly detection by randomly combining labeled samples and unlabeled samples, fully utilizing the existing labeled information to expand the number of labeled samples for model training in the case of insufficient labeled data.

The invention belongs to the technical field of artificial intelligence and network predictive maintenance, relates to a fault prediction method based on a multi-source migration echo state network, and aims to solve the problem of low prediction precision caused by insufficient early samples of a fault, and the method comprises the following steps: firstly, extracting migration knowledge from a plurality of historical fault source domains by using the echo state network; secondly, measuring the similarity between the source domains and the target domain and the difference between the source domains by adopting a dynamic time warping distance, and selecting a plurality of similar source domains with complementary information; then, a plurality of prediction sub-models are established in the target domain in combination with migration knowledge of the selected source domain, and a final prediction model is obtained through integrated learning; and finally, performing future value prediction on the key variables acquired in real time by using the model. According to the method, multi-source historical information can be fully utilized under the small sample condition, the fault prediction precision and generalization ability are remarkably improved, and the method is suitable for fault prediction scenes with multi-source time series data in industrial processes, mechanical equipment, power systems and the like.

The invention discloses an atmospheric numerical mode correction method based on a sample increasing technology, which comprises the following steps: generating a sample generator by using a deep learning model, generating training set samples by using the sample generator, increasing the number of the training set samples, and avoiding an over-fitting phenomenon caused by insufficient sample size when the model is trained.

The application discloses a high-speed moving target recognition method based on data enhancement and belongs to the field of target recognition. The application introduces a high-speed moving target background data set, the confidence of a pseudo target obtained by a classification network after the pseudo target generated by a differentiable generative adversarial network, and the target instances with a confidence greater than a set threshold and the target instances in the initial data set are segmented objects and are enhanced together in the high-speed moving target background data set, solving the problems of insufficient samples and inconsistency between the high-speed moving target training background and the working environment. When the pseudo target generated by the differentiable generative adversarial network is trained, the confidence of the pseudo target and the CIoU value of the positive sample boundary box and the real box are weighted and summed to form a new confidence loss function in the YOLOv7 target detection algorithm, and the improved loss function can more accurately measure the authenticity of the target. The method can realize accurate recognition of a specific type of high-speed moving target under small samples.

The invention discloses an MEMS inertial sensor reliability analysis method, which comprises the following steps: step 1, failure mode and mechanism analysis: carding environmental factors influencing the work of an MEMS inertial sensor, identifying failure modes caused by the environmental factors, and determining internal action mechanisms corresponding to the failure modes; 2, selecting a representative sample, and formulating a sample selection rule according to an application scene, a production batch, a structure type and performance parameter distribution of the sensor; the method has the beneficial effects that the system carding the association relationship between various environmental factors and failure modes, defining the internal mechanism of failure, breaking through the limitation that the existing method only pays attention to a single factor, enabling the analysis result to be more fit with the actual working scene of the sensor, and remarkably improving the comprehensiveness of reliability analysis; representative samples are selected based on scientific sample selection rules, key variables such as production batches, structure types and performance parameters are covered, and the problem of insufficient sample representativeness of an existing method is solved.

This application belongs to a clustering recognition method. Aiming at the technical problems existing in the existing power market collusion behavior recognition methods, such as insufficient reflection of the characteristics of collusion behavior, lack of the ability to deeply mine complex index characteristics, insufficient training samples, and insufficient abnormal collusion sample recognition ability, a power market collusion behavior recognition method and related devices are provided. First, obtain the data corresponding to the discrimination indicators of the power market collusion behavior; the discrimination indicators include electricity price indicators, electricity quantity indicators, and quantity-price relationship indicators. The electricity price indicators include the average value of the declared price fluctuations, and the quantity-price relationship indicators include the similarity of the bid curves. Then, input the data corresponding to the discrimination indicators into the collusion recognition model to obtain the collusion behavior recognition result; the input of the collusion recognition model is the interaction feature data between two market entities. It can analyze the interaction patterns between market entities in different clusters and does not need to rely on collusion label data.

The invention discloses a semi-supervised network traffic anomaly detection method and device based on a stacked auto-encoder, and the method comprises the steps: selecting a training data set, and randomly sampling every two training data sets to form a sample pair, thereby obtaining a sample pair set; training a secondary stacked auto-encoder by adopting the training data set to obtain a feature extractor; extracting and fusing the features of the sample pairs in the sample pair set, and training to obtain an abnormal scoring device; randomly extracting samples from the training data set to form a sample pair with the to-be-detected data, and obtaining an abnormal score; and judging whether the data is abnormal according to the abnormal reference value and the abnormal score. According to the method, the labeled samples and the unlabeled samples are randomly combined, and the existing label information is fully utilized to expand the number of the labeled samples for model training under the condition that label data are scarce, so that the influence of insufficient labeled samples on the model performance and generalization ability can be effectively reduced; and the network flow data anomaly detection precision of the model is improved.

The invention provides a degradation data enhancement method based on combination of a conditional encoder and a diffusion model, which is characterized in that a combined network of the conditional encoder and the diffusion model is designed on the basis of a physical priori guidance generation principle, and a physical rule of monotonically decreasing residual service life is embedded into a potential space through a two-stage conditional encoder network. The high-fidelity degradation data conforming to the physical law is efficiently generated by using the space-time condition diffusion model, the problems of data scarcity and unbalanced distribution faced by the existing data-driven prediction method are solved, the generalization ability and robustness of the prediction model in a complex scene are remarkably improved, and the method can be applied to prediction of the residual service life of the equipment and has a good application prospect. The problem that the prediction precision is reduced due to insufficient samples and unbalanced distribution when a traditional residual service life prediction method faces data scarcity and complex working conditions is solved.

The application discloses an active carbon forming defect online detection method and device, and belongs to the technical field of material defect detection. The method comprises the following steps: collecting process parameters in real time and preprocessing, generating parameter sequences, and self-sensing knowledge identification of new processes, so that the lag start of quality control in the vacuum period is avoided. Once a new process is identified, the unknown defect pre-identification mechanism is triggered immediately, new abnormalities that do not conform to known defects are identified, and an unknown defect candidate set is generated. A multi-level hierarchical detection network is constructed, the unknown defect candidate set is adaptively detected, a benchmark model is established, samples are screened and the benchmark model is updated, a defect identification model suitable for the new process is formed, and the influence of insufficient samples on the detection accuracy is effectively alleviated. Three-dimensional quantitative indicators are set, the transition period is judged, and when the indicators all meet the preset requirements, the full model detection mode is automatically switched to. The quantitative indicators are used for objectively evaluating the identification ability of the detection system for new process defects and the process adaptability.

The invention relates to a pork flavor nondestructive testing method, device and equipment and a medium, and the method comprises the steps: inputting the real spectral reflectivity data of a real pork sample and the corresponding real physicochemical index data into a second data expansion model which is trained to a convergence state, generating simulated spectral reflectivity data and corresponding simulated physicochemical index data through different rounds of output so as to construct a plurality of simulated pork samples, and adding the plurality of simulated pork samples into the first sample data set so as to construct a second sample data set; and inputting the spectral reflectivity data corresponding to the to-be-detected pork sample into a pork flavor detection model trained by adopting the second sample data set to determine a pork flavor label corresponding to the to-be-detected pork sample. The technical problems of insufficient samples, low precision, long consumed time and strong destructiveness in traditional pork flavor detection are solved.

This invention discloses a method for rapid training of neural network models in a small-sample environment, comprising the following steps: S1, acquiring a small-sample dataset containing several categories and preprocessing it to obtain a standardized training sample set, wherein the number of samples in each category of the small-sample dataset satisfies the small-sample distribution characteristics; S2, training an initial neural network classifier based on the training sample set, and constructing prototype point-boundary corridor topology maps for each category in the feature latent space. This invention relates to the field of neural network model training technology. This method for rapid training of neural network models in a small-sample environment, by constructing prototype point-boundary corridor topology maps, can clearly understand the relative positions of each category in the feature latent space and the distribution characteristics of inter-class classification boundaries, thereby accurately identifying boundary gap areas with insufficient sample coverage and prone to misclassification, providing a clear target area for subsequent supplementary sampling operations.

The invention relates to the technical field of hydrologic ecology and water resource protection, in particular to a basin physical attribute fused short sequence hydrologic data ecological flow estimation method, and aims to solve the problems of low ecological flow estimation precision and poor regional transportability caused by insufficient statistical samples and lack of physical mechanisms under short sequence hydrologic data. According to the method, a nonlinear exponential function model with monthly median flow and rainwater collecting area as independent variables and ecological flow as a dependent variable is constructed, the rainwater collecting area extracted by a DEM and the confluence speed inverted by a Manning formula are introduced to serve as physical scale parameters, model parameters are optimized through a Lyon method reference value, and high-precision estimation is achieved. The method has the advantages that the estimation precision equivalent to that of a long sequence can be achieved only through short-sequence data, meanwhile, the regional transportability of the model is remarkably improved by introducing drainage basin physical characteristics, and reliable technical support is provided for ecological flow estimation of regions deficient in data.

The invention provides an abnormal sound sample generation method, and a part abnormal sound fault diagnosis method, device and equipment, and relates to the technical field of abnormal sound fault diagnosis. And generating target noise according to preset parameters, fusing the target noise with the real abnormal sound audio data spectrogram, and generating noisy abnormal sound spectrograms under different noise interferences through a gradual noise adding diffusion process. Sequentially executing N times of feature coding, channel dimension self-adaptive calibration and spatial dimension self-adaptive calibration on the noisy abnormal sound spectrogram, generating N joint calibration feature maps fusing noise suppression and abnormal sound enhancement layer by layer, and after feature coding, executing N times of up-sampling operation in combination with the joint calibration feature maps generated at the corresponding levels, and generating an abnormal sound sample set which is highly close to the real abnormal sound audio data. The problems of weak generalization ability of a diagnosis model and low detection accuracy in diversified actual working conditions caused by insufficient real abnormal sound fault samples are solved, and the accuracy and stability of abnormal sound fault diagnosis of parts are enhanced.

The utility model discloses an accessory structure meeting rapid sample loading of an early pregnancy and ovulation detection pen, which comprises a detection pen main body, a containing cavity is formed in the detection pen main body, one end of the detection pen main body is a liquid inlet end, a plurality of liquid inlet holes are formed in two sides of the liquid inlet end, the liquid inlet holes are communicated with the containing cavity, and the containing cavity is communicated with the containing cavity. And the plurality of liquid inlet holes form a honeycomb structure. According to the accessory structure, the problem that the detection result of the test pen is inaccurate or wrong when urine sample loading is interrupted and the loading amount is insufficient due to misoperation of a user can be avoided, the operation difficulty of the test pen of the user is reduced, and the accuracy and the stability of a chromatographic diagnosis product are improved.

The invention relates to the technical field of sampling methods, and discloses an optimized sampling method based on spatial weight, and the method comprises the steps: dividing a historical time period year by year, and calculating a fire spatial weight value of a to-be-sampled region according to historical fire data; mapping the fire space weight value to an interval range from 0 to 1 by adopting a normalization method to form a plurality of threshold intervals; sampling non-fire samples in each threshold interval range, and combining the fire samples to form a data set of a plurality of samples; carrying out wildfire risk modeling; respectively evaluating the precision of the four wildfire risk models to obtain two models with the highest precision; and analyzing the influence of the driving factors on the two models with the highest precision to obtain the key driving factors of the wildfire occurrence in the sampling area. According to the method, non-fire sample collection is optimized through the fire space weight value, the blank of insufficient sample representativeness in fire danger prediction is filled up, and the effectiveness of a sampling strategy based on the space weight for improving wildfire prediction is verified.