151results about How to "Prevent overfitting" patented technology

The invention relates to the technical field of remote sensing data processing and earth surface parameter inversion, and discloses an earth surface matrix parameter hyperspectral data inversion method and system based on machine learning. Comprising the following steps: constructing a multi-source heterogeneous hyperspectral data set; performing feature screening on the preprocessed hyperspectral data set based on an adaptive band selection algorithm, constructing a dynamic weight matrix by calculating mutual information entropy and inter-class distance measurement between spectral bands to realize intelligent screening of key feature bands, and combining spectral derivative conversion and spectral index calculation to generate an enhanced feature vector; and a multi-task transfer learning neural network model is constructed, and an output layer realizes multi-parameter collaborative inversion based on a multi-task learning architecture. And performing preprocessing and feature enhancement operation which is the same as that of the training data on the hyperspectral image data of the to-be-inverted region, inputting the trained neural network model, and outputting a surface matrix parameter inversion result.

The invention relates to the technical field of intelligent monitoring, in particular to an electronic guide vehicle steering system fault prediction method based on time sequence data analysis, which comprises the following steps: mapping a steering system into perception, decision and execution virtual computing nodes to construct a digital twin model; extracting inter-node time lag correlation through multivariable Granger causality operation, and constructing a virtual causality matrix; fault parameters such as current, angle and bias are injected into the model for numerical simulation, and a response sequence is generated through time domain iteration; using a dynamic time warping algorithm to calculate the morphological difference between the time sequence and the reference sequence so as to generate a time sequence feature; and constructing a digital twinning dynamic graph, inputting the dynamic graph attention network, and extracting spatio-temporal evolution characteristics to generate a fault prediction result. According to the method, accurate deduction of the full-life-cycle fault evolution process of the steering system is realized through digital twinning.

The invention discloses a large-range landslide mass displacement early warning method based on GNSS and radar data fusion, and the method comprises the steps: obtaining GNSS monitoring sequence data in a landslide region based on the position information of a reference station and a GNSS monitoring station in a GNSS monitoring system, and further obtaining the displacement related information of the landslide region; collecting data through a satellite InSAR, and generating deformation field data of a landslide area by comparing radar images at different time points; performing space-time alignment on the obtained data, and performing two types of data fusion S by adopting an adaptive Kalman filtering method after alignment; decomposing time series data of the fused landslide mass displacement by using wavelet transform, and extracting multi-scale features of a time domain and a frequency domain; the proposed features are combined with historical data of landslide mass displacement, a landslide mass displacement trend prediction model using an LSTM method added with a memory decline factor is input, and future trend prediction of landslide displacement is carried out; setting a multi-level threshold value, judging the change trend of the landslide mass displacement according to the displacement rate, the acceleration and the prediction result, and generating a multi-level early warning signal.

The invention discloses a material creep behavior prediction method and device based on a physical mechanism weighted network, and aims to solve the problems of low prediction precision and poor physical interpretability of an existing data-driven model under a small sample condition. The method comprises the following steps: constructing a physical mechanism weighting network which comprises a plurality of weighting mechanism units; each weighting mechanism unit is formed by coupling a weighting layer and a physical mechanism layer, the physical mechanism layer is packaged with a preset creep physical mechanism model, and the weighting layer is configured to represent a competition and conversion relationship of different physical mechanisms under an external load condition through a learnable weighting function; training the physical mechanism weighting network by using creep experiment data of a material so as to determine parameters of the weighting function; and predicting the creep behavior of the target material under a given stress-temperature condition by using the trained physical mechanism weighting network. According to the method, high-precision and physically interpretable creep prediction under small samples and a wide load domain is realized.

The application discloses a work clothes identification method combining continuous learning and effectively resisting forgetting disaster, and comprises the following steps: collecting data in an initial stage, constructing a work clothes classification model, and combining data to perform self-supervised training on the work clothes classification model; the trained work clothes classification model is divided into a deep model and a shallow model, an image self-encoder is constructed between the deep model and the shallow model, and the image self-encoder is trained, the intermediate features output by the shallow model are encoded and compressed through the trained image self-encoder, and compressed features are obtained; in a continuous learning task stage, work clothes category data are collected, the compressed features are decoded, and a new model is constructed; based on the new model, a Tt-stage work clothes classification model is constructed through a gradient boosting method, and after training, compression is performed to obtain a final model for work clothes identification; the application alleviates the forgetting degree of the work clothes classification model to old categories, and improves the plasticity of the model.

This invention provides a clustered federated multi-task learning method and apparatus for the Internet of Things (IoT). By clustering IoT terminal devices, the data distribution within the same cluster becomes more approximate. A federated multi-task learning algorithm is executed within each cluster, with global training and personalized training tasks performed on each IoT terminal device. This achieves data sharing within the cluster while fully utilizing local data from each IoT terminal device for training on personalized tasks, thus efficiently utilizing local data and improving training effectiveness. During local training on each IoT terminal device, the number of training rounds is adjusted based on computing power, fully utilizing the computing resources of each IoT terminal device and improving model training efficiency. Regularization constraints applied to personalized training tasks using the global model effectively prevent overfitting, control the degree of personalization, and improve model quality.

The invention relates to the field of carbon emission prediction, in particular to a subway station construction stage carbon emission prediction method and system. Measuring and calculating carbon emission of the subway integrated station at different stages according to a carbon emission coefficient method, and performing data preprocessing to obtain carbon emission monitoring data; performing feature and target variable separation on the carbon emission monitoring data to obtain a feature matrix and a target variable vector; constructing a carbon emission prediction model to perform feature processing on the carbon emission monitoring data to obtain carbon emission feature data; the method comprises the following steps: constructing a carbon emission prediction integrated model based on LightGBM and XGBoost, identifying carbon emission characteristic data through the carbon emission prediction integrated model, and optimizing parameters in the LightGBM model and the XGBoost model by using grid search to obtain a carbon emission prediction amount in a subway station construction stage.

The invention belongs to the technical field of electric digital data processing, and particularly relates to a PDC drill bit torsion impact signal collecting and processing method which comprises the steps that torsion data are segmented into a plurality of torsion data windows; determining the torsion waveform complexity of the torsion data window; determining the rotation speed slippage strength and the formation meshing complexity of the torsion data window; determining an adaptive order of an AR model of the torsion data window; and performing predictive filtering on the torsion data in the torsion data window to obtain a predictive residual sequence of the torsion data window, and extracting a torsion impact signal according to numerical characteristics of the predictive residual sequence. According to the method, the numerical characteristics of the torsion data and the rotating speed data are analyzed, the order of the AR model is self-adapted, the limitation of the AR model with the fixed order is overcome, background noise is filtered out, and the accuracy and robustness of the torsion impact signal are improved.

The application relates to the computer technical field and discloses a modeling method of a statistical mixed model in a big data distributed scene. The method comprises the following steps: distributing and storing data shards; initializing model parameters; iteratively performing an expectation step and a maximization step; the expectation step is scheduled to a GPU node to perform parallel calculation on posterior probability; the maximization step is scheduled to a CPU node to aggregate statistics and update parameters, perform component merging / deletion, and perform convergence judgment; meanwhile, a memory reuse mechanism based on reference counting and scope analysis is adopted to reduce redundant data transmission. Through heterogeneous task scheduling and memory collaborative optimization, the training speed, resource utilization rate, and model self-adaptation capability are improved.

The application provides a ship type classification prediction method and system based on K-means and XG-Boost, acquires ship data and carries out pretreatment, then adopts a K-means clustering algorithm to respectively cluster each kind of data in the pretreated ship data to obtain multiple clusters, calculates error sum of squares of all data in each cluster, calculates the classification number of each kind of data according to the error sum of squares, selects a certain classification number by using an elbow method, marks the ship type of all clustered ships according to the classification number, then takes the ships with marked ship types as training set samples, adopts an XG-Boost classification algorithm to train the training set samples to obtain multiple classification prediction models, verifies the multiple classification prediction models to obtain an optimal classification prediction model, and predicts the ship types of all ships in the world according to the optimal classification prediction model and marks the ship types. The application can accurately classify all ships in the world, and can avoid overfitting and underfitting of the model while ensuring the accuracy.

Embodiments of the present application disclose a face recognition method, device, equipment and medium. The method comprises: obtaining a to-be-recognized face image; wherein the to-be-recognized face image comprises a to-be-recognized near-infrared image and / or a to-be-recognized visible light image; inputting the to-be-recognized face image into a trained face recognition model to obtain a face recognition result; wherein the face recognition model is trained based on near-infrared sample images and visible light sample images of sample training objects. The above scheme uses the face recognition model trained based on the near-infrared sample images and the visible light sample images to perform face recognition, which is convenient to operate, can simultaneously perform face recognition on the to-be-recognized near-infrared image and the to-be-recognized visible light image, and avoids the case that the face recognition accuracy is not high due to the influence of environmental factors or the lack of detailed texture information when face recognition is performed according to a single to-be-recognized face image, thereby improving the face recognition accuracy.

The present application relates to the technical field of computer data processing, more particularly to a kind of continuous learning method based on nearest neighbor search enhancement.kNN-CL introduces k nearest neighbor search technology, without additional training cost, effectively deal with data imbalance, solve overfitting problem, improve the generalization performance.Different from traditional methods, kNN-CL can retrieve the k nearest neighbors of a test data, and the k nearest neighbors are only related to the test data, thereby realizing selective retrieval of data storage for each task, saving time and resources.

The application relates to the technical field of machine learning, and provides an animal sound classification method, system, device, medium and product. The application trains a classification model through a staged strategy, including pre-training the classification model based on a general audio dataset, end-to-end training the classification model based on an animal sound dataset, and deploying the trained classification model to an intelligent device; the intelligent device collects original audio signals in real time and extracts log-mel spectrum features as feature inputs of the classification model, and the model finally outputs prediction probabilities for each target sound category, so as to determine an animal sound classification result. The staged training method significantly reduces the model training difficulty, the log-mel spectrum features are used as inputs to improve the robustness of the model to background noise, and the generalization ability and overall accuracy of the classification model are effectively improved.

This application provides a method and apparatus for behavioral intent recognition, a training method, a medium, and a device. The method includes: acquiring an image to be recognized; processing the image to be recognized based on a person feature subnet to obtain a person feature sequence; and processing the image to be recognized based on a device feature subnet to obtain a device feature sequence; fusing the person feature sequence and the device feature sequence to obtain a fused feature sequence; and processing the fused feature sequence using an intent recognition model to obtain an intent recognition result. This application uses a person / device subnet and a device feature subnet to process the image to be recognized, respectively. During the network training phase, the person feature subnet and the device feature subnet are trained independently, eliminating the need to collect joint samples of "human-held devices"; it can directly utilize publicly available human / hand datasets and device images collected online, reducing annotation costs.

The invention provides an oil-water separation effect prediction method and device based on a Lonion model. The method comprises the following steps: defining an input independent variable vector f (x) = [x1, x2,..., x8] T; performing feature engineering extension on the input independent variable vector to generate an extended feature vector f (z); the extended feature vector f (z) comprises an original feature vector, a polynomial feature vector and a physical interaction feature vector; inputting the extended feature vector f (z) into a pre-trained prediction model to obtain a predicted value, output by the pre-trained prediction model, of the oil content of the effluent of the air flotation synergistic device; inputting the predicted value into an integrated prediction framework to obtain a final predicted value and uncertainty estimation; displaying the final predicted value and the uncertainty estimation to a user through a display interface; according to the technical scheme, the accuracy, robustness and physical consistency of oil content prediction of the effluent can be improved.

This invention discloses a lidar depth completion method based on hierarchical minimum surface reconstruction, relating to the fields of computer vision and image processing technology. The method includes: performing depth value inversion and morphological dilation on a sparse depth map to obtain a scene dilated depth map and a full-resolution effective mask; then downsampling to obtain a coarse-scale scene dilated map and a coarse-scale mask; using an iterative convolution kernel of the Laplacian operator to obtain a coarse-scale filled map; upsampling the coarse-scale filled map to obtain an upsampled depth map, and then fusing it with the scene dilated depth map to obtain a fused depth map; locating all remaining holes in the fused depth map using connected component analysis, and calculating the mean depth of the effective pixels within the annular pixel band around each remaining hole; using this mean depth to fill the corresponding remaining holes to obtain a hole-free depth map; and after global Gaussian blurring, performing a depth value inversion operation to obtain a scene dense depth map, thereby achieving high-precision 3D reconstruction.

This invention relates to the field of natural language processing and information extraction technology, and discloses a method for automatic generation and iterative optimization of annotation guidelines for event extraction. The method includes: generating initial annotation guidelines based on annotation samples using a large language model and compressing them into lightweight guidelines for predefined event types; employing a two-stage extraction strategy, first identifying event types using the lightweight guidelines, and then recalling complete guidelines based on the types for end-to-end extraction to obtain prediction results; identifying and clustering prediction errors based on a predefined set of event-layer and argument-layer error types to form error clusters; and generating update operations for the annotation guidelines using a large model for each error cluster, and completing iterative optimization after verification of effectiveness. This invention achieves automated generation and continuous evolution of annotation guidelines, and significantly improves the maintainability, generalization ability, and event extraction performance of the guidelines through structured representation and error clustering mechanisms.

This application discloses a model building method, apparatus, electronic device, and storage medium, relating to the field of smart home technology. The method includes: acquiring a first noise sample set, which is collected from different types of fluid power equipment under different operating conditions; expanding the first noise sample set based on preset sound quality feature constraints to obtain a second noise sample set; and constructing a sound quality prediction model for the fluid power equipment based on the second noise sample set. The technical solution provided by this application can effectively expand the sample size and reduce the cost of sample acquisition, solving the pain points of insufficient training data and low prediction reliability under small sample conditions, and providing support for the sound quality optimization of range hoods.

The invention relates to a channel foreign matter image generation and channel foreign matter detection method based on a generative adversarial network, and the method comprises the steps: image collection: employing an unmanned plane for inspection, and collecting a channel image; image noise reduction: carrying out noise reduction processing on the acquired image by adopting a median filtering technology; model improvement: based on the GAN model, introducing an SE attention mechanism, and improving a loss function to obtain an improved GAN model; model training: training the improved GAN model through a training data set; image generation: generating a channel foreign matter image by using the trained foreign matter image generation model, and forming an extended sample set; foreign matter recognition: training a target detection network in combination with the original sample set and the extended sample set, and detecting and recognizing the channel foreign matter based on the trained target detection network; and foreign matter positioning: according to an identification result, based on a GIS system, determining a spatial position of a foreign matter. The method is beneficial to generating images similar to real channel foreign matter features, and the channel foreign matter detection precision is improved.

The invention belongs to the technical field of power distribution network leakage current detection, and particularly relates to a CNN and FiLM-based leakage current type intelligent identification method. Comprising the following steps: S1, collecting real-time operation data of a typical power supply area, obtaining leakage current waveforms and related environment characteristic parameters under different working conditions, and constructing an original leakage current sample data set; s2, a CNN-based leakage current classification model is constructed, a FiLM condition modulation module is introduced, a multi-task learning framework is used at the tail of the network, a main task is leakage current type identification, and an auxiliary task is grounding system discrimination; s3, using a weighted cross entropy loss function to alleviate a class sample imbalance problem; an OneCycleLR dynamic learning rate scheduling strategy is introduced; s4, evaluating the performance of the model on the test set, and using the accuracy and the confusion matrix as evaluation indexes; according to the method, high-precision identification of multiple types of faults such as single-phase grounding, arc type electric leakage and direct current system electric leakage is realized, different grounding systems can be adapted, and the accuracy and robustness of system diagnosis are improved.

This invention discloses an intelligent detection method and system for bid format compliance of sealed bid documents, relating to the field of electronic bidding. The method includes: performing format detection based on the format rules of the bidding documents to obtain a set of violation records and a single format detection report; constructing a format error fingerprint vector, calculating cross-file error similarity, and generating a cross-file error similarity matrix; constructing an initial suspected association graph using each sealed bid document as a node and the similarity values ​​in the cross-file error similarity matrix as edge weights between nodes; dynamically adjusting the connectivity threshold, updating the node connection relationships in the weighted disjoint-set data structure, and defining the groups divided under the optimal connectivity threshold as abnormal similar file groups; calculating a bid-rigging suspicion score, generating a bid-rigging detection report, and outputting the bid compliance detection results. This solves the technical problem in existing technologies where isolated compliance judgments on individual bid documents fail to identify potential bid-rigging behavior.

The invention discloses an ear plate local stress high spatial resolution monitoring method and system based on ultra-weak grating spectral shape demodulation. The system comprises an optical fiber sensing array, a light source module, a detection module, a data acquisition and physical information neural network inversion module and a display alarm module. The optical fiber sensing array is composed of N ultra-weak gratings arranged in a stress concentration area of the ear plate and is used for sensing local strain; the light source and detection module adopts a time division and wavelength division multiplexing strategy to realize synchronous monitoring of multiple ear plates; the data acquisition and inversion module acquires stress distribution along the length direction of the ultra-weak grating based on a physical information neural network inversion model; and the display alarm module realizes local stress threshold alarm of the ear plate. According to the invention, high-precision and real-time monitoring of local stress of metal connection nodes such as ear plates can be realized, the method is suitable for long-term online health monitoring of large-scale structure connection areas such as bridge inhaul cables, steel trussed beams and cable bent towers, and technical support is provided for structural fatigue early warning and safety assessment.

The application discloses a social media-based interpretable dynamic graph network benefit prediction model training method, relates to the technical field of text analysis and market prediction, and obtains market data and related social media data of a financial asset to divide the data by weeks; for the social media data of each week, a topic is taken as a node, and similarity between topics is taken as an edge weight, a topic correlation graph of each week is constructed to obtain weekly graph data; a prediction model performs time sequence updating and market prediction according to the weekly graph data; a text time sequence memory unit obtains a current week topic memory vector according to a current week topic text embedding vector and in combination with a last week topic memory vector; a multi-layer graph attention network captures the correlation features between nodes through an attention mechanism and adopts attention weights for weighted summation to obtain a current week global graph embedding; and a classifier outputs a next week price change prediction result. The application solves the problem that the prior art cannot effectively capture the time sequence evolution of topic sentiment and the correlation between topics.

The application relates to the technical field of computers, in particular to the field of artificial intelligence, and provides a work order classification model training method and device, an electronic device and a storage medium, to improve the accuracy of the model. The method comprises the following steps: obtaining a work order sample set, the work order sample comprising a category label of a corresponding customer service work order and dialogue text information, the dialogue text information being obtained based on a customer service session recorded in the corresponding customer service work order; based on the category label of each work order sample, filtering reference work order samples to be expanded from the work order sample set; based on a preset data augmentation strategy, performing data augmentation on the dialogue text information in the reference work order sample to obtain corresponding expanded work order samples; and based on each work order sample and the obtained expanded work order samples, performing model training to obtain a trained work order classification model. According to the application, the dialogue text information of the customer service work order is subjected to data augmentation, the number of work order samples of various categories is balanced, and the accuracy of the model can be effectively improved.

This disclosure relates to the fields of computer vision and wildlife monitoring technology, and specifically to a method and system for identifying Tibetan macaques in the wild based on deep learning. First, a face detection model is used to detect and crop faces from raw wild images. The detection model is initialized with pre-trained weights and fine-tuned by freezing the backbone network. Then, an individual identification model is constructed to output classification results. During training, a joint loss function is used, combining Focal Loss to suppress easily classified samples and guide attention to difficult-to-classify samples. Supervised contrastive loss is used to enhance the discriminative power of the feature space. Simultaneously, an adaptive hybrid enhancement strategy and a hierarchical Dropout regularization mechanism are introduced to effectively address the problems of small sample sizes, class imbalance, and overfitting in complex wild environments. This disclosure enables high-precision identification of Tibetan macaques, providing an efficient technical means for long-term ecological monitoring of wild primates.

This invention relates to the fields of digital healthcare and artificial intelligence, disclosing a method, device, electronic device, and storage medium for predicting the relationship between diseases and genes. The method includes: constructing a disease similarity network, a gene-disease relationship network, a first integrated network, and a second integrated network; extracting representation vectors from the first integrated network, the second integrated network, and the gene-disease relationship network to obtain a first representation vector, a second representation vector, and relationship node representation vectors; performing a weighted summation of the first representation vector, the second representation vector, and the relationship node representation vectors to obtain a first target representation vector and a second target representation vector; performing a weighted summation of the first target representation vector and the second target representation vector to obtain a target representation vector; and performing linear calculations on the target representation vector to obtain the probability that the gene data to be analyzed and the disease data to be analyzed are correlated in the target representation vector. This invention can improve the intelligence and accuracy of predicting the relationship between genes and diseases.

The application discloses a multi-modal depth perception and grasping system based on transparent objects, and relates to the field of robot operation, comprising a multi-spectral perception module, a depth correction module, a grasping posture generation module and a control module. The application comprehensively acquires visual information and thermal radiation information of the transparent object by combining two perception methods of RGB-D image and thermal imaging (TIR) image, and analyzes systematic errors of the depth map to detect error sources of the RGB-D camera and the TIR camera. The system adopts an encoder-decoder model as a depth correction core, the model extracts complementary features of the RGB-D image and the TIR image through a modal exclusive encoder, and utilizes a feature fusion module to complete feature alignment and integration, thereby effectively improving the depth estimation accuracy on the transparent surface. Meanwhile, a Bayesian optimization method is adopted to optimize hyperparameters of the depth correction model, through hyperparameter optimization and model fitting processing, the system effectively avoids overfitting and underfitting problems, and further improves the robustness of the depth correction model and the grasping accuracy of the transparent object.

This invention relates to a method, apparatus, and medium for rapid sorting of poultry meat percentage based on X-ray imaging. The method includes: Step S1: acquiring X-ray images of the poultry to be tested; Step S2: identifying the sternal region and skin boundary based on the X-ray image; Step S3: drawing a normal at the sternal apex, obtaining the intersection of the normal at the sternal apex and the skin boundary as the first surface point, and obtaining the initial bone-skin distance based on the distance from the sternal apex to the first surface point; Step S4: obtaining the angle between the X-ray machine's visual axis direction and the normal direction of the first surface point as the consistency difference angle, obtaining the sternal integrity ratio and sternal principal axis attitude angle based on the sternal region, and correcting the initial bone-skin distance based on the consistency difference angle, sternal integrity ratio, and sternal principal axis attitude angle to obtain the corrected bone-skin distance; Step S5: obtaining the meat percentage grading result based on the corrected bone-skin distance. Compared with the prior art, this invention can achieve poultry meat percentage sorting based on X-ray imaging.

The invention discloses a data model hybrid driven pine wood nematode disease transmission risk prediction method, and belongs to the technical field of pine wood nematode disease monitoring. The objective of the invention is to improve pine wood nematode disease transmission risk prediction precision. The method comprises the following steps: constructing a multi-source factor data matrix; according to the propagation law of the pine wood nematode disease, a cellular automaton model is constructed, and the spatial diffusion process of the pine wood nematode disease in each county-level administrative division geographic unit in the forest is calculated and simulated; constructing a TabNet neural network model, performing feature learning on the multi-source factor data matrix, and identifying key driving factors of pine wood nematode disease transmission; a PINN physical information neural network framework is constructed, a cellular automaton model is used as a physical constraint module, a TabNet neural network model is used as a data driving module, and the physical constraint module and the data driving module are linked through a composite loss function; and using the trained PINN physical information neural network framework to predict the propagation and diffusion trend of the pine wood nematode disease in the future and realize visualization.

This application provides a model training method, an image detection method, and a storage medium, relating to the field of computer processing technology. The method includes: iteratively optimizing the initial parameters corresponding to the enhancement algorithms in the target enhancement strategy set based on sample data, a target enhancement strategy set, and an initial model, generating optimal enhancement parameters for each algorithm; processing the sample data based on the optimal enhancement parameters to generate optimal sample data; and training the initial model based on the optimal sample data to obtain a target model. The target enhancement strategy set is a collection of multiple enhancement algorithms determined based on the distribution characteristics of the sample data. Because this application employs an iterative optimization mechanism for enhancement parameters based on feedback from sample data and the initial model, it achieves dynamic adaptation between the enhancement strategies and data distribution characteristics, thus significantly improving parameter optimization efficiency, avoiding model overfitting, and enhancing the generalization ability, robustness, and practical adaptability of the target model.