60results about How to "Reduce overfitting" patented technology

The invention relates to an underwater acoustic target recognition system and method based on multi-modal depth feature fusion, and belongs to the field of underwater acoustic target recognition. The method comprises the following steps: firstly, performing preprocessing on an obtained underwater acoustic target original audio and associated metadata, and constructing a multi-modal data set; and then the constructed multi-modal sample is input into an identification model, the model extracts deep representation of each modal through a multi-branch feature coding network, depth alignment and complementary aggregation of different modal features are realized by using a cross-modal cross attention mechanism guided by potential query, and a target identification result is output based on a decision network of mixed experts. According to the method, experimental verification is carried out on two disclosed underwater acoustic data sets, the experimental result verifies the effectiveness of the multi-modal deep fusion and hybrid expert adaptive decision strategy adopted by the method, and through mining the complementary advantages of acoustic features and semantic priori, the multi-modal deep fusion and hybrid expert adaptive decision strategy is obtained. And the robustness and generalization ability of the underwater acoustic target recognition system in the strong-noise and multi-working-condition environment are remarkably improved.

The invention discloses a drilling pump pressure prediction method based on an artificial neural network, and the method comprises the following steps: S1, data collection; s2, data preprocessing; s3, constructing a neural network model; s4, model training and optimization; and S5, model verification and deployment. According to the method, multi-channel data fusion and time synchronization optimization are innovatively adopted, so that the data quality and consistency are improved; in combination with a deep neural network and a self-adaptive optimization strategy, the precision and generalization ability of pump pressure prediction of the drilling pump are improved; and an online updating mechanism is introduced, so that the model can be dynamically optimized according to real-time data, the defects of low prediction precision, poor adaptability and difficulty in real-time updating of a traditional method are overcome, and an efficient and reliable prediction means is provided for intelligent drilling control.

This invention discloses a fruit quality detection method and system based on a fruit quality detection model. First, the original hyperspectral image R of the fruit to be detected is acquired and preprocessed. Then, the preprocessed hyperspectral image R2 is input into the fruit quality detection model to obtain the surface defect results and internal quality index values ​​of the fruit. Compared with the prior art, this invention can achieve simultaneous detection of the internal and external quality of the fruit at low cost and high efficiency, and obtain the fruit quality grade result.

This invention discloses a bearing remaining life prediction method based on frequency domain degradation perception. It constructs an integrated prediction system through a time-frequency decomposition dual-branch feature extraction module, a frequency domain degradation perception weight allocation module, and a life prediction loss function optimization module. First, the time-frequency decomposition dual-branch feature extraction module decomposes the original time series into high and low frequencies and inputs it into an sLSTM and mLSTM dual-branch structure for feature extraction. Second, the frequency domain degradation perception weight allocation module performs degradation perception processing on the fused features, generating dynamic weights and modulating the features to obtain the final fused features, which are then mapped to the prediction space. Finally, the life prediction loss function optimization module calculates the loss, optimizes the model output, and obtains the bearing remaining life prediction result. This invention comprehensively covers the full-stage features of bearings from healthy to severely degraded, improves the sensitivity of early fault detection, and still possesses excellent generalization and robustness under complex operating conditions.

The application discloses a voltage sag identification method based on a triple tuple prototype network under a small sample, and belongs to the technical field of power quality analysis. The method uses a triple tuple feature extractor, a large number of voltage sag triple tuples are constructed, and effective sag features can be extracted by the model under the condition of a small number of training samples. Then, in view of the problem that some voltage sag features are similar and easy to confuse, an efficient channel attention mechanism is integrated into the triple tuple feature extractor, cross-channel feature interaction information is captured under the condition of only a small number of parameters, the model can pay attention to the key feature area, a prototype classifier is finally constructed, representative class prototypes are learned for each class by using the extracted sag features, and the final sample class is determined by comparing the similarity between sample features and class prototypes. Under the condition of limited sample data, the method can realize accurate voltage sag classification effect, and has strong practical application significance.

The application discloses a cross-natural language code retrieval model training method, a cross-natural language code retrieval method, a device, equipment and a medium, and relates to the technical field of artificial intelligence and software engineering. The cross-natural language code retrieval model training method comprises the following steps: obtaining an original corpus database, and constructing training data according to the original corpus database; performing confusion and inversion on main language codes to obtain main language code samples, wherein the main language code samples comprise main language code positive samples and main language code negative samples; and training an initial model through a gradient inversion layer according to the training data and the main language code samples to obtain a target model. According to the application, the natural language-specific "fingerprint" features in the codes can be removed, the embedding space alignment direction can be unified, the sampling distribution deviation in the training process can be reduced, and the consistency and generalization capability of cross-language code retrieval can be improved.

The invention discloses an under-crown temperature inversion method and system and application, and relates to the technical field of forest ecology and microclimate inversion, and the under-crown temperature inversion method comprises the following steps: obtaining flux tower observation data; constructing and training a baseline prediction model to obtain a baseline prediction value of the vertical temperature difference; constructing and training a gating correction function based on the stability index; adding the baseline predicted value and the output value of the gating correction function to obtain a final predicted value of the vertical temperature difference; and obtaining the under-crown temperature through inversion by subtracting the final predicted value of the vertical temperature difference from the on-crown temperature. The invention provides a flux tower scene-oriented under-crown temperature inversion method and system and application, and the method comprises the steps: reconstructing a vertical temperature difference through employing the conventional environment elements of a flux tower and carrying out the inversion of the under-crown temperature under the condition of lacking of the on-crown / under-crown surface temperature or lacking of temperature profile observation; therefore, the availability, comparability and interpretability of the microclimate data of the site are improved.

This invention provides a method for identifying the authenticity of wheat flour based on the fusion of Raman and near-infrared spectroscopy, belonging to the field of wheat flour authenticity identification technology. The method includes: acquiring Raman and near-infrared spectral data of the wheat flour sample to be tested, and preprocessing them separately; constructing and training a fusion detection model, which includes a data-level fusion module, a feature-level fusion module, and a decision-level fusion module; the data-level fusion module generates full-spectrum fusion data; the feature-level fusion module concatenates core features into a comprehensive feature set; the decision-level fusion module inputs the comprehensive feature set into a hybrid model; and the trained fusion detection model processes the comprehensive feature set to output the authenticity identification result of the wheat flour sample to be tested. This invention, through a three-level spectral fusion architecture and a dynamic adaptive adversarial mechanism, can effectively achieve high-precision and high-robust identification of wheat flour authenticity.

The present application relates to the technical field of ship identification detection, and more particularly to a layered attention fusion and edge enhancement SAR ship detection method and system. The method comprises the following steps: obtaining initial ship data images and preprocessing to obtain input feature maps; constructing an enhanced image recognition model, training the enhanced image recognition model according to the input feature maps to obtain a trained image recognition model; the enhanced image recognition model comprises a multi-scale edge information selection module, a cross-domain feature gating module and a layered attention fusion block module; and inputting an actual SAR image into the trained image recognition model to obtain a ship detection result. By designing and using the multi-scale edge information selection module, the cross-domain feature gating module and the layered attention fusion block module, the present application realizes better training stability and generalization ability in a few-shot detection scenario, effectively alleviating the overfitting problem of the baseline model when data is scarce.

The application discloses a face forgery detection method and system based on federal incremental learning, comprising the following steps: constructing a federal incremental learning framework, each client is provided with a local face forgery detection model, and a global server is provided with a global face forgery detection model; the local face forgery detection model is subjected to cross-entropy training in a basic training stage, generates a forgery substitute sample in an incremental training stage, and is used as training input together with a real face sample and a forged face sample; after training, the model parameters are uploaded to the global server for aggregation; the global face forgery detection model is updated based on a difference perception aggregation strategy; an adversarial perturbation is trained based on the global face forgery detection model converged in the present stage and the real face sample, and a perturbation pool is refreshed until a preset termination condition is reached; and a to-be-tested face image is identified as real or fake based on the trained global face forgery detection model. The application can relieve catastrophic forgetting and improve cross-task generalization ability and robustness.

The invention relates to a new energy output prediction method and system based on an NHITS model, and the method comprises the steps: obtaining the input characteristics of a to-be-tested new energy object, forming new energy time series data, and enabling the new energy to comprise photovoltaic or wind power; the input characteristics corresponding to photovoltaic output prediction comprise time, solar irradiance, air temperature, air pressure, humidity and historical power data, and the input characteristics corresponding to wind power output prediction comprise time, meteorological variables and wind speeds and wind directions at different heights; new energy output prediction is carried out by adopting an NHITS prediction model, the NHITS prediction model carries out multi-time scale decomposition on a time sequence through a hierarchical recursive structure, the time sequence is divided into a plurality of modules stacked according to layers, and each module carries out partial interpretation on input new energy time sequence data under different time scales and outputs a prediction component; and superposing the prediction components of each layer to obtain a final prediction result. Compared with the prior art, the method can maintain high prediction precision and stability.

The application provides a multimedia information recommendation model training method and device and electronic equipment, the method comprises the following steps: extracting a pre-training sample set based on the basic historical data; training a basic recommendation model based on the pre-training sample set to obtain model parameters of the basic recommendation model; obtaining industry historical data in a multimedia information recommendation environment; and determining the model parameters of the multimedia information recommendation model according to the industry historical data, thereby enhancing the accuracy and relevance of multimedia information recommendation and improving the generalization of the multimedia information recommendation model. The embodiments of the application can also be applied to various scenes such as cloud technology, artificial intelligence, intelligent transportation and auxiliary driving.

The application discloses a kind of weakly supervised image semantic understanding methods based on multi-task learning, comprising the following steps: obtaining task missing image, constructing multi-level task sharing encoder, extracting high-level semantic information layer by layer, input corresponding decoder branch;Construct public space-task space feature mapping module, through the unaligned task fusion module and task interaction mapping module, update each subtask feature by mapping;Task adaptive feature update module is constructed, and multi-level iterative update unaligned task feature;Task adaptive weakly supervised image semantic understanding framework is constructed, model loss function is established, image data with task missing is input into model, and obtains multi-task prediction result such as semantic segmentation, depth estimation, surface normal estimation.The application is according to the data information of task label unaligned, through the mapping interaction of public space and task space, fully fuses unaligned task feature, iteratively generates high-quality multi-task prediction result, can effectively handle weakly supervised problem with task missing, and simultaneously improves each task prediction accuracy.

The invention relates to a three-dimensional reconstruction method based on Gaussian point information redistribution and geometric structure constraint. The method comprises the following steps: acquiring target scene image data, and extracting a visual Gaussian point attribute parameter set based on a 3D Gaussian sputtering framework; randomly rejecting part of Gaussian points, determining a neighborhood point set of the Gaussian points, and constructing an approximate model to obtain a gradient item set; constructing an incidence matrix and calculating an information distribution coefficient, and compensating the opacity and color information of the rejected points to neighborhood points to update a parameter set; constructing a total loss function containing basic reconstruction, local direction consistency and main direction alignment loss, and iteratively optimizing parameters; and dynamically adjusting the Gaussian point shielding rate and modifying opacity associated parameters, maintaining physical constraints, and repeating information redistribution until training is completed, thereby realizing high-quality three-dimensional reconstruction. By adopting the method, the definition of the geometric structure can be improved.

The application provides a multi-branch graph adaptive network for individualized motor imagery electroencephalogram signal classification and a method thereof. Through an adaptive matching technology, a graph convolutional neural network (GCN) is used to learn a mapping relationship between an EEG signal and an optimal time-frequency domain processing method, and an optimal processing method is matched for each individual. The multi-branch network comprises an original data branch, a CWT branch and an STFT branch, and corresponding feature extraction networks are respectively designed. A data enhancement method based on a super-resolution generative adversarial network is used to expand a training data set and improve the generalization ability of the model. Experimental results show that the method significantly improves the classification accuracy of individualized EEG signals and has practical application value.

The present disclosure provides a training method of an image extraction model and an image extraction method, which can be applied to the fields of image processing and pattern recognition. The image extraction model comprises a linear embedding network, an encoding network and a decoding network. The training method comprises: processing a sample image obtained by using the linear embedding network to obtain an embedding feature map; processing the embedding feature map by using the encoding network to obtain i encoding feature maps; processing the i encoding feature maps by using the decoding network to obtain a weight fusion feature map, wherein the decoding sub-network is constructed based on a local-global attention layer and a weight feature fusion layer; performing segmentation head mapping processing on the weight fusion feature map to obtain an image segmentation result, wherein the image segmentation result represents a geological change attribute of a target geographical environment region; and training the image extraction model according to the image segmentation result and label data corresponding to the image segmentation result to obtain a trained image extraction model.

The invention discloses a bearing fault identification method based on multi-scale space-time synchronization attention and space-time alignment. The method comprises the following steps: carrying out sliding window segmentation and normalization preprocessing on an original vibration signal; constructing an adjacent matrix by using a K-nearest neighbor algorithm, and extracting local spatial features by using a graph convolutional neural network; bidirectional sequence features are extracted through a bidirectional gating loop unit network, and weighted fusion is carried out in combination with a global attention mechanism; and after space and time sequence features are spliced, fault prediction is realized through adaptive average pooling and a full-connection classifier. According to the method, graph modeling, GCN, BiGRU and an attention mechanism are fused, cooperative extraction of space-time double-path features is realized, the problems of weak space modeling and lack of time sequence dependence in a traditional method are effectively solved, and the bearing fault recognition precision and generalization ability in a multi-working-condition and strong-noise environment are remarkably improved.

This invention discloses a PolSAR image classification method and related apparatus based on a complex convolutional Kolmogorov-Arnold network, belonging to the field of land cover classification technology. The method includes: acquiring a PolSAR image to be processed and performing preprocessing; inputting the preprocessed PolSAR image into a preset complex convolutional Kolmogorov-Arnold network to obtain a classification result; wherein the preset complex convolutional Kolmogorov-Arnold network includes complex KAN convolutional layers, multi-branch complex KAN convolutional blocks, and CV-PolyLoss. This invention has promising prospects for practical remote sensing application deployment.

The application discloses a kind of deep learning landslide identification method and system of fusing geology and mining prior knowledge, it is related to geological disaster remote sensing identification and deep learning technical field, the method includes: constructing the geological and mining prior knowledge graph of target coal mine area, and according to prior knowledge graph generates prior feature map;Acquire the remote sensing image to be identified, prior feature map is registered with remote sensing image in space, the center point of landslide candidate area in the remote sensing image is extracted using lightweight network, and landslide candidate area is obtained by adaptive cropping;The segmentation model after target domain meta migration learning training is input into landslide candidate area, and the landslide probability graph of each candidate area is obtained;Through channel attention mechanism, the landslide probability graph under multiple prior assumptions is adaptively fused, and the final landslide identification mask is generated, and the landslide disaster identification result is output according to landslide identification mask.In this way, the recognition accuracy and generalization ability are improved under the condition of sample scarcity.

This invention relates to the field of artificial intelligence technology, and in particular to a method and apparatus for detecting and locating multiple sound events based on a neural network model. The method includes: innovatively designing time-frequency multi-scale residual convolutional blocks, which, together with a Conformer module and a cross-stitch unit module, form a network model to extract features at multiple scales, enhance long sequence modeling, and promote task-based collaborative optimization, thereby improving performance and accuracy; in terms of data processing, pre-emphasis and frame-by-frame windowing improve feature quality, audio channel swapping and spectrum enhancement increase data diversity and reduce overfitting, and SALSA-Lite features are used to enhance feature representation; in terms of training strategy, a multivariate loss function is used to accelerate convergence while considering task requirements, and hyperparameters are flexibly adjusted using a validation set. This makes the method highly efficient in training, has excellent practical performance, strong generalization ability on unknown data, and can accurately cope with complex and ever-changing real-world scenarios, effectively overcoming the shortcomings of traditional methods.

This invention discloses a method for determining the elemental content of iron ore using variable importance-artificial neural network-assisted laser-induced breakdown spectroscopy. The method includes the following steps: S1: Collect LIBS spectral data and elemental content data from at least 13 batches of iron ore in at least four categories. Use radiometric analysis (RF) to measure the importance of LIBS spectral features. A variable importance threshold is used to optimize the input variables of the BP-ANN model. The variable importance threshold and the number of neurons in the BP-ANN model are optimized using the determination coefficients and root mean square error of five-fold cross-validation to establish a VI-BP-ANN model. S2: Using the VI-BP-ANN model from S1, input the LIBS spectral data. The model ranks the features according to variable importance and calculates the elemental content in the iron ore. This invention's detection method can rapidly detect the total iron content, calcium content, magnesium content, aluminum content, and silicon content in iron ore.

The application discloses an ICS intrusion detection system and method fusing reinforcement learning and feature selection optimization, carries out binary coding and population initialization on data feature selection of an industrial control system (ICS) historical data set, carries out offline training through SVM-reinforcement learning, takes the accuracy obtained on a verification set as a fitness function, designs cross operation and mutation operation based on cumulative probability to update the population, and obtains an optimal feature set after iterative optimization; feature selection is carried out on an ICS real-time data set based on the optimal feature set, online intrusion detection testing is carried out on the real-time data set through support vector machine (SVM)-reinforcement learning, and thus an intrusion detection performance index is obtained. The application adopts a new mode based on SVM-reinforcement learning, and fuses intelligent optimization of optimal feature selection on this basis, and improves the intelligent design level and the precision of intrusion detection of the ICS intrusion detection system.

The invention provides a plant leaf chloroplast three-dimensional imaging axial resolution enhancement method and system, and the method comprises the steps: constructing an in-vitro model sample through fluorescent microspheres, and the in-vitro model sample comprises a non-scattering true value image and an axial stretching distortion image caused by scattering of the same visual field imaging; pre-training the deep learning network by using the in-vitro model sample, and establishing a pre-training model; constructing a plant real sample of plant leaf chloroplast, and training the pre-training model to obtain an axial resolution enhancement model; and inputting a to-be-processed chloroplast three-dimensional fluorescence image collected in a complete leaf into the axial resolution enhancement model, and outputting a chloroplast image used for accurate three-dimensional reconstruction after axial resolution enhancement. According to the method, hardware does not need to be transformed, the axial deformation of the plant leaf chloroplast in three-dimensional imaging is specifically corrected through an innovative training data preparation strategy and supervised deep learning, and the real three-dimensional form of the chloroplast is recovered.

This invention relates to a cross-machine few-sample equipment fault diagnosis method based on transmission path enhancement, comprising: 1. acquiring the vibration signal of the transmission system, segmenting it into fixed-length segments, and preprocessing it; 2. constructing a diagnostic guidance perturbation pair for each vibration signal segment; 3. applying the perturbation to obtain an enhanced signal; 4. extracting local temporal representations and global semantic representations from the enhanced signal; 5. performing self-supervised pre-training based on the representations to obtain a pre-trained encoder; 6. performing few-sample fine-tuning to obtain a target domain fault diagnosis model; 7. inputting the operating signal of the target domain equipment to be diagnosed into the target domain fault diagnosis model and outputting the fault category or abnormal state identification result. This invention can effectively alleviate the domain offset problem caused by differences in equipment structure, changes in transmission paths, and complex operating condition perturbations in high-end equipment transmission systems, and improve the accuracy, stability, and cross-platform generalization ability of fault diagnosis under extremely low sample conditions.

The application relates to the technical field of gear detection, and discloses a gear life prediction method based on a multi-modal space-time coupling graph neural network, which comprises the following steps: collecting multi-modal data for pretreatment, and extracting multi-modal feature sequences according to sliding time windows. In each sliding time window, a collection node and a measured gear node are taken as a node set, physical coupling edges and data coupling edges are determined, edge weights are fused, and a space-time coupling graph is constructed. Time correlation information is extracted, feature propagation is carried out on the space-time coupling graph, cross-modal attention weight distribution is carried out on different modes, and a degradation state representation of the measured gear is obtained. A parameter set is obtained based on data with a remaining life label. In the running process, newly formed multi-modal feature sequences and the space-time coupling graph are input, and the remaining life of the measured gear is output. When the remaining life is less than a maintenance threshold, a maintenance suggestion is generated. The application realizes stable and reliable prediction of the remaining life of the measured gear.

The application provides a human pose estimation method based on attention and adversarial network, a medium and equipment; wherein the method is: inputting an image to be predicted into a data enhancement module based on an adversarial network for data enhancement processing to obtain a data enhanced image; inputting the data enhanced image into a human pose estimation network based on multi-scale spatial attention to obtain an output heat map; converting the output heat map into spatial coordinates of each key point to generate a human pose, and then obtaining a human pose estimation result. The method can obtain spatial attention from different scale features, simultaneously perform multi-scale fusion on the attention, combine global and local attention features, finally generate more fine spatial attention, and can solve the prediction problem of occluded images and complex background images.

This invention discloses a computer program product and its application for data processing devices in the field of bioinformatics for ctDNA variant detection. The technical problem this invention aims to solve is how to detect ctDNA variants in early-stage cancer or postoperative minimal residual disease (MRD) under conditions of no UMI library construction and moderate sequencing depth. This invention constructs a set of supporting sequences for candidate variant sites, generating a sequence feature tensor and a fragment physical feature vector (including normalized fragment length). The former is input into a first neural network branch to extract sequence representation, and the latter into a second neural network branch to extract physical representation. A fusion module combines the sequence representation and physical representation, and a gating unit calculates the gating weight based on the physical representation and dynamically adjusts the contribution of the sequence representation, outputting the probability of the true ctDNA variant. This invention utilizes physical laws to suppress sequencing noise and can be applied to monitor MRD under conditions without molecular barcodes.

The application provides a skeleton data action recognition method based on a multi-relation graph convolutional neural network, which extracts high-level features of skeleton data from different scales by considering human natural connection relations, symmetric relations and global cooperation relations, and effectively fuses the high-level features by a relation attention mechanism. In this way, the network pays more attention to key part information in different actions, and does not lose effective information of other parts. In order to solve the overfitting and oversmoothing problems commonly existing in the graph convolutional neural network, the application proposes a new regularization method: Drop-Relation. The traditional method often discards single graph nodes or block graph nodes, which cannot prevent the node information from continuing to spread in the graph. Drop-Relation makes the whole relation matrix inactivate, effectively prevents the information in the relation graph from spreading in the network, and can inhibit the dependency between relations, effectively alleviates the overfitting and oversmoothing problems of the graph convolution.

The embodiment discloses an image generation and neural network training method, device, equipment and medium, and the method comprises the steps of obtaining an original image; inputting the original image into a trained generative adversarial network; processing the original image by using the trained generative adversarial network to obtain a first generated image; the training data of the generative adversarial network comprises a sample image set and a second generated image; the second generated image comprises at least part of images of a generated image set; and the generated image set represents generated images obtained by processing the sample image set by using the generative adversarial network.

This invention relates to the field of precipitation fusion technology and discloses a multi-source precipitation data fusion method based on metaheuristics and machine learning, comprising the following steps: collecting ground precipitation observation data, multi-source satellite precipitation product data, and auxiliary variable data, and performing data preprocessing; dividing the preprocessed ground precipitation observation data, multi-source satellite precipitation product data, and auxiliary variable data into training and validation sets according to a certain ratio; setting the initial value range of hyperparameters of a multilayer perceptron model, and inputting them along with the training set into the multilayer perceptron model for training; optimizing the hyperparameters of the multilayer perceptron model using a sparrow search method to obtain a trained multilayer perceptron model; inputting the validation set into the trained multilayer perceptron model to identify the precipitation observation values ​​of the validation set, thereby obtaining the predicted ground precipitation observation data, i.e., the fused precipitation data; this method improves the accuracy of fused precipitation data while avoiding reliance on assumptions that are not valid in reality.