41 results about "Probability vector" patented technology

This application provides a method, system, medium, and device for human-centered video scene reconstruction and separation. The method includes: for a first-person perspective video sequence, initializing a 3D Gaussian set covering the background, hands, and objects based on a priori knowledge of structure recovery from motion; assigning a learnable dynamic category probability vector to each Gaussian point in the 3D Gaussian set; constructing a dedicated deformation branch; according to the learnable dynamic category probability vector of each Gaussian point, assigning each Gaussian point to the dedicated deformation branch for processing through a preset soft-hard two-stage routing mechanism, determining the Gaussian points processed by the dedicated deformation branch; rendering the Gaussian points processed by the dedicated deformation branch to determine the 4D scene reconstruction image and the decomposed reconstruction images of the background, hands, and objects. This application achieves 4D scene reconstruction of human-centered video and explicit, fine-grained separation of the background, hands, and objects.

The application provides a kind of intelligent production line fault prediction method and system based on random forest enhancement, by generating production line equipment association graph containing device node, connection edge and running state archive;Continuously receive real-time state signal stream encapsulated with adaptive time window from sensor array, and dynamically adjust the collection granularity according to the device fault sensitivity;The state stream input is introduced into the random forest enhanced classifier of fault sample distribution weight, and the membership degree probability vector set of each device node belonging to the preset fault mode is output by traversing output step by step decision;Traverse the device node in the graph, and take the fault mode exceeding the confidence threshold as the activated node to perform breadth-first search, record the fault propagation direction sequence and time delay distribution;Finally, the propagation sequence and delay distribution are encoded into fault evolution path graph, the fault arrival probability and expected impact range of propagation path are calculated, and the prediction warning instruction is generated.The application realizes the accurate prediction and propagation tracing of production line fault.

The present disclosure provides a model training method of a joint data set, an X sheet identification method and related equipment. Different categories of data sets are combined during model training, thereby expanding the sample categories. Then, the network parameters of the feature extraction network corresponding to each category of data set are updated using the common loss value of the combined data set, which is beneficial to considering other categories of data sets when the feature extraction network corresponding to different categories of data sets is used, thereby improving the generality and accuracy of the identification model obtained by training. And according to the preset demand set by the user, the probability vector is analyzed to obtain the to-be-identified result of the to-be-identified X sheet. Not only can the to-be-identified X sheet be accurately identified, but also the actual needs of the user can be met, and the flexibility is high.

This invention discloses a method for proactive defense and resilient reconfiguration of distribution networks integrating spatiotemporal graph neural networks, belonging to the field of distribution network cyber-physical security. This invention aims to address the limitations of traditional passive defense methods in responding to spatiotemporally coordinated FDIA attacks and their insufficient reconfiguration resilience. First, a spatiotemporal graph model of the distribution network's CPS is constructed, and a spatiotemporal graph neural network is used to predict the FDIA attack probability vector. Second, an incomplete information stochastic game model is constructed, using the probability vector as the observation input to solve for the optimal defense resource deployment strategy. Finally, a multi-objective resilient reconfiguration model is constructed with the objectives of maximizing source-load spatiotemporal matching and restoring critical loads. A deep reinforcement learning algorithm is used to solve for the optimal switching sequence based on the game situation or physical state. This invention can effectively predict and defend against spatiotemporally coordinated attacks, significantly improving the proactive defense capability and operational resilience of the distribution network.

This application discloses a method and system for comprehensive prediction and grading assessment of food safety risks, mainly relating to the field of comprehensive risk prediction technology. It addresses the problems of conventional data preprocessing methods struggling to construct robust feature representations for multi-source heterogeneous data and the failure of most models to effectively integrate prior domain knowledge. The method includes: averaging the dimensionality-reduced feature matrix over the time dimension to aggregate it into a feature vector, thereby obtaining a Laplace score; using the Laplace score and the dimensionality-reduced feature matrix to obtain an enhanced feature matrix; inputting the enhanced feature matrix into a trained model, weighting and aggregating the temporal features in the enhanced feature matrix to obtain a risk-aware context vector; fusing the risk-aware context vector with a preset external condition feature vector through a modulation gating mechanism to obtain a condition-aware decision feature vector; and performing hierarchical risk classification on the condition-aware decision feature vector to obtain a risk level probability vector.

This invention discloses a classification method for ultra-high carbon steel microcrystals based on spatial attention and ensemble prediction, relating to the field of metallic materials technology. The method includes: acquiring K improved EfficientNet-B7 models; these K improved EfficientNet-B7 models are obtained by training with K-fold cross-validation based on the initial improved EfficientNet-B7 model structure; the initial improved EfficientNet-B7 model adds a spatial attention module between the output of the sixth feature extraction stage and the input of the seventh feature extraction stage; classifying the microcrystal image of ultra-high carbon steel to be classified using each improved EfficientNet-B7 model to obtain the probability distribution of each category; arithmetically averaging the multiple probability distributions predicted by the multiple improved EfficientNet-B7 models to obtain a fused probability vector; and selecting the category corresponding to the maximum probability in the fused probability vector as the category of the microcrystal image to be classified. This method improves the classification accuracy of ultra-high carbon steel microcrystal structures.

The application provides a rail vehicle fault early warning method, device, equipment and storage medium, comprising: collecting current and temperature signals of a low-voltage electrical appliance cabinet power supply loop; sending the current high-frequency component into an arc detection channel, performing time-frequency analysis and outputting an arc fault probability vector through a first neural network model; sending the current low-frequency component and the temperature signal into an overheating detection channel, extracting time sequence features and outputting an overheating fault probability vector through a second neural network model; calculating a fusion weight based on the correlation of the two types of faults, weighting and fusing the two probability vectors to obtain a final fusion probability vector, and determining a vehicle warning level accordingly. The above scheme ensures the recognition accuracy of a single fault type and improves the stability and accuracy of the overall fault determination.

The invention relates to a method for automatic segmentation of a dental arch that comprises acquiring a three-dimensional surface of the dental arch, in order to obtain a three-dimensional representation comprising a set of vertices, generating virtual views from the three-dimensional representation, projecting the three-dimensional representation onto each two-dimensional virtual view, in order to obtain an image representing each vertex on the virtual view, processing each image by means of a deep learning network, carrying out inverse projection of each image in order to assign, to each vertex of the three-dimensional representation, one or more pixels of the images in which the vertex appears and to which it corresponds, and assigning one or more probability vectors to each vertex, determining the class of dental tissue to which each vertex most probably belongs based on the probability vector or vectors.

The application relates to a causal DAG discovery method for an online service system based on fusion of soft priori. The method comprises the following steps: obtaining observation data and text meta-knowledge of the service system, preprocessing to form a standardized sample set, identifying variable types and semantics and outputting; generating a natural language description according to the variable semantics, querying a large language model for an ordered variable pair, analyzing to obtain three types of causal probability vectors, and calibrating to obtain an edge-level priori probability. An appropriate conditional independence test method is selected, high-confidence independent / dependent sentences are divided, and weights are assigned. A candidate directed acyclic graph is selected as an initial structure, parameters are estimated by linear regression, and data fitting scores are calculated, language priori scores, conditional independence penalty terms and counterfactual self-consistency penalty terms are calculated. Fusion is carried out into a hybrid score function, discrete optimization is carried out under the constraint of a directed acyclic graph, and a causal graph structure with the optimal score is output. The method can improve the efficiency and accuracy of a smart operation and maintenance system.

The application discloses a kind of test time self-adaptive target detection methods based on pre-training large model, pre-trained text encoder and image encoder are used to extract the features of input text and image respectively, then input pre-training large model to obtain initial target proposal set, after refining initial target proposal set, the visual label of each refined target proposal is generated, then it is spliced with text features to obtain extended text features, the bounding box of new target proposal is obtained by inputting extended text features and original image features into pre-training large model again, the classification probability vector and confidence based on prediction score are calculated, and the final target proposal set is obtained, finally, post-processing is carried out to obtain the final target detection result.The application realizes fine-grained, context-aware visual-language real-time alignment in the test stage, so as to improve the positioning accuracy and classification robustness of target detection.

The present application relates to a fraud risk prevention and control method and system, wherein the method inputs multi-source heterogeneous data into a pre-trained channel risk assessment model for channel risk assessment, splices the generated channel risk assessment result with the individual feature vector of the user, inputs the obtained risk portrait feature into a Gaussian mixture model for unsupervised clustering to obtain K potential fraud scenarios and a probability vector of each user belonging to each potential fraud scenario. Based on the real-time context feature vector of all users, the risk portrait feature and the probability vector of each user belonging to each potential fraud scenario, fraud risk assessment is carried out, a heterogeneous evidence graph is constructed for fraud reasoning to generate a fraud risk list, and the risk prevention and control feedback data is used for reverse optimization to realize dynamic risk prevention and control. Thus, the present application can quickly respond to the changing new fraud scenarios while realizing the source perception of fraud, and realize accurate fraud risk prevention and control.

This application relates to the field of data processing technology, and discloses a graph structure generation method and a memory computing accelerator based on a memory computing accelerator. The graph structure generation method based on the memory computing accelerator includes: a node generation module calculating the hidden state of the current node and sending it to an edge probability generation module; the edge probability generation module outputting a probability vector based on the hidden state and sending it to a Bernoulli sampling module; the Bernoulli sampling module outputting a decision vector based on the probability vector and sending it to an edge storage module; and the edge storage module receiving the decision vectors corresponding to all nodes and generating a target graph structure based on the decision vectors corresponding to all nodes. This invention can reduce computational complexity, reduce data movement, save storage space, and improve energy efficiency, thereby improving the graph structure generation effect.

The present application relates to a micron-level coating process modeling and precision control method based on digital twinning, comprising the following steps: S1: adopting a hybrid modeling strategy of physical information neural network and intrinsic orthogonal decomposition, constructing a multi-layer composite substrate proxy model; S2: using Monte Carlo Dropout and active learning strategy, quantifying the uncertainty of key parameters of the multi-layer composite substrate proxy model, and obtaining a reduced order model; S3: based on the reduced order model and online sensor data, constructing a dynamic adaptive edge twin, and outputting a predicted film thickness field; S4: according to the predicted film thickness field and the measured film thickness, based on the micro-scale error source sparse decomposition of the graph neural network, the error source probability vector is obtained; S5: taking the error source probability vector as the basis for feedforward compensation, using the feedforward-feedback composite compensation based on robust model predictive control. The present application effectively improves the efficiency and reliability of micron-level coating process control.

This invention discloses a method for reconstructing tunnel stratigraphic structures based on a physical information neural network, belonging to the field of tunnel engineering and geotechnical engineering geological exploration technology. The method includes: acquiring observation data within the tunnel engineering area; constructing a physical information neural network with spatial coordinates as input and lithology probability vectors as output; constructing an objective function containing a data consistency term and a physical regularization term, wherein the physical regularization term applies anisotropic constraints to the spatial gradient of the lithology probability vector, including continuity constraints along the tunnel axis and layer interface preservation constraints along the vertical direction; training the neural network using the observation data by minimizing the objective function; inputting the spatial coordinates into the trained network to obtain the lithology probability vector and determine the stratigraphic lithology prediction result. This invention transforms discrete lithology prediction into a continuous probability field reconstruction problem, achieving stable reconstruction of the stratigraphic structure by constraining the axial continuity and vertical layering of the geological features of the strata through anisotropic structural regularization.

Method and a system for identifying target messages are provided. The method comprises: during a first phase: training a plurality of prediction models to generate respective predictions of whether a given in-use message is a target one or not; generating, based on the respective predictions of the plurality prediction models, respective training consolidated probability vectors for a plurality of training messages; using the respective training consolidated probability vectors, training a decision tree model to determine whether the given in-use message is a target one or not; and during a second stage, following the first phase: using the plurality of prediction models and the decision tree model to classify in-use messages on online platforms; in response to determining that a given in-use message is a target message, causing execution of a remedial action.

This invention relates to an artificial intelligence-based method for identifying the specifications and models of waste cables, belonging to the field of artificial intelligence technology. It includes the following steps: acquiring image data of waste cables and constructing a dataset; processing the core region in the waste cable images to obtain a normalized core-focused image; obtaining a multi-scale texture feature map through adaptive scale selection wavelet transform and directional filter bank transform; constructing a deep neural network based on cable core focusing, including a spatial attention weight map generation module, a channel attention weight vector generation module, a core focusing convolution module, and a prediction result output module; the multi-scale texture feature map is processed by the network to obtain a predicted probability vector; and the deep neural network based on cable core focusing is trained and optimized using a composite loss function that integrates core region alignment constraints and multi-scale discriminative feature structure constraints. This invention can improve the accuracy of waste cable specification and model identification.

The present application provides a kind of based on Markov chain's sandstone and conglomerate source supply capacity quantitative prediction method, comprising: step 1, understand the basic geological profile of study area, carry out Cenozoic standard layer structural diagram interpretation;Step 2, determine the prediction object and initial state probability vector and time series vector, establish Markov prediction model;Step 3, apply statistical estimation method, establish one-step transition probability matrix;Step 4, realize multiple state probability prediction by one-step transition probability matrix;Step 5, based on source supply intensity factor K, establish the functional relationship between source supply intensity factor K and fan body advancing (retreating) distance.The sandstone and conglomerate source supply capacity quantitative prediction method based on Markov chain has high practical value, strong operability, high prediction result accuracy, and has significant progress compared with the traditional gentle slope zone sandstone and conglomerate source area source supply capacity quantitative prediction.

The application discloses a multi-temporal SAR coherent change detection image false alarm suppression method and system, and relates to the technical field of synthetic aperture radar image processing. The method establishes a hidden Markov model suitable for multi-temporal SAR coherent change detection, defines an initial state probability vector, a state transition probability matrix and an emission probability density function; performs radiation calibration, sub-pixel registration and coherent coefficient calculation on multi-temporal SAR images to obtain a coherent coefficient observation sequence; obtains optimal hidden states of each pixel by using a Viterbi algorithm through initialization, recursive updating, termination judgment and backtracking; distinguishes states according to the time sequence characteristics of unchanged areas, target changes and CCD clutter, realizes matching of transient targets and persistent clutter according to state transition probability configuration, and finally removes clutter to complete false alarm suppression. The application can reduce the false alarm rate while maintaining the sensitivity of subtle target detection, and improve the accuracy and robustness of multi-temporal SAR coherent change detection.

The application provides a screen defect detection method, device, equipment and medium, which are used in the technical field of screen defect detection, and include the following steps: cropping a target image to be detected and identifying each image block to obtain multiple defect connected domains and corresponding geometric information of each defect connected domain; weighting and fusing probability vectors to obtain a target probability vector, and determining a target defect category according to a category index corresponding to the target probability vector; obtaining a single defect detection result based on the geometric information, the target defect category and a preset threshold value, and obtaining a bright spot defect detection result based on a corresponding associated dark picture, an associated dust picture and a preset defect threshold value; obtaining a dark spot defect detection result based on multiple color channel images and a preset number threshold value, and determining a screen detection conclusion corresponding to the screen to be detected according to the single defect detection result, the bright spot defect detection result and the dark spot defect detection result; in this way, the efficiency and accuracy of screen defect detection are improved.

The application provides an electrode three-dimensional reconstruction segmentation method and device and electronic equipment. The method comprises: collecting multi-modal data; wherein the multi-modal data comprises morphological data and element data; constructing an implicit neural field model; wherein the input of the implicit neural field model is a spatial three-dimensional coordinate, and the output of the implicit neural field model is a probability vector of the spatial three-dimensional coordinate belonging to each phase; training the implicit neural field model based on a composite loss function of the multi-modal data; wherein the composite loss function comprises a data fidelity loss term and a physical prior loss term; and performing electrode three-dimensional reconstruction based on the trained implicit neural field model. In this way, the conductive agent, the binder and the pore phase with similar gray values in the composite electrode can be accurately segmented through multi-modal data fusion and physical prior constraints, so that the low-contrast image data obtained by using a focused ion beam-scanning electron microscope can be subjected to high-precision three-dimensional reconstruction.

The present application relates to artificial intelligence technology, disclose a kind of text abstract generation method, device, equipment and medium, the method includes: the article to be processed of acquisition is carried out word segmentation processing, constructs dictionary according to word segmentation;The input vector of multiple is obtained by encoding dictionary;Input vector is carried out data enhancement, and the first enhancement data, second enhancement data and third enhancement data corresponding to input vector are obtained;First enhancement data, second enhancement data and third enhancement data are carried out network calculation, and the update vector of input vector is obtained;When the number of data enhancement reaches preset number, update vector is associated with dictionary, and the vector calculation is carried out to the dictionary after association, and the probability vector corresponding to update vector is obtained;Global probability path is constructed according to probability vector, and maximum probability path is obtained according to global probability path, and the abstract of the article to be processed is formed.The present application can improve the efficiency of text abstract generation.

This application provides a multi-granularity electrocardiogram (ECG) diagnostic system, method, device, readable storage medium, and program product. The system includes an ECG signal input module for inputting ECG signals; a backbone network module for outputting the temporal continuous features and multi-lead spatial features corresponding to the ECG signals; a first granularity module for outputting a first granularity probability vector and a first granularity diagnostic label; a second granularity module for determining second granularity features and outputting the second granularity probability vector and the second granularity diagnostic label; and a third granularity module for determining third granularity features and outputting the third granularity diagnostic label corresponding to the third granularity features. The multi-granularity ECG diagnostic system of this application employs a cascaded conditionalization mechanism, enabling significant and balanced performance improvements at each diagnostic level. This solves the problem of conflicting prediction results between different levels in existing parallel multi-head structures, improving the accuracy of multi-granularity ECG diagnosis.

The application relates to the technical field of computers and discloses a pilot training operation automatic evaluation method and system based on rule guidance and artificial intelligence fusion, which comprises the following steps: a probability semantic state machine model is constructed; a benchmark operation rule set constraint is converted into a semantic topological space with a state transition probability matrix; an artificial intelligence recognition algorithm is used to obtain multi-modal features of a target operation behavior of an operation subject, and an observation probability vector sequence representing operation probability distribution at each moment is output; the observation probability vector sequence is mapped to the semantic topological space, a state alignment relative entropy is calculated, an optimal alignment path is recursively searched, and an operation deviation cost is generated, and evaluation indexes are determined according to the operation deviation cost; and the semantic topological space with state transition weights is established, and the fault tolerance of evaluation logic to uncertainty of a perception layer is improved.

This invention relates to the field of land planning technology, specifically to a land use planning optimization method and system based on spatiotemporal big data. A land use planning optimization system based on spatiotemporal big data includes: a historical land use data sequence acquisition module, a change characteristic calculation module, and an iterative optimization module. This invention acquires multi-temporal land use remote sensing raster map sequences of the land to be planned, constructs a pixel-level land use change dataset, comprehensively characterizes the land use evolution trends and change sensitivity of each location at different historical stages, and ensures that the model can identify the spatial distribution characteristics of stable areas and high-frequency change areas. Based on land type change statistics, a global transition matrix and trend probability vector are constructed to achieve quantitative modeling of land use evolution patterns, and this is used for trend-guided construction of the initial population of a genetic algorithm, improving the directionality and convergence efficiency of the optimization.

The present invention relates to a Bernoulli sampling-based interpretable CNN training method and apparatus, and a medium. The method comprises the following steps: inputting an image into a CNN to obtain response feature maps of filters; for the response feature maps, performing Bernoulli sampling to obtain a binarized assignment matrix; calculating average weight matrices of the filters for image categories on the basis of the binarized assignment matrix, and calculating the sum of pairwise differences; calculating a Hadamard product of an assignment vector of the binarized assignment matrix and the response feature maps to obtain masked feature maps; separately inputting the response feature maps and the masked feature maps into a CNN fully-connected layer to respectively obtain classification prediction probability vectors, and respectively calculating cross entropy losses between the prediction results and a ground-truth label; and on the basis of the sum of the pairwise differences, and the cross entropy losses, using a stochastic gradient descent method to implement network training to obtain an interpretable CNN for image classification. Compared with the prior art, the present invention has the advantages such as high adaptability and high interpretability.

This application relates to the field of artificial intelligence technology. It provides a method, system, device, and medium for intelligent screening of financial anomaly data. The method includes: performing multi-layer reinforcement learning anomaly detection based on heterogeneous financial graphs to obtain node-level anomaly probability vectors; generating a graph evolution operation instruction set based on the node-level anomaly probability vectors and the heterogeneous financial graphs; generating adversarial subgraphs based on the graph evolution operation instruction set and the heterogeneous financial graphs; implementing graph reconstruction defense on the adversarial subgraphs to obtain defense-enhanced graph structure data; performing meta-learning adaptation based on the defense-enhanced graph structure data to obtain updated meta-controller parameters; and performing fusion decision-making on the node-level anomaly probability vectors and the graph evolution operation instruction set to output anomaly group labels. This aims to improve adversarial attack defense capabilities, reduce response time for new fraud patterns, and enhance the accuracy of identifying decentralized fraud groups.

The application relates to the technical field of knowledge graph reasoning, in particular to a user intention reasoning system based on a knowledge graph. The system obtains user business state data and generates a state vector and an intention validity vector, constructs a heterogeneous graph model containing state nodes and intention nodes, analyzes an excitation weight based on a fault level and a historical business handling time length, and obtains an intention activation probability vector based on a business background through a random walk algorithm; a confidence factor is dynamically adjusted based on the amount of valid information in the text, the text semantics and the business background reasoning result are fused, and a hierarchical strategy is executed according to the intention comprehensive evaluation index after fusion. The application converts business background common sense into a calculable probability feature by constructing a state-intention heterogeneous graph and introducing a dynamic fusion mechanism, effectively reduces the failure problem of NLP technology in the semantic fuzzy and multi-state coupling scene, and improves the intention recognition accuracy and service efficiency in the complex business scene.

The application discloses a multilingual machine translation low-rank adaptive sharing strategy learning method and system, relates to the technical field of multilingual neural machine translation, and comprises the following steps: constructing a LoRA architecture in a large language model; defining a language gate vector and performing normalization to obtain a probability vector; generating a mask through Bernoulli sampling based on the probability vector; introducing the mask between two linear transformations of the LoRA architecture to calculate a final output result '; performing approximate gradient calculation through a pass-through estimator to obtain the gradient of the probability vector; training and reasoning the large language model; evaluating the multilingual translation performance of the large language model; dynamically learning and optimizing a cross-language sharing strategy of the LoRA rank, effectively relieving language knowledge interference, and improving the translation performance of the model.

The application discloses a kind of based on the multi-dimensional early warning method of slope of LSTM neural network, it is related to geotechnical engineering technical field, the application includes collection data, data is preprocessed, forms time series dataset;LSTM neural network model is constructed and trained, with time series dataset as input, output each original probability vector of early warning level;Real-time calculation surface displacement and internal displacement rate and acceleration characteristics, construct deformation evolution stage discriminant factor, generate adaptive weight vector according to discriminant factor;Original probability vector and current normalized displacement value output by LSTM model are weighted and integrated using adaptive weight vector, obtain comprehensive risk score;Comprehensive risk score is compared with preset multilevel threshold, trigger corresponding level early warning and publish information.The application extracts multi-source time series characteristics by LSTM neural network, and adaptively generates weight in combination with deformation evolution stage identification mechanism, can continuously adjust the contribution of each parameter according to freeze-thaw evolution process.

A computer-implemented method for classifying manoeuvres performed by an aircraft, including: acquiring a data structure including at least one unknown data matrix including a plurality of time series of samples of quantities related to the flight of the aircraft, the samples being relative to a succession of instants of time; applying to the unknown data matrix a neural network generating a corresponding probability matrix including, for each instant of time of the succession of instants of time, a corresponding probability vector including, for each class of a plurality of classes of manoeuvres, a corresponding estimate of the probability that, in the instant of time, the aircraft has performed a manoeuvre belonging to the class; and selecting, for each instant of time of the succession of instants of time, a corresponding class of manoeuvres, based on the probability estimates of the corresponding probability vector.