70 results about "Probit model" patented technology

The invention discloses a physical information neural network-driven ocean island typhoon scene generation method and system. The method comprises the steps of collecting multi-source heterogeneous meteorological data and performing space-time alignment preprocessing; constructing a coarse-scale space-time probability prediction model, capturing space correlation of meteorological elements by using a graph topology learning network, efficiently processing long-time-sequence dependence of typhoon evolution by integrating a state space model with linear complexity, and generating a probabilistic typhoon scene with coarse resolution through a multivariable joint distribution probability model; further constructing a physical downscaling model, taking a coarse-scale prediction result as condition input, and performing physical consistency downscaling on a coarse-scale scene by embedding an atmospheric fluid mechanics equation in a loss function as a physical hard constraint; and finally, outputting a high-resolution typhoon scene with probability reliability and physical authenticity.

The invention relates to a lithofacies intelligent pattern classification method based on a conditional diffusion probability model, and belongs to the technical field of deep learning, pattern recognition and big data processing. Comprising the following steps: step (1), collecting and preprocessing big data; step (2), constructing a conditional diffusion probability model; step (3), model training and parameter optimization; step (4), generating a category balance sample; (5) enhancing data quality evaluation; and (6) training and verifying the pattern classification model. According to the method, shale lithofacies data enhancement and intelligent mode classification based on the conditional diffusion probability model and the deep learning technology are realized, the lithofacies identification problem under the condition of training set category imbalance is effectively solved, and the minority class lithofacies identification precision and the overall classification performance of the model are remarkably improved; and a reliable technical method is provided for marine shale oil and gas reservoir evaluation and sweet spot prediction.

The invention provides a soil moisture sensor data quality inspection and interpolation method, which comprises the following steps: screening multivariable soil moisture time sequence data based on a preset core physical feature list, carrying out abnormal value detection through physical rule constraint and an isolation forest algorithm, and carrying out data labeling by creating a complete time axis; generating a training sample from the preprocessed data through sliding window sampling, and performing deep feature learning by using a denoising network based on a space-time diffusion probability model; performing interpolation on missing values in the original data by using the trained space-time diffusion probability model, generating a noisy data sample through a forward noise adding process, and performing conditional data interpolation based on a known observation value and a mask matrix in a reverse denoising process to generate a preliminary interpolation result; and performing post-processing correction on the preliminary interpolation result, wherein the post-processing correction comprises clamping correction based on a monthly historical range and correction based on interlayer physical logic.

This application provides a runoff prediction method, apparatus, electronic device, and computer-readable storage medium. The method includes: acquiring the forecast meteorological time series of a target watershed during the prediction period, historical meteorological time series, and historical runoff time series for historical periods; inputting the historical meteorological time series and historical runoff time series into an attention model to extract global contextual features of the target watershed; inputting the forecast meteorological time series, global contextual features, and initial noise data into a conditional diffusion probability model, performing multiple backdiffusion processes to obtain multiple predicted runoff time series of the target watershed during the prediction period; and calculating a specified quantile for each moment in the prediction period based on the multiple predicted runoff time series to construct a confidence interval, thereby obtaining runoff prediction information containing a risk probability distribution. This method avoids gradient vanishing when processing long-sequence data and outputs the probability distribution of the prediction results.

A method for dependency task prioritization is disclosed. The method includes providing base data including task identifiers, failure values, and dependency data values, where each dependency data value is associated with a pair of tasks. The method also includes generating a probabilistic model using a directed acyclic graph, where each task is associated with a network node and each dependency data value is associated with a network edge of the directed acyclic graph, where each dependency data value represents a conditional probability related to a respective per-time-unit failure rate. Additionally, the method includes determining, for each task in the directed acyclic graph, a posterior marginal probability value representing a probability of the task experiencing a failure, and selecting, based on the posterior marginal probability values, a sequence of tasks that are most likely to fail the fastest.

The invention discloses a method and a system for predicting migration and blockage of fracture particles based on machine learning, and the method comprises the following steps: firstly, presetting fracture roughness coefficient (JRC), flow velocity, equivalent particle size of particles and particle quantity data, constructing a fracture model with a real shape, and carrying out a visual migration test; recording fracture blockage label data; inputting the parameters and the labels into a neural network dichotomy probability model taking a multi-layer perceptron as a core, and training to obtain a blockage probability prediction function; on the basis of the trained model, sensitivity analysis and feature importance evaluation are carried out, the influence sequence of the blockage probability on JRC, the flow velocity, the equivalent particle size of particles and the particle number is output, and prediction of the particle blockage event in the fracture is achieved; according to the invention, the particle blocking mechanism in the crack under the multi-factor coupling effect is researched, and a scientific basis is provided for design and construction of geotechnical engineering.

The invention relates to the technical field of day-ahead electricity price prediction, in particular to a day-ahead electricity price prediction method, device and equipment based on a diffusion model and a computer readable storage medium, and the method comprises the steps: obtaining day-ahead market transaction data, real-time transaction data and auxiliary information in a fixed time period every day, and writing the data into a market information database; based on the market information database, aligning the market information database according to unified time granularity to obtain a multi-dimensional condition feature sequence; inputting the multi-dimensional condition feature sequence into the trained de-noising diffusion probability model, and generating an electricity price prediction sample set at discrete time points in the next day through reverse multi-step Markov de-noising sampling; and obtaining a statistical value of the electricity price prediction sample set at each time point, constructing a prediction curve, taking a preset quantile to obtain a probability interval, and writing a result into prediction data and an evaluation database. Probabilistic generation is carried out on the day-ahead electricity price by adopting a diffusion model, and the prediction precision in an extreme fluctuation scene is remarkably improved.

This invention discloses a spacecraft pose estimation and uncertainty modeling method based on intrinsic space, belonging to the field of spacecraft pose estimation technology. The method involves directly constructing a hierarchical Bayesian probability model on the rotating manifold SO(3) and translation space, using Fisher and Gaussian distributions respectively, and introducing conjugate priors to achieve the fundamental decomposition and quantification of accidental and cognitive uncertainties. An end-to-end multi-task neural network is used to jointly learn the pose probability model parameters, key points, and segmentation information, and an iterative optimization module is employed to improve estimation accuracy. During the training phase, marginal negative log-likelihood and evidence regularization are jointly optimized; during the inference phase, the probability distribution of pose prediction is obtained through analytical marginalization, and the two types of uncertainty are distinguished. This invention improves pose estimation accuracy while outputting well-calibrated uncertainties, providing a reliable basis for the autonomous and safe operation of spacecraft in orbit.

The invention discloses an adjustment capability evaluation and improvement method, system and device based on new energy and energy storage cooperation and a medium. The adjustment capability evaluation and improvement method comprises the following steps: constructing a probability density function and a confidence interval of adjustment capacity based on a new energy output probability prediction model; a probability density function of capacity adjustment is remodeled through an energy storage cooperation strategy, and a confidence interval is narrowed; constructing a joint probability model of the adjustment instruction and the adjustment capacity; and deriving the probability distribution of the adjustment precision and the shortest high-density confidence interval according to the joint probability model. According to the method, the probability distribution of the new energy adjustment capacity is actively remodeled and optimized by using the energy storage system, the confidence interval width of the adjustment capacity is effectively narrowed under the same confidence level, and the improvement effect of energy storage on the new energy adjustment precision is accurately quantified.

The invention discloses an underwater explosion shock wave load probability characterization method based on Bayesian inference, and relates to the field of underwater explosion load calculation, and the method comprises the steps: obtaining previous test data, carrying out the analysis processing of the test data, obtaining sample data, taking a Cole empirical model as a prior model of Bayesian inference, and carrying out the calculation of the probability of the underwater explosion shock wave load. Performing uncertainty analysis on load characterization parameters and calculation errors of the prior model in combination with the sample data to obtain target probability distribution of the load characterization parameters and the calculation errors; and by taking the target probability distribution as priori knowledge, carrying out updating calculation on the load characterization parameters and calculation errors by adopting a Bayesian inference method to obtain a Bayesian probability model of the underwater explosion shock wave load. And carrying out probability representation on the underwater explosion shock wave load based on the Bayesian probability model of the underwater explosion shock wave load. According to the method, the uncertainty of the underwater explosion shock wave load is effectively represented, and random input considering the load variability is provided for the anti-explosion reliability design of the underwater structure.

The invention discloses a new energy power generation prediction error analysis method and system considering meteorological conditions, and belongs to the technical field of computer data processing and prediction.The new energy power generation prediction error analysis method includes the steps that new energy power generation historical data, prediction error data and meteorological data are obtained and preprocessed, an initial data set is generated, and the statistical magnitude of prediction errors is calculated; combining the meteorological data in the initial data set, using a kernel density estimation method to estimate the joint probability density of the meteorological data and the prediction error data, generating joint probability density distribution, using a Bayesian formula to calculate the conditional probability distribution of the prediction error under a preset meteorological condition, and generating a conditional probability model; and performing multi-dimensional conditional probability modeling on the conditional probability model based on the climate and the position to generate a multi-dimensional conditional probability model. According to the method, kernel density estimation and the Bayesian theory are combined, and multi-dimensional space-time factors are fused to carry out refined modeling, so that the uncertainty of new energy power generation prediction can be accurately quantified, and prospective risk early warning can be realized.

This invention discloses a structural topology optimization design method considering the uncertainty of multi-peak loads, belonging to the field of uncertain structural optimization design. It mainly includes three parts: establishing a multi-peak load uncertainty model, solving for the random response, and solving for the optimization formula. This invention describes load uncertainty using a Gaussian mixture model, solves for the Gaussian mixture model coefficients based on the EM algorithm, and establishes a multi-peak load distribution probability model. By decorrelating random variables, sparse grid technology is used to solve for the mean, standard deviation, and sensitivity of the response, thereby solving for the topology optimization model considering the multi-peak load uncertainty. This method addresses the problem of low structural product reliability that may occur in structural designs considering multi-peak load uncertainty by establishing an accurate probabilistic model of multi-peak load uncertainty. This method is simple and easy to implement, readily applicable in engineering, and can improve the design efficiency of structural designers considering complex load conditions.

This work classification estimation method includes: a step for acquiring the likelihood of a candidate classification from a likelihood model that is a machine learning model; a step for acquiring the appearance probability of the candidate classification from a probability model that is a statistical model; a step for estimating the classification of work executed by a caregiver, on the basis of the likelihood and the appearance probability of the candidate classification; and a step for outputting the results of estimating the classification of the work.

The present application relates to the technical field of network security, and particularly relates to a network intrusion detection adversarial sample defense method and device, which comprises the following steps: extracting features of input network traffic to obtain traffic feature vectors of clean samples, which are then used to train a denoising diffusion probability model to obtain a target denoising diffusion probability model; inputting the traffic feature vectors of multiple clean samples into the target denoising diffusion probability model for reconstruction, outputting reconstruction loss corresponding to each clean sample, and determining an adversarial detection threshold based on the statistical distribution of the reconstruction loss corresponding to each clean sample; inputting the traffic feature vector of a to-be-detected sample into the target denoising diffusion probability model to output the reconstruction loss of the to-be-detected sample; determining whether the reconstruction loss of the to-be-detected sample is greater than the adversarial detection threshold, and if yes, determining that the to-be-detected sample is an adversarial sample, otherwise, determining that the to-be-detected sample is a clean sample; and performing adversarial detection based on the denoising diffusion probability model to effectively recover the sample features affected by adversarial perturbations.

The invention relates to a Monte Carlo sampling-based data asset income distribution method. The method comprises the following steps of S1, establishing a data asset income distribution model; s2, establishing a participant selection probability model; s3, establishing a dynamic income distribution scheme; step S4, carrying out Monte Carlo sampling; step S5, estimating a Shapley value; step S6, carrying out service distribution; and S7, updating the model. The method has the advantages that the thought of the Shapley value is adopted to carry out data asset income distribution, the mutual influence between related parties is fully considered, the limitation that the service contribution degree is evaluated only according to the data size is avoided, and the accuracy and fairness of service distribution are improved.

This invention discloses a machine learning-based method and system for predicting particle migration and blockage in fractures. The method includes the following steps: First, preset data on fracture roughness coefficient (JRC), flow velocity, equivalent particle size, and particle number; construct a fracture model with realistic shape and conduct a visualization migration experiment; record fracture blockage label data; input the parameters and labels into a neural network binary classification probability model with a multilayer perceptron as its core, and obtain a blockage probability prediction function through training; based on the trained model, perform sensitivity analysis and feature importance evaluation, and output the ranking of the impact of blockage probability on JRC, flow velocity, equivalent particle size, and particle number, thereby achieving the prediction of particle blockage events within fractures. This invention studies the particle blockage mechanism inside fractures under the coupling effect of multiple factors, providing a scientific basis for the design and construction of geotechnical engineering.

A multi-view user and entity behavior analysis (UEBA) system ("system") constructs and maintains interchangeable modules for predicting the likelihood of anomalous user behavior at the range of organized participants (i.e., users or entities) over a time period. Each module includes a probabilistic model and / or a machine learning model as a sub-module that models participant behavior at various granularity levels with respect to the use of a software as a service application. The system generates anomaly scores by decorrelating the likelihoods output by each sub-module, and uses the anomaly scores to monitor and perform corrective actions based on anomalous participant behavior to maintain a security situation across organizations.

The invention relates to an index construction method for representing sensitivity of a multi-dimensional joint probability model. The index construction method comprises the following steps: step 1, carrying out generalized extreme value distribution edge fitting on each single-variable sample; according to the edge distribution, T-year once-encounter extreme values in different return periods are calculated; converting the sample into a [0, 1] standard probability space; 2, constructing a three-dimensional joint probability model; 3, based on each three-dimensional joint probability model, generating three-dimensional environment contour surfaces in different return periods by adopting an inverse first-order reliability method to obtain a joint extreme load combination; and 4, constructing a composite disaster index based on a univariate extreme value and a combined load combination. According to the method, the problem of significant uncertainty of an existing multi-dimensional joint probability model in extreme marine environment load inference can be solved, and the defect that different models are lack of unified sensitivity evaluation indexes is overcome. And a stable and reliable basis is provided for environmental load selection of structures such as offshore wind power and offshore oil and gas platforms.

The application discloses a physical guidance-based spatio-temporal hypergraph condition denoising diffusion probability model construction method, and relates to the field of ecological environment modeling and artificial intelligence prediction. The application innovatively combines the physical heat diffusion process and the spatio-temporal modeling of multi-modal environmental factors by introducing a temperature driving module based on heat conduction and a spatio-temporal perception fusion module. In addition, a lightweight convolution block attention module and a hierarchical hypergraph attention mechanism are embedded in the diffusion denoising network to realize high-order correlation modeling and dynamic fusion of multi-scale spatial features. The method can effectively capture the spatio-temporal evolution characteristics and ecological dependence of species distribution while maintaining reasonable computational complexity, thereby significantly improving the accuracy of distribution prediction. The experimental results on different types of data sets prove the superiority and robustness of the proposed PSTH-CDPM.

The invention discloses a three-dimensional model generation method, electronic equipment and a storage medium. The method comprises the following steps: respectively training a denoising diffusion probability model used for coupling input expected generation information into rough features of a three-dimensional model and a detail prediction model used for predicting detail features of the three-dimensional model; inputting expected generation information to the de-noising diffusion probability model to obtain rough feature parameters of a target three-dimensional model; inputting the rough feature parameters into the detail prediction model to obtain detail feature parameters of the target three-dimensional model; and generating a target three-dimensional model according to the rough feature parameters and the detail feature parameters. In the process, rough feature parameters of the three-dimensional model can be adjusted and refined according to expected generation information, and then matching and controllability of a final generation result of the three-dimensional model are achieved.

Randomization methods for healthcare scheduling optimization using perioperative stages. Poor scheduling of surgical appointments and procedures in operating rooms can lead to unnecessary downtime, and therefore loss of efficiency. The randomization methods include various probability models, Monte Carlo simulations, and stochastic optimization is used to optimize procedure scheduling in operating rooms. The optimized schedule may be based on estimated procedure duration, estimated turn-around-time, estimated cancellation frequency, forecasted emergency operating room usage, estimated surgeon utilization, and hospital site configuration. A probabilistic machine learning model may be trained based on historic data and ongoing performance data to automate the optimization process and increase accuracy based on up-to-date information and statistics.

The invention discloses a spacecraft pose estimation and uncertainty modeling method based on an intrinsic space, and belongs to the technical field of spacecraft pose estimation. The method comprises the following steps: directly constructing a hierarchical Bayesian probability model on a rotating manifold SO (3) and a translation space, respectively using Feisnow distribution and Gaussian distribution, and introducing conjugate prior to realize principle decomposition and quantification of accidental uncertainty and cognitive uncertainty. Through an end-to-end multi-task neural network, pose probability model parameters, key points and segmentation information are jointly learned, and an iterative optimization module is adopted to improve the estimation precision. In the training stage, joint optimization is carried out through edge negative logarithm likelihood and evidence regularization; in the reasoning stage, probability distribution of pose prediction is obtained by analyzing marginalization, and two types of uncertainty are distinguished. According to the method, the pose estimation precision is improved, meanwhile, the uncertainty of good calibration can be output, and a reliable basis is provided for on-orbit autonomous safe operation of a spacecraft.

A method of estimating the probability of a seizure in a subject, the method comprising: receiving historical data associated with epileptic events experienced by the subject over a first time period, the historical data comprising physiological data associated with each epileptic event and a time at which each epileptic event occurred; generating a temporal probability model of future epileptic events based on the time of each of the epileptic events, the temporal probability model representing a probability of a future seizure occurrence in each of a plurality of time windows; generating a probabilistic model based on the physiological data associated with each epileptic event; weighting the probabilistic model based on the temporal probability model to generate a weighted probabilistic model of future seizure activity; and outputting an estimate of seizure probability in the subject using the weighted probabilistic model.

A method for sequential point cloud forecasting is described. The method includes training a vector-quantized conditional variational autoencoder (VQ-CVAE) framework to map an output to a closest vector in a discrete latent space to obtain a future latent space. The method also includes outputting, by a trained VQ-CVAE, a categorical distribution of a probability of V vectors in a discrete latent space in response to an input previously sampled latent space and past point cloud sequences. The method further includes sampling an inferred future latent space from the categorical distribution of the probability of the V vectors in the discrete latent space. The method also includes predicting a future point cloud sequence according to the inferred future latent space and the past point cloud sequences. The method further includes denoising, by a denoising diffusion probabilistic model (DDPM), the predicted future point cloud sequences according to an added noise.

This invention provides a method and system for patient clustering and survival risk prediction based on a deep probabilistic model, belonging to the field of survival analysis technology. The invention first acquires user health data under right censoring conditions and preprocesses this data. Then, based on a feature extractor within a multi-task learning framework, it extracts feature representations of the preprocessed right-censored user health data. Based on these feature representations, a deep probabilistic model is used to obtain data distribution information, which is then reconstructed and optimized. Next, without considering parameter assumptions, a category-level risk function is estimated based on a survival prediction multi-task sub-network, and individual risk estimation results are obtained based on this risk function. Finally, based on the optimized data distribution information and the variational evidence lower bound of the individual risk estimation results, clustering prediction results and event time prediction results are obtained. This invention improves the accuracy of clustering and event time prediction results.

The invention discloses a Bayesian neural network prior construction method based on feature tag data, and relates to the technical field of machine learning, and the method comprises the steps: carrying out the standardized preprocessing of input feature tag data, calculating a covariance matrix between features, and adding a regularization term to guarantee the numerical stability; constructing structured prior distribution of a first layer of weight of the network by using the covariance matrix, and fusing feature correlation information into a covariance structure of weight parameters through a Kronecker product; and on the basis, a complete Bayesian neural network probability model is established, posterior distribution approximation is performed by adopting a variational inference method, and finally a prediction result is obtained through sampling. According to the method, the adaptive prior distribution is constructed by using the covariance structure in the feature tag data, so that the complex relevance between the features can be effectively captured, and the structural information is fused into the regularization constraint of the model parameters.

A system and method for automatic construction of probabilistic deep neural network (DNN) architectures are provided. The framework of the present invention automatically explores the most relevant probabilistic modes of the basis of a dataset for variational Bayesian inference. The present invention provides a method for using heterogeneous, irregular, and mismatched beliefs in probabilistic sampling for intermediate representations in DNNs, with the ability of automatic tuning mechanisms of posterior, prior, and likelihood models to enable accurate generative and uncertainty models for machine learning tasks. The system further enables an adjustable discrepancy measure to regularize the intermediate representations by variants of divergence metrics, including Renyi's alpha, beta, and gamma divergence. The present invention enables diverse mixed combinations of probabilistic models for misspecified and unspecified probabilistic relationships in an automatic manner. Thus, the representation capabilities of variational autoencoders, variational information bottlenecks, denoising diffusion probability models, and other probabilistic DNNs are improved.

Parameters of a probabilistic model comprising Gaussian processes are learned. A system (100) is disclosed for learning a set of parameters of a probabilistic model having layers of multiple Gaussian processes (e.g., deep GPs) from a training data set. The set of parameters includes at least induced locations for the multiple Gaussian processes, and parameters of a probability distribution approximating outputs of the multiple Gaussian processes at the multiple induced locations. The probability distribution comprises a multivariate normal probability distribution having a covariance matrix defined by a Kronecker product of a first matrix indicating similarities between the multiple Gaussian processes and a second matrix indicating similarities between the multiple induced locations. A system (200) is also disclosed for determining one or more samples of an output of the probabilistic model for a given input using the set of parameters, e.g., to determine a mean and / or uncertainty estimate for the probabilistic model.

The invention relates to the technical field of building structure safety monitoring and risk assessment, in particular to a building structure safety assessment method and system based on artificial intelligence, and the method comprises the steps: firstly obtaining multi-source monitoring data of a target building in real time, automatically recognizing an abnormal bump moment when the settlement rate suddenly changes, and extracting subsequent data as core modeling data; then, prior distribution hyper-parameters are determined based on historical data before salient points, and a hierarchical Bayesian probability model is constructed in combination with core data; then, calculating to obtain posterior probability distribution of a state after building sudden change by using the model, and evaluating a current safety state level and confidence based on national specifications; then, performing forward probability simulation in combination with the current state and posterior distribution, and generating future safe service life probability distribution expressed in a risk time window form; and finally, outputting graded early warning and optimization maintenance decision suggestions. According to the invention, evaluation from abnormal intelligent perception and uncertainty quantification to risk prospective decision making is realized.

The application belongs to the cross technical field of artificial intelligence and marine weather prediction, and discloses a sea wave probability prediction method and system. The method obtains historical wind field data and sea wave spectrum data of a target sea area, pre-processes and organizes the data, and constructs a training data set. A hybrid expert probability model is constructed, which includes a gating network and multiple expert networks. The hybrid expert probability model is trained using the training data set. The trained model receives input wind field data and outputs a hybrid probability distribution through the synergistic effect of the gating network and the expert networks. The hybrid probability distribution is sampled to obtain a predicted sea wave spectrum set, realizing sea wave risk probability prediction. The application can obtain complete distribution information through one forward calculation, does not require a numerical mode set, significantly reduces the computing power and energy consumption overhead, and facilitates high-frequency updating and quasi-real-time business application in resource-limited environments such as shipborne, buoy and near-shore sites.