11 results about "Power iteration" patented technology

The present application relates to the technical field of data processing, in particular to a kind of spectrum radius feature extraction and Hamiltonity determination method for mixed graph structure.It includes the following steps: obtaining mixed graph data and constructing adjacency matrix, Laplacian matrix and unsigned Laplacian matrix;Three spectrum radii are extracted using power iteration acceleration algorithm, and a spectrum radius pyramid is constructed by multi-layer coarsening, forming a multi-scale spectrum feature vector;The feature vector is input into the classification model to output the Hamiltonity probability value;The confidence score is calculated by multiple random dropout, and it is dynamically decided whether to activate the accurate determination engine;Finally, the determination result is output.The present application combines multi-scale spectrum features and statistical features, uses graph neural networks to achieve high-precision probability prediction, and improves the reliability of boundary samples through confidence evaluation and cascade determination architecture, suitable for graph data analysis in the fields of communication networks and bioinformatics.

The invention discloses a Ritchey-Commann surface shape detection method and system based on two-angle Power iterative fitting. Belongs to the technical field of optical detection and particularly relates to the technical field of large-aperture plane mirror surface shape error detection. The objective of the invention is to efficiently process detection data of Ritchey, Common and Miwn angles and calculate an accurate and reliable real surface shape of a plane mirror. Under the condition that the defocusing error amount is unknown, the optimal Power coefficient is calculated automatically and iteratively, the detection operation process is effectively simplified, the dependence of data processing on accurate measurement of the length value of the system is reduced, and meanwhile, the precision of plane mirror surface shape recovery solution is ensured.

The invention belongs to the technical field of non-linear dynamic numerical calculation and rigid differential equation solution, and discloses a non-matrix rigid self-adaptive non-linear dynamic numerical calculation method and a non-matrix rigid self-adaptive non-linear dynamic numerical calculation system. According to the method, equivalent product calculation of the Jacobian matrix and the vector is realized based on the directional derivative through a matrix-free calculation framework, and a complete Jacobian matrix does not need to be stored; high-efficiency and high-precision estimation of the spectral radius is realized through quantum annealing driven Krofft power iteration; constructing an adaptive mixing precision calculation engine based on a spectral radius result to reduce memory occupation; and realizing stable solution of the extreme rigid system through rigid self-adaptive step length control. According to the method, the solution convergence rate and the calculation efficiency of the extreme rigid nonlinear system can be effectively improved, the memory occupation and the long-term integral error are greatly reduced, and the method can be widely applied to the fields of aerospace, mechanical engineering, intelligent equipment, chemical reaction kinetics and the like.

The invention discloses a frequency domain neutron noise problem subcritical degree diagnosis and parameter selection method, and belongs to the nuclear reactor physics field, and the method comprises the following steps: 1, modeling: based on reactor parameters, inputting reactor core geometry, cross section, material and kinetic parameters, a phase shift method and a pseudo-section method (different hyper-parameters) are adopted to construct neutron noise characteristic value problem physical models of different frequencies; 2, calculation: scanning calculation is carried out under different methods and frequencies, and the subcriticality set by different calculations under different frequencies is calculated through a power iteration method; 3, judging: sequentially judging subcritical states / supercritical states at different frequencies according to criticality, excluding supercritical calculation settings, and preferentially selecting the calculation settings with lower subcriticality; and 4, application: performing normal noise calculation by using the optimized calculation setting. According to the method, method selection and parameter optimization basis are provided for neutron noise calculation in different frequency ranges, the calculation power consumption of a reactor core noise analysis database is greatly reduced, and reactor safety diagnosis and operation state monitoring are supported.

The application provides a weakly supervised video anomaly detection method and system based on a multi-head spectral residual gating. The method comprises the following steps: segmenting an input video, and extracting features by using a pre-trained feature extractor; dividing a global feature sequence into a predetermined number of subspace heads along a channel dimension; constructing a local autocorrelation matrix, and estimating a background principal component direction corresponding to each subspace by using a differentiable power iteration algorithm; inputting a spectral residual feature into a semantic perception gating network to generate a mask, and enhancing the global feature sequence by using the mask to obtain an enhanced output feature. In the feature manifold, the application models a normal background as a low-rank principal component subspace, and models an anomaly as a sparse and high-energy spectral residual, so that the semantic decoupling of the background and the anomaly is effectively realized.

This invention discloses a method, apparatus, and medium for multi-scale wavelet analysis of brain networks based on manifold learning. Using T1-weighted MRI and DW-MRI images, combined with Desctrieux mapping and probabilistic fiber tractography based on surface seeds, an initial adjacency matrix is ​​obtained. The average adjacency matrix is ​​calculated. Based on the node degree, betweenness, PageRank, and assignment coefficient of the average adjacency matrix, several nodes are selected from the brain network. Masks at different scales are calculated on the nodes. Multi-scale wavelets are initialized. The eigenvectors of the Laplacian matrix of the average adjacency matrix are solved using a power iteration method to obtain the optimal multi-scale wavelet. Protein signals in the brain network are projected onto this wavelet to obtain new biomarker signals from the brain network. Therefore, this embodiment of the invention uses manifold learning to calculate the mean of a brain network group, which better maintains the geometric topology of the network and, considering the hierarchical modularity and centrality of network nodes, better uncovers some potential physiological and pathological mechanisms in brain diseases.

Embodiments of the present application provide a beamforming method, network device, apparatus and storage medium, wherein the method comprises: determining a third matrix corresponding to a target user equipment based on a first matrix corresponding to the target user equipment and a second matrix corresponding to the target user equipment in paired users; the first matrix is a channel estimation matrix corresponding to the target user equipment, the second matrix is determined based on power iteration vectors corresponding to multiple data streams of the target user equipment, and the third matrix is an equivalent channel matrix corresponding to the target user equipment; performing orthogonalization processing on the third matrix corresponding to the target user equipment based on third matrices corresponding to other user equipments except the target user equipment in the paired users; updating the second matrix corresponding to the target user equipment based on the third matrix after the orthogonalization processing; iterating the foregoing steps, and determining a beamforming factor of multiple data streams of the target user equipment and performing beamforming on the target user equipment when a preset iteration number is reached.

This application provides a resource scheduling method, apparatus, electronic device, and readable storage medium. The method is applied to a target device, which is any device in a distributed multiprocessor system, and includes: acquiring multiple influencing factor variables of the target device; using these variables to characterize the operating cost of the target device; determining the cost function of the target device and its Newton descent direction; the Newton descent direction being the direction in which the cost function converges fastest during iteration; performing multiple iterations until convergence based on the updated influencing factor variables, the Newton descent direction, a preset computing power allocation constraint range, and a preset computing power iteration formula; and using the convergence result as the computing power allocation for the target device to complete computing power resource scheduling based on the allocation. Thus, by using the Newton descent direction during multiple iterations, the convergence speed of the iterations can be improved, thereby increasing the resource scheduling speed.

The application discloses a method for designing inter-pulse waveform amplitude and phase agility against range folding clutter, which comprises the following steps: firstly, under the amplitude and phase agility airborne radar system, a slow-time echo model of a target, clutter and noise is established, and mathematical expressions of target sidelobe and clutter energy are derived; then, considering the constraints of transmitting power and peak-to-average power ratio (PAR), an inter-pulse amplitude and phase agility waveform optimization design problem based on minimizing target sidelobe and clutter energy is established; finally, the accelerated power method of least squares iteration (PMLI) algorithm is used to optimize the inter-pulse amplitude and phase parameters, adjust the clutter energy distribution of the airborne radar, improve the signal-to-clutter-and-noise ratio of a long-distance target under the near-distance folding clutter of the airborne radar, and complete the solution of the optimization problem. The method can suppress the clutter base in the region of interest, realize the target detection against the range folding clutter, and has good robustness and strong realizability without relying on the clutter prior knowledge such as the clutter covariance matrix, and has certain generalization for the clutter suppression of different terrains.

The invention provides a multi-head spectrum residual gating-based weak supervision video anomaly detection method and system, and the method comprises the steps: carrying out the segment division of an input video, and carrying out the feature extraction through a pre-training feature extractor; segmenting the global feature sequence into a predetermined number of subspace heads along the channel dimension; constructing a local autocorrelation matrix, and estimating a background principal component direction corresponding to each subspace by using a micro-power iterative algorithm; and inputting the spectral residual features into a semantic perception gating network to generate a mask, and enhancing the global feature sequence by using the mask to obtain enhanced output features. According to the method, a normal background is modeled as a low-rank principal component molecular space in a feature manifold, and an anomaly is modeled as a sparse and high-energy spectrum residual, so that semantic decoupling of the background and the anomaly is effectively realized.

The application provides a kind of optical fiber communication system carrier phase recovery method and device.Method includes using linear time domain data filtering to carry out predicted receiving signal carrier phase noise in the receiving end of optical fiber communication system;According to the predicted carrier phase noise, the receiving signal is pre-rotated, and the carrier phase noise is pre-compensated;Residual error of the receiving signal after pre-rotation is estimated by power iteration method, for extracting the error between predicted carrier phase noise and actual phase noise;According to the error between the predicted phase noise and the actual phase noise extracted, the next state quantity of linear time domain data filtering is updated, and the carrier phase recovery is realized.The method can effectively solve the carrier phase recovery problem in optical fiber communication system, and has good algorithm robustness under low signal-to-noise ratio condition, and low computational complexity.