12 results about "Matrix group" patented technology

The present application relates to the technical field of distributed system testing and operation and maintenance monitoring, and particularly relates to a server test environment automatic monitoring method and system, comprising: collecting a micro-service cluster directed call topology structure, constructing a distributed cell layer with node and edge vector spaces; parsing data serialization rules to extract constraint restriction mapping matrices, assembling a Hodge-Dirac operator; obtaining real-time business state vector groups to splice into zero-order co-chains, using the operator to solve an algebraic Laplace matrix to calculate a quadratic form energy functional; when determining that the quadratic form energy functional is greater than zero, generating an early warning signal, constructing and solving a discrete layer diffusion dynamics equation based on the algebraic Laplace matrix, obtaining a state compensation update matrix; finally, analyzing the state compensation update matrix into reverse rollback and cache erasure instructions, and submitting the instructions to a service mesh data plane proxy component for execution. The present application realizes code-free cross-node semantic conflict quantification and distributed consistency state self-healing closed loop.

The embodiment of the invention belongs to the technical field of computers, and particularly relates to an operator execution method, a processor and computer equipment. The method is applied to a processor, and the processor is provided with a control unit, a vector operation unit and a matrix operation unit. The control unit obtains a plurality of matrix groups of operators to be executed. And the control unit controls the vector operation unit and the matrix operation unit to execute the first operation and the second operation included in the operators on the plurality of matrix groups in batches so as to obtain an execution result of executing the operators on each matrix group. Wherein when the first operation and the second operation are executed for the Nth time, the at least one vector operation unit executes the first operation on the first matrix group, the at least one matrix operation unit executes the second operation on the second matrix group, the first matrix group and the second matrix group are different matrix groups, and N is greater than 1 and less than M; m is the total batch for executing the first operation and the second operation. By adopting the method, the execution efficiency of the processor on the operator can be improved.

A robust machine learning system for robust machine learning is provided. The robust machine learning system includes: a processor; and a memory storing a robust machine learning program and a group representation network (GRepsNet), where the robust machine learning program runs on the processor. In this case, the processor is configured to: receive, via an input layer connected to the GrepsNet, an input data set represented by a first group representation; the input dataset is communicated / transmitted by a GRepsNet configured to generate an output dataset represented by an output group representation from the input dataset, where the output group is equivariant with the first group, and the output dataset is generated from an output layer connected to the GRepsNet.

The application provides a general information extraction method based on task-specific mixed low-rank adaptation, comprising the following steps: selecting a pre-trained model as a base model and initializing the base model; initializing a first matrix group and a second matrix group for a weight matrix of the base model; performing forward propagation calculation on all low-rank decomposition matrices; constructing a gating matrix, fusing features and a fourth output vector to generate a fifth output vector through vector splicing; constructing a joint loss function for optimization; combining the matrices to generate a target model; inputting a natural language text information task into the target model to output a structured information extraction result. The application has the beneficial effects of effectively alleviating the task interference and model over-coupling problems of the existing multi-task learning method, thereby improving the performance of the model in the general information extraction task; integrating multiple low-rank matrices and gating matrices to improve the adaptability of the model to different task characteristics and ensure effective cross-task knowledge transfer.

The invention discloses a concrete uniaxial compressive strength prediction method based on multi-dimensional matrix energy, and the method comprises the steps: instantiating original nondestructive testing data into a three-dimensional measurement matrix group, extracting a multi-dimensional matrix energy feature set which can distinguish the macroscopic compactness and local non-uniformity of a material, and taking the multi-dimensional matrix energy feature set as an independent variable to drive a model for iterative optimization. Through the mode, the deviation degree of the secondary effect relative to the dominant effect is effectively measured, and a unified feature scale suitable for samples of different strength grades is constructed, so that the physical interpretability, the robustness and the prediction precision of a concrete strength prediction model are remarkably improved.

The present application belongs to the technical field of photoelectric countermeasure, and relates to a detection technology for a hidden camera device, and particularly provides a long-distance detection method for a hidden camera device based on sparse representation, which is used to overcome the problem of low recognition accuracy in the prior art; in the present application, after sparse representation of an image through wavelet transform, a measurement matrix group is obtained by compressing and projecting the sparse image using a random measurement matrix; a measurement matrix without static background and dynamic background information is obtained by performing a series of difference addition operations on the matrix group; then, the orthogonal matching pursuit algorithm is used to reconstruct and recover the measurement matrix to obtain a recovered image; finally, the cat-eye target area is screened out according to the image pixel value and the connected area area judgment method, so as to realize the cat-eye effect target recognition method based on sparse representation; in this way, the image data redundancy is greatly reduced, the recognition accuracy of the long-distance cat-eye target is effectively improved, and the anti-detection of the hidden camera device is realized, so as to eliminate the potential information leakage danger.

An information processing method and device, wherein each n columns of an LDPC base matrix correspond to a sub-matrix group, each sub-matrix group includes W sub-matrices, and the W sub-matrices can be obtained from a first matrix, that is, the LDPC base matrix can be obtained from the first matrix. In this way, the LDPC base matrix can not be stored, but the first matrix can be stored, the first matrix is much smaller than the LDPC base matrix in size, and thus the occupation of the storage space can be reduced. Since the LDPC base matrix can be obtained from the first matrix, the first matrix can be obtained in a smaller search space when constructing the matrix, and then the LDPC base matrix is obtained according to the first matrix, and thus the difficulty of matrix construction can be reduced.

Disclosed are an information processing method and apparatus. Every n columns of an LDPC base matrix correspond to one sub-matrix group, each sub-matrix group comprises W sub-matrices, and the W sub-matrices can be obtained from a first matrix, that is, the LDPC base matrix can be obtained from the first matrix. In this way, the LDPC base matrix may not be stored, but the first matrix is stored. The size of the first matrix is much smaller than that of the LDPC base matrix, thereby reducing the occupied storage space. As the LDPC base matrix may be obtained from the first matrix, during matrix construction, the first matrix may be obtained in a small search space, and then the LDPC base matrix is obtained on the basis of the first matrix, thereby reducing the difficulty of matrix construction.

The application discloses a kind of multi-energy complementary system double-layer energy optimization scheduling method, comprising: based on Nataf conversion obtains the correlation coefficient matrix of each variable in normal space and obtains sample matrix group;With sample matrix group as input variable, 2PEM evaluation and series expansion fitting are carried out, and probability density function is obtained;Quadratic programming model is constructed, and quadratic programming model is solved based on improved efficient set method;Based on risk aversion strategy, construct robust optimization model, avoid the influence of wind light load uncertainty on energy storage configuration result;MPC rolling optimization model is constructed to reduce deviation;Based on MPC rolling optimization model and robust optimization model, construct double-layer energy optimization scheduling framework;Based on double-layer energy optimization scheduling framework, correct day-ahead scheduling plan.The application adopts MPC to carry out rolling optimization, and according to actual situation, deviation is corrected on the basis of day-ahead IGDT robust scheduling plan, and the shortcomings of IGDT open-loop control are improved.

The invention relates to the technical field of model optimization, in particular to an aggregation Markov model dimension reduction method and device for dynamic system monitoring, and the method comprises the steps: firstly obtaining a probability transfer matrix Q, an observation probability matrix O and an initial distribution vector p of a target model; generating a diagonal matrix based on O, and constructing an intermediate matrix M fusing dynamic and initial information of the model; calculating the maximum matching number E of the bipartite graph constructed by the M, and comparing the maximum matching number E with the model dimension n to judge whether the dimension of the model can be reduced or not; if the dimension can be reduced, for each observation category, constructing a conversion sub-matrix based on a bipartite graph matching result of the sub-matrix blocks, and combining to form a conversion matrix T; and finally, constructing a dimensionality reduction aggregation Markov model with lower dimensionality by utilizing T and a left inverse matrix thereof. The invention provides a determinable automatic dimension reduction mechanism, which can obviously reduce the calculation complexity on the premise of ensuring the function equivalence of the model, and is particularly suitable for edge calculation scenes with limited resources.

The present application relates to a water turbine unit equipment rotation method based on group representation theory and rotation voting algorithm, comprising: establishing a permutation group having an isomorphic relationship with the unit equipment rotation logic; constructing a linear space and a base vector corresponding to the permutation group; determining the state and level of the unit equipment; determining the conjugate class of the permutation group and the corresponding matrix group element; determining the trigger condition of the unit equipment rotation; and calculating the new order of the main and standby equipment in a descending order k-rotation manner. The equipment rotation voting algorithm of the present application realizes the scientific and reasonable rotation and standby of the water turbine unit equipment, and the number of equipment used for rotation is not limited, improving the reliability of the water turbine unit operation, and solving the problems of algorithm complexity, program logic redundancy, poor readability, and difficult to find and troubleshoot logical errors in the design of the equipment rotation logic algorithm relying on experience and enumeration method.

A kind of impedance model computer automatic construction method for multi-terminal flexible system.The method first obtains the steady state operating point of each element of system by power flow analysis, then establishes mathematical model for common elements of multi-terminal flexible system, and linearizes, discretizes based on steady state operating point, to establish element small signal state space model;Subsequently, the generalized incidence matrix group compatible with ac-dc hybrid system is used to realize element aggregation, obtain subsystem model and calculate its z domain impedance model, finally generate s domain impedance model by bilinear transformation.The invention realizes the computer-aided automatic construction of multi-terminal flexible system impedance model, supports the impedance fast analysis of any ac / dc port, is flexible and efficient;And the topological sensitivity of system is low, when the structure changes, only the corresponding part of generalized incidence matrix group needs to be modified, and the scalability is high.Example analysis shows that the invention can accurately obtain the ac / dc port impedance of multi-terminal flexible system, and effectively improve the analysis efficiency.