146 results about "Block matrix" patented technology

The invention provides a large model interaction method and system based on multi-model collaborative dialogue, and relates to the technical field of artificial intelligence, and the method comprises the steps: dividing multi-modal data into a pixel block matrix and extracting feature vectors, constructing a hexagonal honeycomb-shaped coding matrix to store data information, and generating an attention score matrix by adopting a multi-head attention mechanism; constructing a model communication network based on the transmission data unit, and calculating a professional degree score through graph convolution; according to the method, performance data is collected to generate an evaluation graph, and an optimal node cooperation strategy is output in combination with dominant learning, so that the efficiency and accuracy of multi-model cooperation dialogue are improved, and the robustness and expandability of the system are enhanced.

The invention provides a quantum circuit optimization method, device and equipment for a ZUC algorithm subcomponent, and the method comprises the steps: encoding a signal transmitted by communication into quantum bits, and inputting the quantum bits into a ZUC algorithm structure for quantum implementation; representing a first linear subcomponent in the ZUC algorithm structure as a matrix in a binary field; after a matrix in the binary field is expressed as a block matrix, a low-dimensional matrix M in the binary field is derived from the block matrix, and quantum implementation of the M is searched by using a search algorithm designed based on a matrix decomposition principle; and according to a quantum gate parallel rule, carrying out optimized multi-layer division on the quantum of the M again, obtaining the quantum of the M, carrying out optimized layering operation, and optimizing the quantum of the first linear subcomponent. According to the method provided by the invention, the quantum circuit of the subcomponent of the ZUC algorithm is optimized, so that the resource overhead of quantum implementation of the ZUC algorithm is saved.

The invention discloses a multi-core-particle optical matrix calculation framework, which is characterized in that a block matrix is adopted to decompose a large-scale optical matrix calculation chip into a plurality of small-scale optical matrix calculation core particles, and active and passive separation is adopted to decompose the optical matrix calculation chip into modulator array core particles, detector array core particles and optical matrix calculation core particles. Decomposing the M * N matrix A into m * n p * q matrixes Aij; an input N-dimensional vector X is decomposed into n-channel q-dimensional vectors Xj, and an output M-dimensional vector Y is decomposed into m p-dimensional vectors Yi. Input light passes through the n modulator array core particles to form n channel q-dimensional vectors, m * n q-dimensional vectors are formed after m beam splitting copying, the m * n q-dimensional vectors are input into the m * n optical matrix calculation core particles respectively and multiplied by Aij to obtain m * n p-dimensional vectors, m p-dimensional vectors are obtained after the light of the n channels passes through n channel multiplexing (wavelength division multiplexing, polarization multiplexing, mode multiplexing and the like and multi-dimensional multiplexing), and m * n p-dimensional vectors are obtained. And then detection is carried out, and large-scale matrix-vector multiplication calculation Y = AX is realized.

The invention provides a BLAS3 structured operator accelerated computing system based on a Hopper architecture GPU, and relates to the technical field of computers. The system comprises: a calculation unit discrimination module for determining a calculation unit used by a current operator during operation, and estimating the maximum row dimension upper bound of the current operator in a tensor core execution path; an instruction sensing block parameter determination module dynamically determines the optimal block size and number of the input matrix in real time; the block matrix loading and aligning module divides an input matrix and a matrix to be updated into sub-matrixes by taking the block size as a basic block and completes loading of the corresponding sub-matrixes; the operator kernel function execution module completes shared memory structured parallel loading and storage of a double-precision floating-point number array of a sub-matrix corresponding to the input matrix, and calls a tensor core to carry out multiply-add accumulation calculation; and the assembly line and concurrent scheduling module adds the block calculation tasks into corresponding task sets and performs multi-stream concurrent scheduling on the task sets.

The invention discloses a solving optimization method and device for a trigonometric linear equation set. The method comprises the following steps: determining a to-be-solved trigonometric linear equation set as a target trigonometric linear equation set; a plurality of task blocks obtained by dividing a result vector are circularly distributed to a computing core of a processing platform; distributing the inverse matrix of each triangular block matrix on the diagonal line of the input matrix in the target triangular linear equation set and the input vector block of the input vector to the calculation core; and carrying out iterative solution by utilizing the calculation core until a result vector block of each task block is obtained, and determining a target input vector as a solution vector. According to the method, the technical problems that the core carrying waiting time and the calculation time are too long in the process of solving the trigonometric linear equations by using a traditional method in related technologies, and solving is performed one by one in a forward iteration mode on a processing platform, so that the solving efficiency is relatively low are solved.

The invention discloses an electromagnetic field simulation optimization method and system based on deep learning. The method comprises the following steps: discretizing a calculation area, and dividing the calculation area into a limited number of units; constructing a finite element matrix for the electric field of each unit; the finite element matrix is subjected to block processing, and two sub-block matrixes describing the electric field gradient and the electric field rotation of the triangular surface area are decomposed into an upper triangular matrix and a lower triangular matrix through an incomplete-like Cholesky decomposition model; inversing the upper triangular matrix and the lower triangular matrix to obtain approximate inverse of the finite element matrix; substituting the approximate inverse of the finite element matrix into a conjugate gradient algorithm, and solving to obtain an electric field of each unit; and obtaining a corresponding magnetic field according to the electric field of each unit so as to obtain an electromagnetic simulation result. According to the method, through a technical path of block dimension reduction-graph structure learning-physical constraint enhancement, the bottleneck problem that preprocessing efficiency and adaptability in electromagnetic field simulation are difficult to consider at the same time is solved.

The embodiment of the invention provides a quantification method and reasoning method of an industry large model, equipment and a storage medium. When an original weight matrix of a network layer in an industry large model is quantified, values of matrix elements on diagonals of an original Hessian matrix are obtained based on an input data matrix of the network layer and can represent activation values, and the positions of columns in the original weight matrix and the original Hessian matrix of the network layer are reordered according to an activation value descending mode. Carrying out matrix partitioning processing on the reordered first weight matrix and the first Hessian matrix respectively, and quantizing the first sub-weight matrix based on the quantization parameter of each sub-weight matrix in the first weight matrix in sequence by taking the partitioned matrix as granularity; the method comprises the steps of quantizing a first weight matrix, updating other first weight sub-matrixes which are not quantized based on the quantized first weight sub-matrix, and finally determining a quantization result of an original weight matrix of a network layer based on a quantization result of each first weight sub-matrix in the first weight matrix, so that the size of an industry large model is compressed.

The invention discloses a heterogeneous parallel fast direct solution method for accelerating the solution of complex electromagnetic problems, comprising: an MPI process on a CPU side reads a model file and model parameters after three-dimensional geometric modeling of a target; all MPI processes are evenly divided into two groups, all basis functions obtained are divided into M groups, and M×M unit block matrices are generated by using the grouped basis functions to merge and form a block matrix Z; each MPI process obtains basis function information contained in the unit block matrix required by itself according to an acquisition rule related to the MPI process group; all MPI processes complete the LU decomposition process of the block matrix Z in a heterogeneous parallel decomposition method jointly used by a CPU side and a GPU side based on the basis function information obtained by themselves; a current coefficient matrix is ​​solved by back-substitution of a matrix equation group according to the decomposition result of the obtained block matrix Z; and the three-dimensional geometric model of the target is calculated according to the current coefficient matrix to obtain near-field data and far-field data.

The invention relates to the technical field of telephone communication, and discloses a robust recursive least square adaptive filter based on inverse QR decomposition and a storage medium, and the method comprises the steps: calculating an inner product of an adaptive weight vector at a moment and an input signal vector, and obtaining an output signal at the moment of the adaptive filter; calculating a difference value between the expected signal and the output signal at the moment, obtaining an estimation error signal, and constructing a nonlinear weighting function; recursively updating and acquiring an inverse correlation matrix of the moment input signal vector based on the moment input signal vector and transpose thereof, a nonlinear weighting function, the inverse correlation matrix of the input signal vector and the covariance of the estimated error signal, and constructing a block matrix containing the inverse correlation matrix of the moment input signal vector; inverse QR matrix decomposition and unitary rotation operation are sequentially carried out on the block matrix, an array before inverse QR decomposition and an array after inverse QR decomposition at the moment are obtained, a gain vector at the moment is constructed, and a self-adaptive weight vector at the moment is updated and obtained.

The method for determining the maximum allowable communication data delay of a multi-agent formation system solves the problem of how to effectively determine the maximum allowable communication data delay, and belongs to the field of multi-agents. The present invention includes: S1, obtaining the parameters and iteration step ΔT of the pilot-follower multi-agent formation system; S2, calculating the block matrices C and E; S3, initializing the number of iterations f=1, setting the initial value of the time lag variable d0=ΔT; S4, substituting d0 into the determination condition, calculating the feasible solution of the determination condition, if there is a solution, go to S5, if there is no solution, reduce the iteration step ΔT, go to S2, or replace the Laplace matrix of the communication topology and the communication matrix of the pilot and follower, and go to S3; S5, updating the iteration step f=f+1; S6, d f‑1 =fΔT, d f‑1 Substitute the judgment conditions and calculate the feasible solutions to the judgment conditions. If there is a solution, go to S5. If there is no solution, the maximum communication delay d M is (f‑1)ΔT.

The invention relates to a power network structure weakness detection method based on multi-diagonal-block matrix decomposition, and belongs to the technical field of super-large-scale integrated circuits. According to the method, an original large-scale sparse matrix is converted into a band edge diagonal block structure and divided into a plurality of sub-matrixes capable of being solved independently by constructing a layering and blocking strategy for eliminating tree drive, and redundancy of full-matrix calculation is avoided. Meanwhile, a local approximate inverse algorithm of column norm truncation is designed, and target elements are calculated on the premise that a preset error threshold value is met. According to the method, the parallel computing architecture of the multi-core processor is fully utilized, the computing complexity is reduced by 2-3 orders of magnitude, the computing efficiency is effectively improved, the memory occupation is remarkably reduced, and a high-precision and high-efficiency solution is provided for detecting the weakness of the power network structure of the super-large-scale integrated circuit.

The invention provides an artificial intelligence model accelerated training reasoning method and device, and belongs to the field of artificial intelligence, and the method comprises the steps: firstly carrying out the grouping disassembly of a large-dimension matrix in an artificial intelligence model through employing a lamination grouping technology, and converting the large-dimension matrix into a lamination block matrix structure; then, carrying out compression decomposition on the non-diagonal block matrix by utilizing an ACA-SVD matrix decomposition technology, and representing that two small matrixes are multiplied; then diagonal matrix blocks of the obtained laminated block matrix structure are extracted in sequence, the matrix is expressed as multiplication of a plurality of block diagonal matrixes, and the block diagonal matrixes are stored respectively; when model training and reasoning are carried out, a matrix multiplication process can be converted into multiplication of a plurality of block diagonal matrixes, a matrix inversion process can be converted into inversion of the plurality of block diagonal matrixes, the block diagonal matrixes are mutually independent, a recursive relationship does not exist in an operation process, and highly parallel operation can be realized.

The invention provides an acceleration processing method and device for sparse matrix vector multiplication. The method comprises the following steps: acquiring a sparse matrix; dividing the sparse matrix into segments, and distributing threads for the segments; a matrix multiplication-accumulation instruction in a preset instruction set architecture is called, a tensor calculation core Tensor Core is used for carrying out matrix multiplication calculation of small blocks on the fragments, result data are obtained, and the result data are used for representing vector data objects of the linear equation set solution vectors. According to the method, the obtained sparse matrix is divided into the fragments and the threads are distributed, so that Tensor Core concurrent calculation is realized, the calculation efficiency is greatly improved, and the calculation time is remarkably shortened especially for solving a large-scale linear equation set.

The invention provides a quantization method and a quantization device, which can perform singular value decomposition on a first matrix formed by word embedding vectors to obtain the maximum characteristic value of each block matrix in M block matrixes, and can determine the quantity of quantization bits of each block matrix according to the maximum characteristic value of each block matrix. The method comprises the following steps of: dividing each block matrix into a plurality of block matrixes, quantifying each block matrix to obtain each quantized block matrix, combining each quantized block matrix to obtain a quantized first matrix, and representing the amount of semantic information carried in each block matrix by a maximum characteristic value of each block matrix to a certain extent, according to the embodiment of the invention, the quantized bit number can be allocated to each block matrix according to the amount of semantic information carried by each block matrix, so that different bit numbers can be allocated to the block matrixes carrying different semantic information, differential bit number allocation can be realized, semantic information loss can be reduced, and user experience can be improved.

The invention provides an image processing method and device which can be applied to the technical field of image processing and data security. The method comprises the steps of scrambling an image matrix representing a to-be-processed image in a matrix form based on a plurality of chaotic sequences corresponding to the to-be-processed image to obtain an out-of-order matrix; according to a preset size, performing block processing on the out-of-order matrix to obtain a plurality of block matrixes; for each block matrix, scrambling the element layout in the block matrix to obtain an out-of-order sequence; performing encryption and layout transformation on the out-of-order sequence by using a biological encryption method corresponding to the block matrix to obtain an encryption matrix, the biological encryption method corresponding to the block matrix being determined according to image information of sub-images corresponding to the block matrix in the to-be-processed image; and merging the encryption matrixes corresponding to the plurality of block matrixes to obtain an encrypted to-be-processed image.

The invention discloses a method for generating an amplitude preparation circuit and a quantum state preparation method and device, and the method comprises the steps: setting a main diagonal element of a block matrix as a to-be-prepared amplitude, and setting other elements as 0, so that the block matrix is equivalent to a diagonal matrix, thereby facilitating the construction of a unitary matrix through employing the block matrix, and in the unitary matrix, the unitary matrix is not damaged; the to-be-prepared amplitude is arranged at the upper left corner of the unitary matrix, so that the amplitude preparation of the to-be-prepared quantum state can be carried out by utilizing a circuit corresponding to the unitary matrix, and the preparation of the to-be-prepared quantum state by utilizing a large-scale operator can also be understood, thereby being beneficial to more efficiently and accurately preparing the to-be-prepared quantum state.

The invention discloses a rapid and direct solving method for a metal medium composite target based on a block sparse row matrix. The method comprises the following steps: approximately representing a low-rank non-diagonal block matrix as multiplication of two decomposition matrixes U and V by adopting a self-adaptive crossover algorithm; cascading the decomposition matrix of each sub-block to obtain an integral decomposition matrix; carrying out compression storage on the decomposition matrixes U and V by adopting a block sparse row matrix, and removing zero blocks generated by matrix cascading; performing inversion on the block diagonal matrix of the finest layer through a conjugate gradient method, and solving an inverse matrix of the block diagonal matrix of the non-finest layer according to an SMW formula; and finally, obtaining a current solving target RCS through vector multiplication of an inverse matrix of a series of block diagonal matrixes and a matrix of a right vector. According to the method, the electromagnetic scattering characteristics of the complex target are solved, and the calculation efficiency of analyzing the electromagnetic single-station scattering characteristics of the complex target can be improved.

The embodiment of the invention discloses a processor, a chip product, computer equipment and a data processing method. The processor comprises a block matrix calculation unit, a storage unit and a target register, wherein the storage unit is used for storing an input matrix to be processed and inverse quantization data corresponding to the input matrix; the block matrix calculation unit is used for reading the block matrix from the storage unit and reading the last target operation result from the target register; performing accumulation operation on the last target operation result and the block matrix to obtain a candidate operation result; and according to the inverse quantization parameter corresponding to the block matrix read from the storage unit, performing data precision adjustment on each element in the candidate operation result to obtain a current target operation result, and updating and storing the current target operation result in a target register.

Apparatuses and methods for multi-register block-wise matrix multiplication are disclosed. An example processor executes an instruction having an operation object field to indicate that a multi-register cross product matrix multiplication is to be performed with a first plurality of submatrices and a second plurality of submatrices to generate a third plurality of submatrices. An execution circuit executes the instruction to generate each of the third plurality of submatrices in a corresponding vector register of a third plurality of vector registers by multiplying a submatrix in a corresponding vector register of the first plurality of vector registers with a submatrix in a corresponding vector register of the second plurality of vector registers.

The invention discloses a near-field rotation invariant parameter estimation method based on linear fitting, and the method comprises the steps: calculating the cross-correlation between the receiving data of an array element in a first array and the receiving data of an array element in a second array on a signal source three-dimensional space precise propagation model based on a symmetric cross array, and obtaining a cross-correlation matrix; obtaining a virtual receiving data matrix by vectorizing the cross-correlation matrix; performing eigenvalue decomposition on the covariance matrix of the virtual receiving data matrix to obtain a signal subspace matrix; estimating parameters of a narrowband signal source by using an ESPRIT algorithm, specifically, determining a virtual uniform area array according to a covariance matrix, and partitioning a signal subspace matrix into blocks; the rotation invariance is recovered through the subspace relation of the adjacent block matrixes; obtaining the phase difference between the adjacent sub-arrays through the rotation invariance matrix, building a linear observation model, and obtaining the parameter estimation value of the narrowband signal source; the method has the advantages of low calculation complexity and high universality.

The application discloses an operation method and device based on an artificial intelligence model, equipment, a medium and a product, and relates to the technical field of computers.The method comprises the following steps: acquiring source data of target operation; wherein the source data comprises a first matrix and a second matrix; judging whether the target operation is block calculation; if yes, determining a first block size of the first matrix and a second block size of the second matrix; storing the first matrix as a plurality of first block matrices according to the first block size, and storing the second matrix as a plurality of second block matrices according to the second block size; and in the process of executing the target operation, the plurality of first block matrices and the plurality of second block matrices are read in sequence to realize the target operation and obtain a target operation result.The application improves operation efficiency and saves hardware resources.

The invention discloses a high-performance decoding method for quantum low-density parity check codes, and belongs to the technical field of quantum computing. Comprising the following steps: solving a transformation matrix and a permutation matrix through a matrix decoupling strategy optimized by an off-line satisfiability modulus theory, and equivalently transforming an original check matrix into a diagonal block matrix and any sparse matrix; the method comprises the following steps: decomposing an error mode into a left error corresponding to a diagonal block matrix and a right error corresponding to any sparse matrix through an online hierarchical greedy decoding algorithm, guessing an error bit of the right error, and then calculating an error symptom corresponding to the left error based on any sparse matrix; then the error symptom is divided according to the diagonal block matrixes for parallel decoding, decoding results of all the block matrixes are combined to form a left error, and finally the left error and the right error are spliced to obtain a complete error mode. According to the method, accurate and real-time decoding of the qLDPC code can be realized, and the method is applied to the frontier fields of quantum communication, quantum calculation and the like.

The invention discloses a privacy protection matrix multiplication calculation method based on a threshold Paillier cryptosystem. According to the method, the two participants can safely calculate the multiplication of the matrix under the condition that the data of the participants are not leaked. And the participant P1 divides the matrix into a plurality of sub-matrixes based on each element in the Paillier encryption matrix, and cooperatively executes the security matrix multiplication protocol with the participant P2 by using the block matrix. According to the method, a security matrix multiplication protocol based on a threshold Paillier cryptosystem is designed, a plurality of ciphertexts are packaged into one ciphertext, the communication traffic is reduced, parallel execution of matrix block multiplication is supported, and the calculation efficiency is improved.

Disclosed in the present invention are an image denoising method and apparatus. The image denoising method comprises: S110, initializing a pixel block matrix, an interpolation position matrix, an interpolation weight matrix, a contrast-to-noise model ratio matrix, a detection coefficient matrix and a pixel point coefficient matrix; S120, successively updating the contrast-to-noise model ratio matrix, the detection coefficient matrix and the pixel point coefficient matrix, so as to process an input current image frame; S130, on the basis of the difference between an updated previous image frame and the current image frame and an updated pixel point coefficient matrix, acquiring an updated current image frame output; and S140, repeatedly executing steps S120-S130.

The invention discloses a loess mechanical parameter spatio-temporal interpolation prediction model construction method based on big data, and relates to the technical field of geotechnical engineering data analys.The loess mechanical parameter spatio-temporal interpolation prediction model construction method comprises the following steps that loess mechanical parameter data, environmental factor data and geologic structure data are collected through a multi-source sensor, and spatio-temporal unified multi-modal data are constructed; designing a collaborative kernel function fusing spatial heterogeneity and time dynamics on the basis of multi-modal data, and quantifying relevance between mechanical parameters and space-time positions and environmental factors; according to the collaborative kernel function, dynamically optimizing a space weight, a time attenuation coefficient and a multi-source data fusion weight, and generating a parameter-adaptive interpolation model; carrying out distributed training and calculation acceleration on the interpolation model by adopting a block matrix approximation and parallel calculation architecture; through the trained model, a mechanical parameter prediction value of a target space-time position is calculated in real time, and the calculation efficiency is improved through a tensor expansion algorithm.

The present invention discloses a multi-core optical matrix computing architecture, which uses a block matrix to decompose a large-scale optical matrix computing chip into several small-scale optical matrix computing cores, and uses active-passive separation to decompose the optical matrix computing chip into modulator array cores, detector array cores, and optical matrix computing cores. The M×N matrix A is decomposed into m×n p×q matrices A ij The input N-dimensional vector X is decomposed into n-channel q-dimensional vector X j , the output M-dimensional vector Y is decomposed into m p-dimensional vectors Y i The input light passes through n modulator array core particles to form n channel q-dimensional vectors, and after m beam splitting and replication, it forms m×n q-dimensional vectors, which are input into m×n optical matrix calculation core particles respectively, and A ij After multiplication, m×n p-dimensional vectors are obtained. After the light of n channels is multiplexed through n channels (wavelength division multiplexing, polarization multiplexing, mode multiplexing, etc. and multidimensional multiplexing), m p-dimensional vectors are obtained. Then, detection is performed to realize large-scale matrix-vector multiplication calculation Y=AX.

Provided in the embodiments of the present application are an optimization method for implementing high-performance single-precision matrix multiplication by using Ascend-based half-precision computing units, and a related device. The method comprises: first, acquiring a first half-precision matrix of a first single-precision matrix, a second half-precision matrix of a second single-precision matrix, a second error matrix between the second single-precision matrix and the second half-precision matrix, and a plurality of second block error matrices in the second error matrix, and storing the second block error matrices in an L1 buffer of parallel computing hardware; and acquiring first block matrices in the first half-precision matrix one by one, storing the first block matrices in the L1 buffer, further performing a matrix operation to obtain a first result matrix, acquiring a second result matrix and a third result matrix, summing the first result matrix, the second result matrix and the third result matrix to obtain a single-precision target matrix, and using the single-precision target matrix as the result of performing a matrix multiplication operation on the first single-precision matrix and the second single-precision matrix. Thus, the accuracy and computational speed of single-precision matrix multiplication are effectively improved.

Provided in the embodiments of the present disclosure are a data repair method and apparatus. The data repair method comprises: for a distributed storage node cluster, constructing an initial check matrix for data repair; updating a target block matrix in the initial check matrix according to a matrix full-rank strategy, so as to obtain a check matrix corresponding to the distributed storage node cluster, wherein the target block matrix meets a full-rank condition of the matrix full-rank strategy; when the distributed storage node cluster has a faulty node, selecting from the check matrix a matrix element associated with the faulty node; determining a target node from the distributed storage node cluster on the basis of the matrix element, and acquiring node storage data corresponding to the target node; and calculating repair data of the faulty node on the basis of the matrix element and the node storage data, and uploading the repair data to the faulty node. A repair function for faulty nodes is implemented while reducing node repair bandwidth and hard disk consumption.

The application discloses a kind of pre-decarbonization device operation anomaly detection method based on instant ARRG model, to establish adaptive AR model in real time using the online sampling data of pre-decarbonization device to generate error, then real-time error generated is used to complete operation anomaly detection task.Specifically, the method of the present application establishes an instant ARRG model in real time for each sampling time block matrix, and then calculates the error using the corresponding conversion vector and AR coefficient vector, so as to implement the operation anomaly detection of the pre-decarbonization device by the up and down fluctuation of the error.Different from the traditional method which relies on fixed time sequence relationship to describe the model, the error monitored in real time by the method of the present application is generated by the adaptive instant ARRG model, and is targeted for the sample data of each sampling time, can adapt to different characteristics of different sample data, and can always adaptively generate appropriate error for anomaly detection.

The present invention discloses a method for estimating the rotation parameters of small celestial bodies by integrating a camera and a lidar, comprising the following steps: defining a coordinate system, establishing a measurement model for integrating the camera and the lidar, defining an observation block matrix, and fusing various observation information through the block matrix to realize the estimation of the rotation parameters of small celestial bodies. The present invention improves the estimation accuracy of the spin parameters of small celestial bodies by introducing the depth information of feature points, and simultaneously calculates the position and velocity of the small celestial body relative to the detector.